diff --git a/.gitignore b/.gitignore
deleted file mode 100644
index 325d7206680f0197cdd448578d09bbc4c7455afd..0000000000000000000000000000000000000000
--- a/.gitignore
+++ /dev/null
@@ -1,20 +0,0 @@
-*~
-*.swp
-*.pyc
-bin
-eggs
-parts
-.installed.cfg
-.mr.developer.cfg
-*.egg-info
-src
-develop-eggs
-sphinx
-dist
-.nfs*
-.gdb_history
-build
-.coverage
-record.txt
-miniconda.sh
-miniconda/
\ No newline at end of file
diff --git a/MANIFEST.in b/MANIFEST.in
deleted file mode 100644
index 82d409243bd964985baf51af5dbe8c803f935684..0000000000000000000000000000000000000000
--- a/MANIFEST.in
+++ /dev/null
@@ -1,2 +0,0 @@
-include README.rst buildout.cfg COPYING version.txt requirements.txt
-recursive-include doc *.rst *.png *.ico *.txt
\ No newline at end of file
diff --git a/README.rst b/README.rst
deleted file mode 100644
index d96a5574f45da701ba9b1a0f28d18ce13a72cac1..0000000000000000000000000000000000000000
--- a/README.rst
+++ /dev/null
@@ -1,48 +0,0 @@
-.. -*- coding: utf-8 -*-
-
-.. image:: https://img.shields.io/badge/docs-stable-yellow.svg
- :target: https://www.idiap.ch/software/bob/docs/bob/bob.paper.synthetic_dataset/stable/index.html
-.. image:: https://img.shields.io/badge/docs-latest-orange.svg
- :target: https://www.idiap.ch/software/bob/docs/bob/bob.paper.synthetic_dataset/master/index.html
-.. image:: https://gitlab.idiap.ch/bob/bob.paper.synthetic_dataset/badges/master/pipeline.svg
- :target: https://gitlab.idiap.ch/bob/bob.paper.synthetic_dataset/commits/master
-.. image:: https://gitlab.idiap.ch/bob/bob.paper.synthetic_dataset/badges/master/coverage.svg
- :target: https://www.idiap.ch/software/bob/docs/bob/bob.paper.synthetic_dataset/master/coverage/index.html
-.. image:: https://img.shields.io/badge/gitlab-project-0000c0.svg
- :target: https://gitlab.idiap.ch/bob/bob.paper.synthetic_dataset
-.. image:: https://img.shields.io/pypi/v/bob.paper.synthetic_dataset.svg
- :target: https://pypi.python.org/pypi/bob.paper.synthetic_dataset
-
-
-=============
- New package
-=============
-
-This package is part of the signal-processing and machine learning toolbox Bob_.
-
-.. todo::
-
- **Complete the sentence above to include one phrase about your
- package! Once this is done, delete this to-do!**
-
-
-Installation
-------------
-
-Complete bob's `installation`_ instructions. Then, to install this
-package, run::
-
- $ conda install bob.paper.synthetic_dataset
-
-
-Contact
--------
-
-For questions or reporting issues to this software package, contact our
-development `mailing list`_.
-
-
-.. Place your references here:
-.. _bob: https://www.idiap.ch/software/bob
-.. _installation: https://www.idiap.ch/software/bob/install
-.. _mailing list: https://www.idiap.ch/software/bob/discuss
\ No newline at end of file
diff --git a/buildout.cfg b/buildout.cfg
deleted file mode 100644
index dd66717567e7575cbb71956f289803ab127ccbf8..0000000000000000000000000000000000000000
--- a/buildout.cfg
+++ /dev/null
@@ -1,14 +0,0 @@
-; -*- coding: utf-8 -*-
-; Thu Apr 8 16:01:16 2021
-
-[buildout]
-parts = scripts
-develop = .
-eggs = bob.paper.synthetic_dataset
-extensions = bob.buildout
-newest = false
-verbose = true
-
-[scripts]
-recipe = bob.buildout:scripts
-dependent-scripts = true
\ No newline at end of file
diff --git a/compute_latent_directions.py b/compute_latent_directions.py
deleted file mode 100644
index feb2305c2666f7182784c56cd8bfdc1f0d767875..0000000000000000000000000000000000000000
--- a/compute_latent_directions.py
+++ /dev/null
@@ -1,60 +0,0 @@
-import os
-from bob.extension import rc
-import click
-from bob.extension.scripts.click_helper import (
- ResourceOption,
- ConfigCommand,
-)
-import pickle
-from lib.latent import analysis
-
-@click.command(
- cls=ConfigCommand,
- help="Compute and save latent directions starting from precomputed latent projections of MultiPIE",
-)
-@click.option(
- "--projections_dir",
- "-i",
- type=str,
- default=rc['bob.synface.multipie_projections'],
- help="Root of directory containing Multipie projections",
- cls=ResourceOption,
-)
-@click.option(
- "--output_path",
- "-o",
- type=str,
- default=rc['bob.synface.latent_directions'],
- help="Path where to store the latent directions (pickle file)",
- cls=ResourceOption,
-)
-@click.option(
- "--force",
- "-f",
- is_flag=True,
- cls=ResourceOption,
- help="`Activate flag to overwrite computed directions if the file already exist",
-)
-def compute_latents(projections_dir=rc['bob.synface.multipie_projections'],
- output_path=rc['bob.synface.latent_directions'],
- force=False,
- **kwargs):
-
- if (not os.path.exists(output_path)) or force:
- out = {}
- for covariate in ['illumination','expression','pose']:
- print('Analyzing {} covariate ...'.format(covariate))
- print(' Loading projected latents ...')
- df = analysis.load_latents(covariate)
- print(' SVM fitting ...')
- result = analysis.covariate_analysis(covariate, df)
- out[covariate] = result
-
- with open(output_path, 'wb') as f:
- pickle.dump(out, f)
-
- else:
- print('Computed directions are already found under {}. Use the --force flag to overwrite them.'.format(output_path))
-
-if __name__ == "__main__":
- compute_latents()
\ No newline at end of file
diff --git a/config/db_gen/syn_multipie.py b/config/db_gen/syn_multipie.py
deleted file mode 100644
index 2e173127e1296a814daf9a90da3e471ac5c30506..0000000000000000000000000000000000000000
--- a/config/db_gen/syn_multipie.py
+++ /dev/null
@@ -1,7 +0,0 @@
-ict=0.1
-seed=0
-pose_scaling=1
-illumination_scaling=1
-expression_scaling=1
-color_variations=0
-num_identities=64+65
\ No newline at end of file
diff --git a/config/db_gen/uniqueness_ict_0.1.py b/config/db_gen/uniqueness_ict_0.1.py
deleted file mode 100644
index 7c9d31d4bf3ba8447ae9c3a570e7449399408ce2..0000000000000000000000000000000000000000
--- a/config/db_gen/uniqueness_ict_0.1.py
+++ /dev/null
@@ -1,4 +0,0 @@
-ict=0.1
-seed=0
-task='references'
-num_identities=11000
\ No newline at end of file
diff --git a/config/db_gen/uniqueness_no_ict.py b/config/db_gen/uniqueness_no_ict.py
deleted file mode 100644
index 350a6c8659e0921df83b3dfe1a1f48efb3f1ad98..0000000000000000000000000000000000000000
--- a/config/db_gen/uniqueness_no_ict.py
+++ /dev/null
@@ -1,4 +0,0 @@
-ict=0.0
-seed=0
-task='references'
-num_identities=11000
\ No newline at end of file
diff --git a/config/project/multipie_E.py b/config/project/multipie_E.py
deleted file mode 100644
index d43eac9a824d8a215c66fadf048fca6f086f023d..0000000000000000000000000000000000000000
--- a/config/project/multipie_E.py
+++ /dev/null
@@ -1,14 +0,0 @@
-import os
-from bob.extension import rc
-from bob.bio.face.database import MultipieBioDatabase
-
-
-database = MultipieBioDatabase(
- original_directory=rc['bob.db.multipie.directory'],
- annotation_directory=rc['bob.db.multipie.annotations'],
- protocol='E_lit',
- training_depends_on_protocol=True,
- )
-
-seed=0
-num_steps=1000
diff --git a/config/project/multipie_P.py b/config/project/multipie_P.py
deleted file mode 100644
index aee3e2de5574d4a7eb40d6c6a8651ff5d5f9fd1e..0000000000000000000000000000000000000000
--- a/config/project/multipie_P.py
+++ /dev/null
@@ -1,16 +0,0 @@
-import os
-from bob.extension import rc
-from bob.bio.face.database import MultipieBioDatabase
-
-cameras= ["04_1", "05_0", "05_1", "13_0", "14_0"] + ["08_0", "19_0"] # Frontal + Half profile, no elevation
-
-database = MultipieBioDatabase(
- original_directory=rc['bob.db.multipie.directory'],
- annotation_directory=rc['bob.db.multipie.annotations'],
- protocol='P_lit',
- training_depends_on_protocol=True,
- all_files_options={'cameras': cameras}
- )
-
-seed=0
-num_steps=1000
\ No newline at end of file
diff --git a/config/project/multipie_U.py b/config/project/multipie_U.py
deleted file mode 100644
index 694ac317d53be0b9a57a45825421fe4bfecfa586..0000000000000000000000000000000000000000
--- a/config/project/multipie_U.py
+++ /dev/null
@@ -1,13 +0,0 @@
-import os
-from bob.extension import rc
-from bob.bio.face.database import MultipieBioDatabase
-
-database = MultipieBioDatabase(
- original_directory=rc['bob.db.multipie.directory'],
- annotation_directory=rc['bob.db.multipie.annotations'],
- protocol='U',
- training_depends_on_protocol=True,
- )
-
-seed=0
-num_steps=1000
diff --git a/develop.cfg b/develop.cfg
deleted file mode 100644
index 21fd3010dbd107f8c12a0f604401c9b8f35112c8..0000000000000000000000000000000000000000
--- a/develop.cfg
+++ /dev/null
@@ -1,25 +0,0 @@
-; -*- coding: utf-8 -*-
-; Thu Apr 8 16:01:16 2021
-
-[buildout]
-parts = scripts
-develop = .
- src/bob.db.multipie
-
-eggs = bob.paper.synthetic_dataset
- bob.db.multipie
-
-extensions = bob.buildout
- mr.developer
-
-newest = false
-verbose = true
-auto-checkout = *
-
-[scripts]
-recipe = bob.buildout:scripts
-dependent-scripts = true
-
-
-[sources]
-bob.db.multipie = git git@gitlab.idiap.ch:bob/bob.db.multipie
\ No newline at end of file
diff --git a/doc/conf.py b/doc/conf.py
deleted file mode 100644
index 9e8a3c6efb27fda0cdf2c56890ebc05f0dc19f32..0000000000000000000000000000000000000000
--- a/doc/conf.py
+++ /dev/null
@@ -1,239 +0,0 @@
-#!/usr/bin/env python
-# -*- coding: utf-8 -*-
-
-import os
-import time
-
-import pkg_resources
-import sphinx_rtd_theme
-
-# For inter-documentation mapping:
-from bob.extension.utils import link_documentation
-from bob.extension.utils import load_requirements
-
-# -- General configuration -----------------------------------------------------
-
-# If your documentation needs a minimal Sphinx version, state it here.
-needs_sphinx = "1.3"
-
-# Add any Sphinx extension module names here, as strings. They can be extensions
-# coming with Sphinx (named 'sphinx.ext.*') or your custom ones.
-extensions = [
- "sphinx.ext.todo",
- "sphinx.ext.coverage",
- "sphinx.ext.ifconfig",
- "sphinx.ext.autodoc",
- "sphinx.ext.autosummary",
- "sphinx.ext.doctest",
- "sphinx.ext.graphviz",
- "sphinx.ext.intersphinx",
- "sphinx.ext.napoleon",
- "sphinx.ext.viewcode",
- "sphinx.ext.mathjax",
- # 'matplotlib.sphinxext.plot_directive',
-]
-
-# Be picky about warnings
-nitpicky = True
-
-# Ignores stuff we can't easily resolve on other project's sphinx manuals
-nitpick_ignore = []
-
-# Allows the user to override warnings from a separate file
-if os.path.exists("nitpick-exceptions.txt"):
- for line in open("nitpick-exceptions.txt"):
- if line.strip() == "" or line.startswith("#"):
- continue
- dtype, target = line.split(None, 1)
- target = target.strip()
- nitpick_ignore.append((dtype, target))
-
-# Always includes todos
-todo_include_todos = True
-
-# Generates auto-summary automatically
-autosummary_generate = True
-
-# Create numbers on figures with captions
-numfig = True
-
-# If we are on OSX, the 'dvipng' path maybe different
-dvipng_osx = "/Library/TeX/texbin/dvipng"
-if os.path.exists(dvipng_osx):
- pngmath_dvipng = dvipng_osx
-
-# Add any paths that contain templates here, relative to this directory.
-templates_path = ["_templates"]
-
-# The suffix of source filenames.
-source_suffix = ".rst"
-
-# The encoding of source files.
-# source_encoding = 'utf-8-sig'
-
-# The master toctree document.
-master_doc = "index"
-
-# General information about the project.
-project = "bob.paper.synthetic_dataset"
-
-copyright = "%s, Idiap Research Institute" % time.strftime("%Y")
-
-# Grab the setup entry
-distribution = pkg_resources.require(project)[0]
-
-# The version info for the project you're documenting, acts as replacement for
-# |version| and |release|, also used in various other places throughout the
-# built documents.
-#
-# The short X.Y version.
-version = distribution.version
-# The full version, including alpha/beta/rc tags.
-release = distribution.version
-
-# The language for content autogenerated by Sphinx. Refer to documentation
-# for a list of supported languages.
-# language = None
-
-# There are two options for replacing |today|: either, you set today to some
-# non-false value, then it is used:
-# today = ''
-# Else, today_fmt is used as the format for a strftime call.
-# today_fmt = '%B %d, %Y'
-
-# List of patterns, relative to source directory, that match files and
-# directories to ignore when looking for source files.
-exclude_patterns = ["links.rst"]
-
-# The reST default role (used for this markup: `text`) to use for all documents.
-# default_role = None
-
-# If true, '()' will be appended to :func: etc. cross-reference text.
-# add_function_parentheses = True
-
-# If true, the current module name will be prepended to all description
-# unit titles (such as .. function::).
-# add_module_names = True
-
-# If true, sectionauthor and moduleauthor directives will be shown in the
-# output. They are ignored by default.
-# show_authors = False
-
-# The name of the Pygments (syntax highlighting) style to use.
-pygments_style = "sphinx"
-
-# A list of ignored prefixes for module index sorting.
-# modindex_common_prefix = []
-
-# Some variables which are useful for generated material
-project_variable = project.replace(".", "_")
-short_description = "New package"
-owner = ["Idiap Research Institute"]
-
-# -- Options for HTML output ---------------------------------------------------
-
-
-html_theme = "sphinx_rtd_theme"
-
-# Theme options are theme-specific and customize the look and feel of a theme
-# further. For a list of options available for each theme, see the
-# documentation.
-# html_theme_options = {}
-
-# Add any paths that contain custom themes here, relative to this directory.
-html_theme_path = [sphinx_rtd_theme.get_html_theme_path()]
-
-# The name for this set of Sphinx documents. If None, it defaults to
-# "<project> v<release> documentation".
-# html_title = None
-
-# A shorter title for the navigation bar. Default is the same as html_title.
-# html_short_title = project_variable
-
-# The name of an image file (relative to this directory) to place at the top
-# of the sidebar.
-html_logo = "img/bob-logo.png"
-
-# The name of an image file (within the static path) to use as favicon of the
-# docs. This file should be a Windows icon file (.ico) being 16x16 or 32x32
-# pixels large.
-html_favicon = "img/bob-favicon.ico"
-
-# Add any paths that contain custom static files (such as style sheets) here,
-# relative to this directory. They are copied after the builtin static files,
-# so a file named "default.css" will overwrite the builtin "default.css".
-# html_static_path = ['_static']
-
-# If not '', a 'Last updated on:' timestamp is inserted at every page bottom,
-# using the given strftime format.
-# html_last_updated_fmt = '%b %d, %Y'
-
-# If true, SmartyPants will be used to convert quotes and dashes to
-# typographically correct entities.
-# html_use_smartypants = True
-
-# Custom sidebar templates, maps document names to template names.
-# html_sidebars = {}
-
-# Additional templates that should be rendered to pages, maps page names to
-# template names.
-# html_additional_pages = {}
-
-# If false, no module index is generated.
-# html_domain_indices = True
-
-# If false, no index is generated.
-# html_use_index = True
-
-# If true, the index is split into individual pages for each letter.
-# html_split_index = False
-
-# If true, links to the reST sources are added to the pages.
-# html_show_sourcelink = True
-
-# If true, "Created using Sphinx" is shown in the HTML footer. Default is True.
-# html_show_sphinx = True
-
-# If true, "(C) Copyright ..." is shown in the HTML footer. Default is True.
-# html_show_copyright = True
-
-# If true, an OpenSearch description file will be output, and all pages will
-# contain a <link> tag referring to it. The value of this option must be the
-# base URL from which the finished HTML is served.
-# html_use_opensearch = ''
-
-# This is the file name suffix for HTML files (e.g. ".xhtml").
-# html_file_suffix = None
-
-# Output file base name for HTML help builder.
-htmlhelp_basename = project_variable + "_doc"
-
-# -- Post configuration --------------------------------------------------------
-
-# Included after all input documents
-rst_epilog = """
-.. |project| replace:: Bob
-.. |version| replace:: %s
-.. |current-year| date:: %%Y
-""" % (
- version,
-)
-
-
-# Default processing flags for sphinx
-autoclass_content = "class"
-autodoc_member_order = "bysource"
-autodoc_default_options = {
- "members": True,
- "undoc-members": True,
- "show-inheritance": True,
-}
-
-
-sphinx_requirements = "extra-intersphinx.txt"
-if os.path.exists(sphinx_requirements):
- intersphinx_mapping = link_documentation(
- additional_packages=["python", "numpy"] + load_requirements(sphinx_requirements)
- )
-else:
- intersphinx_mapping = link_documentation()
\ No newline at end of file
diff --git a/doc/img/bob-128x128.png b/doc/img/bob-128x128.png
deleted file mode 100644
index 787d77b3ec4f143f8dd66162227b5f385f5183f1..0000000000000000000000000000000000000000
Binary files a/doc/img/bob-128x128.png and /dev/null differ
diff --git a/doc/img/bob-favicon.ico b/doc/img/bob-favicon.ico
deleted file mode 100644
index 4cc3264302627d40868261add69eb755856611b6..0000000000000000000000000000000000000000
Binary files a/doc/img/bob-favicon.ico and /dev/null differ
diff --git a/doc/img/bob-logo.png b/doc/img/bob-logo.png
deleted file mode 100644
index b60858a7068bf45c1ed8e3da12fe244ccdcfe85d..0000000000000000000000000000000000000000
Binary files a/doc/img/bob-logo.png and /dev/null differ
diff --git a/doc/index.rst b/doc/index.rst
deleted file mode 100644
index 6f21e670179c1a25f64c2cf7b16886945bd25bb4..0000000000000000000000000000000000000000
--- a/doc/index.rst
+++ /dev/null
@@ -1,25 +0,0 @@
-.. -*- coding: utf-8 -*-
-
-.. _bob.paper.synthetic_dataset:
-
-=============
- New package
-=============
-
-.. todo ::
- Write here a small (1 paragraph) introduction explaining this project. See
- other projects for examples.
-
-
-
-Users Guide
-===========
-
-.. toctree::
- :maxdepth: 2
-
- api
-
-.. todolist::
-
-.. include:: links.rst
\ No newline at end of file
diff --git a/doc/links.rst b/doc/links.rst
deleted file mode 100644
index d22f71b499c99e9c5fde02659e44c5b93b689d76..0000000000000000000000000000000000000000
--- a/doc/links.rst
+++ /dev/null
@@ -1,8 +0,0 @@
-.. -*- coding: utf-8 -*-
-
-.. This file contains all links we use for documentation in a centralized place
-
-.. _idiap: http://www.idiap.ch
-.. _bob: http://www.idiap.ch/software/bob
-.. _installation: https://www.idiap.ch/software/bob/install
-.. _mailing list: https://www.idiap.ch/software/bob/discuss
\ No newline at end of file
diff --git a/environment.yml b/environment.yml
deleted file mode 100644
index 997a39f38c8b3396123a3d876d037eed728fbbd4..0000000000000000000000000000000000000000
--- a/environment.yml
+++ /dev/null
@@ -1,25 +0,0 @@
-name: synface
-channels:
- - https://www.idiap.ch/software/bob/conda/label/beta
- - defaults
- - https://www.idiap.ch/software/bob/conda
-dependencies:
- - python=3.7
- - bob.buildout=2.2.4
- - bob.bio.face=4.0.3
- - bob.bio.face_ongoing=1.0.6
- - bob.learn.tensorflow=1.2.0
- - numpy=1.16.4
- - pandas
- - jupyter
- - matplotlib
- - tensorflow-gpu=1.14
- - scikit-learn=0.23.2
- - scikit-image=0.16.2
- - bob.io.image
- - bob.ip.dlib
- - pillow
- - requests
- - gridtk
-
-prefix: /idiap/temp/lcolbois/miniconda3/envs/synface
diff --git a/generate_db.py b/generate_db.py
deleted file mode 100644
index de1a03045761f2b4e794d27579f59fa481658790..0000000000000000000000000000000000000000
--- a/generate_db.py
+++ /dev/null
@@ -1,208 +0,0 @@
-from bob.extension import rc
-from bob.bio.face.preprocessor import FaceCrop
-from bob.bio.face_ongoing.configs.baselines.msceleb.inception_resnet_v2.centerloss_rgb import extractor
-from lib.generate.db_generator import DBGenerator
-from lib.stylegan2.generator import StyleGAN2Generator
-from lib.stylegan2.dnnlib import tflib
-from lib.latent import analysis, editing
-from lib.utils import get_task
-import numpy as np
-
-import pickle as pkl
-from scipy.spatial.distance import cdist
-import h5py
-import os
-import time
-
-import click
-from bob.extension.scripts.click_helper import ConfigCommand, ResourceOption
-
-
-def get_postprocessing_fn():
- SG2_REYE_POS = (480, 380)
- SG2_LEYE_POS = (480, 650)
-
- SIZE = 182
- cropped_size = (SIZE, SIZE)
- IMG_REYE_POS = (SIZE / 3, SIZE / 3)
- IMG_LEYE_POS = (SIZE / 3, 2 * SIZE / 3)
- postprocessor = FaceCrop(
- cropped_image_size=cropped_size,
- cropped_positions={"reye": IMG_REYE_POS, "leye": IMG_LEYE_POS},
- color_channel="rgb",
- fixed_positions={"reye": SG2_REYE_POS, "leye": SG2_LEYE_POS},
- )
- postprocess_fn = lambda x: postprocessor(x).astype("uint8")
- return postprocess_fn
-
-
-def get_cropper():
- SIZE = 182
- IMG_REYE_POS = (SIZE / 3, SIZE / 3)
- IMG_LEYE_POS = (SIZE / 3, 2 * SIZE / 3)
- CROP_REYE_POS = (46, 53)
- CROP_LEYE_POS = (46, 107)
- cropper = FaceCrop(
- cropped_image_size=(160, 160),
- cropped_positions={"reye": CROP_REYE_POS, "leye": CROP_LEYE_POS},
- color_channel="rgb",
- fixed_positions={"reye": IMG_REYE_POS, "leye": IMG_LEYE_POS},
- )
- return cropper
-
-
-def initialization(
- ict,
- pose_scaling,
- illumination_scaling,
- expression_scaling,
- color_variations,
- output_dir,
- task_id,
- seed,
-):
- tflib.init_tf({"rnd.np_random_seed": seed, "rnd.tf_random_seed": "auto"})
-
- cropper = get_cropper()
- generator = StyleGAN2Generator(randomize_noise=True, batch_size=4)
- image_postprocessor = get_postprocessing_fn()
-
- image_dir = os.path.join(output_dir, "image")
- ref_metadata_dir = os.path.join(output_dir, "metadata", "identities")
- aug_metadata_dir = os.path.join(
- output_dir, "metadata", "samples.{}".format(task_id)
- )
-
- for folder in [image_dir, ref_metadata_dir, aug_metadata_dir]:
- if not os.path.exists(folder):
- os.makedirs(folder)
-
- covariates_scaling = {
- "pose": pose_scaling,
- "illumination": illumination_scaling,
- "expression": expression_scaling,
- }
-
- db_generator = DBGenerator(
- generator=generator,
- image_postprocessor=image_postprocessor,
- cropper=cropper,
- extractor=extractor,
- ict=ict,
- covariates_scaling=covariates_scaling,
- color_variations=color_variations,
- image_dir=image_dir,
- references_metadata_dir=ref_metadata_dir,
- augmentations_metadata_dir=aug_metadata_dir,
- seed=seed,
- )
-
- return db_generator
-
-
-@click.command(
- cls=ConfigCommand,
- help="Generate a synthetic database using semantic augmentation in the latent space.",
-)
-@click.option(
- "--num-identities",
- "-n",
- required=True,
- type=int,
- help="Number of identities in the database",
- cls=ResourceOption,
-)
-@click.option(
- "--ict",
- "-i",
- type=float,
- required=True,
- help="Minimum interclass distance between distinct identities",
- cls=ResourceOption,
-)
-@click.option(
- "--pose-scaling",
- "-ps",
- type=float,
- help="Strength of the pose augmentation",
- cls=ResourceOption,
-)
-@click.option(
- "--illumination-scaling",
- "-is",
- type=float,
- help="Strength of the illumination augmentation",
- cls=ResourceOption,
-)
-@click.option(
- "--expression-scaling",
- "-es",
- type=float,
- help="Strength of the expression augmentation",
- cls=ResourceOption,
-)
-@click.option(
- "--color-variations",
- "-c",
- type=int,
- help="Number of color variations per image",
- cls=ResourceOption,
-)
-@click.option(
- "--output-dir",
- "-o",
- type=str,
- required=True,
- help="Root of the output directory tree",
- cls=ResourceOption,
-)
-@click.option(
- "--seed",
- "-s",
- type=int,
- help="Random seed. Fix for reproducibility.",
- cls=ResourceOption,
-)
-@click.option(
- "--task",
- "-t",
- type=click.Choice(["full", "references", "augmentations"]),
- help="Subtask to execute",
- cls=ResourceOption,
-)
-def db_gen(
- num_identities,
- ict,
- output_dir,
- seed,
- task,
- pose_scaling=1.0,
- illumination_scaling=1.0,
- expression_scaling=1.0,
- color_variations=0,
- **kwargs
-):
- current_task, num_tasks = get_task()
- db_generator = initialization(
- ict=ict,
- pose_scaling=pose_scaling,
- illumination_scaling=illumination_scaling,
- expression_scaling=expression_scaling,
- color_variations=color_variations,
- output_dir=output_dir,
- task_id=current_task,
- seed=seed,
- )
-
- identities = list(range(num_identities))[current_task::num_tasks]
-
- if task == "references":
- db_generator.create_references(identities)
- elif task == "augmentations":
- db_generator.augment_identities(identities)
- else:
- db_generator.create_database(identities)
-
-
-if __name__ == "__main__":
- db_gen()
diff --git a/lib/__init__.py b/lib/__init__.py
deleted file mode 100644
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..0000000000000000000000000000000000000000
diff --git a/lib/generate/__init__.py b/lib/generate/__init__.py
deleted file mode 100644
index 8b6b33ce599e38c8f22565cdaa5d91d63fa612a2..0000000000000000000000000000000000000000
--- a/lib/generate/__init__.py
+++ /dev/null
@@ -1,28 +0,0 @@
-
-import numpy as np
-from pkgutil import extend_path
-__path__ = extend_path(__path__, __name__)
-
-# General utility functions
-def db_gen_runtime_estimate(embedding_time,
- num_identities,
- samples_per_identity,
- average_new_identity_rejects,
- average_new_sample_rejects):
- """
- Estimate the total runtime of a database generation given the provided requirements and specifications.
- Inputs:
- `embedding_time` : time to generate embedding from latent vector
- `num_identities` : number of different identites to generate
- `samples_per_identity` : required number of samples per identity
- `average_new_identity_rejects` : expected number of rejects for new identities (when to close from preexisting one)
- `average_new_sample_rejects` : expected number of rejects for new samples (when to far from target identity)
-
- Returns:
- estimated total runtime in the same units as the input `embedding_time`
- """
-
- total = 0.0
- total += (average_new_identity_rejects + 1) * num_identities * embedding_time
- total += (average_new_sample_rejects + 1) * (samples_per_identity - 1) * num_identities * embedding_time
- return total
diff --git a/lib/generate/db_generator.py b/lib/generate/db_generator.py
deleted file mode 100644
index 45cabe83183683176f643bbbc39923c954962b7b..0000000000000000000000000000000000000000
--- a/lib/generate/db_generator.py
+++ /dev/null
@@ -1,432 +0,0 @@
-import os
-import h5py
-import numpy as np
-import time
-import pickle as pkl
-from bob.extension import rc
-from .face_wrapper import Face
-from ..latent import editing
-from scipy.spatial.distance import cdist
-
-COVARIATES = ["expression", "pose", "illumination"]
-MAX_FILES_PER_FOLDER = 1000
-
-
-class DBGenerator(object):
- def __init__(
- self,
- generator,
- image_postprocessor,
- cropper,
- extractor,
- ict,
- covariates_scaling,
- color_variations,
- image_dir,
- references_metadata_dir,
- augmentations_metadata_dir,
- seed,
- covariates_analysis_path=rc["bob.synthface.multipie_latent_directions"],
- ):
- """[summary]
-
- Parameters
- ----------
- generator : :synthface.generate.stylegan2_generator.StyleGAN2Generator:
- The core StyleGAN2 generator used for face image generation
- image_postprocessor : callable
- Function to apply on the generated images before saving & embedding
- computation. Typically : downscaling/cropping from 1024x1024 to lower resolution.
- cropper : bob.bio.face.Preprocessor
- Cropper stage of the embedding computation
- extractor : bob.bio.face.Extractor
- Extractor stage of the embedding computation
- ict : float
- Inter-class threshold requirement. During generation of the references for each identity in the
- database, the generator will ensure that each new identity has is at least at a distance `ict` from
- the previous identities, in the embedding space.
- covariates_scaling : dict
- Scaling factors for each type of semantic augmentation. Example:
- ```
- covariates_scaling = {
- "pose" : 3/4,
- "illumination": 5/4,
- "expression": 1.0
- }
- ```
- color_variations : int
- Number of color variations of each image to generate.
- If set to 0, then no color variation is applied (only the semantic augmentations are computed)-
- image_dir : str
- Root directory where to store the images.
- references_metadata_dir : str
- Path to the folder where is stored the metadata on the references (latents, embeddings, statistics)
- augmentations_metadata_dir : str
- Path to the folder where is stored the metadata on the augmented faces (latents, embeddings, statistics)
- seed : int
- Seed of the random generator for reproducibility
- covariates_analysis_path: str
- Path to pickled covariaties analysis results (that contains the latent directions)
- """
-
- self.generator = generator
- self.image_postprocessor = image_postprocessor
- self.cropper = cropper
- self.extractor = extractor
- self.seed = seed
- with open(rc["bob.synthface.multipie_latent_directions"], "rb") as f:
- self.covariates_analysis = pkl.load(f)
-
- self.ict = ict
- self.covariates_scaling = covariates_scaling
- self.color_variations = color_variations
- self.image_dir = image_dir
- self.references_metadata_dir = references_metadata_dir
- self.augmentations_metadata_dir = augmentations_metadata_dir
- self.references_hdf5_path = os.path.join(
- self.references_metadata_dir, "representations.h5"
- )
- self.augmentations_hdf5_path = os.path.join(
- self.augmentations_metadata_dir, "representations.h5"
- )
-
- def _get_subdir(self, identity):
- return "{:05}".format(identity // MAX_FILES_PER_FOLDER)
-
- def get_embedding(self, img):
- return self.extractor(self.cropper(img))
-
- def get_color_augmented_faces(self, identity, w_augmented, labels):
- """ Generate faces from w-latents using style-mixing to add some color variation.
- The identity specific latents are used for coarse resolution inputs (4x4 to 64x64) of the synthesis network,
- while another randomly generated latent is used for fine resolutions (128x128 to 1024x1024).
-
- Parameters
- ----------
- identity : int
- Tag for the identity that is currently processed.
- w_augmented : np.array of shape (num_latents, latent_dim)
- Base w-latents for the current identity (including the semantically augmented latents and optionally the reference latent)
- labels : iterable of str
- Sample tag for each latent in w_augmented
-
- Returns
- -------
- List of all Faces obtained after computing the synthetic image and associated face embeddings.
- """
-
- # Compute resulting image with style mixing
- # By choice, the style_mixing_depth and truncation_psi are hardcoded as carelessly chosen values
- # can lead the generated images to not all match to the same identity.
-
- images, all_base_latents, _ = self.generator.run_with_style_mixing_augmentation(
- w_augmented,
- num_variations=self.color_variations,
- style_mixing_depth=10,
- truncation_psi=0.85,
- return_w_latents=True,
- )
- images = [self.image_postprocessor(img) for img in images]
-
- # Update labels to differentiate variations
- labels = [
- ["{}_var_{}".format(label, variation) for label in labels]
- for variation in range(self.color_variations)
- ]
- labels = sum(labels, [])
-
- # Extract embeddings
- embeddings = [self.get_embedding(img) for img in images]
-
- # Encapsulate as faces
- faces = [
- Face(
- z_latent=None,
- w_latent=w,
- image=img,
- embedding=embedding,
- identity=identity,
- sample=label,
- )
- for w, img, embedding, label in zip(
- all_base_latents, images, embeddings, labels
- )
- ]
- return faces
-
- def get_faces(self, identity, w_augmented, labels):
- """
- Generate faces from w_latents, in particular including the computation of face embeddings.
-
- Parameters
- ----------
- identity : int
- Tag for the identity that is currently processed.
- w_augmented : np.array of shape (num_latents, latent_dim)
- Base w-latents for the current identity (including the semantically augmented latents and optionally the reference latent)
- labels : iterable of str
- Sample tag for each latent in w_augmented
-
- Returns
- -------
- List of all Faces obtained after computing the synthetic image and associated face embeddings.
- """
- # Compute images and embeddings
- images = self.generator.run_from_W(w_augmented)
- images = [self.image_postprocessor(img) for img in images]
- embeddings = [self.get_embedding(img) for img in images]
-
- # Encapsulate as faces
- faces = [
- Face(
- z_latent=None,
- w_latent=w,
- image=img,
- embedding=embedding,
- identity=identity,
- sample=label,
- )
- for w, img, embedding, label in zip(w_augmented, images, embeddings, labels)
- ]
- return faces
-
- def create_reference(self, identity, compared_embeddings):
- """ Create the reference Face for the required identity by randomly sampling a new latent vector,
- ans computing the face image as well as the face embedding.
- Embeddings of preexisting identities must be provided in order to apply
- an inter-class distance constraint.
-
- Parameters
- ----------
- identity : int
- Tag of the currently generated identity
- compared_embeddings : iterable of np.arrays
- Embeddings of all preexisting identities in the database (embeddings have dimensionality (128, )).
-
- Returns
- -------
- :synthface.generate.face_wrapper.Face:
-
- """
-
- # For reproducibility : random generator is reseeded for each identity
- # using a composite seed that depends on the generator seed, and on the
- # identity tag itself.
- np.random.seed([self.seed, identity, 0])
-
- accepted = False
- num_candidates = 0
- # Generate candidates until the ICT requirement is satistfied
- while not accepted:
- num_candidates += 1
- # Sample a new identity from the Z space
- ref_z = np.random.randn(1, self.generator.latent_dim)
- _, ref_w = self.generator.run_from_Z(ref_z, return_w_latents=True)
-
- # Neutralize it (make it frontal view, frontal lightning, neutral expression)
- ref_w = editing.latent_neutralisation(ref_w, self.covariates_analysis)
-
- # Compute the image and embeddings
- raw_images = self.generator.run_from_W(ref_w)
- ref_img = self.image_postprocessor(raw_images[0])
- ref_embedding = self.get_embedding(ref_img)
-
- # Compare the new embedding to previous embeddings, and accept the candidate only
- # if the ICT constraint is satisfied
- if len(compared_embeddings) > 0:
- distances = (
- 0.5
- * cdist([ref_embedding], compared_embeddings, metric="sqeuclidean")[
- 0
- ]
- )
- accepted = np.all(distances >= self.ict)
- else:
- accepted = True
-
- # Encapsulate as a Face
- ref_face = Face(
- z_latent=ref_z,
- w_latent=ref_w,
- image=ref_img,
- embedding=ref_embedding,
- identity=identity,
- sample="reference",
- )
- return ref_face, num_candidates
-
- def augment_identity(self, reference_face):
- """ Generate all semantic augmentations for the provided reference.
-
- Parameters
- ----------
- reference_face : :synthface.generate.face_wrapper.Face:
- Reference Face
-
- Returns
- -------
- list of :synthface.generate.face_wrapper.Faces:
- Faces for each semantic augmentations of the reference
- """
- # Compute semantic augmentations
- w_augmented, labels = editing.latent_augmentation(
- reference_face.w_latent, self.covariates_analysis, self.covariates_scaling
- )
-
- # Optionally, also add color variation
- if self.color_variations == 0:
- # Exclude the first latent (which is the latent from the reference), as it already is in the database
- faces = self.get_faces(reference_face.identity, w_augmented[1:], labels[1:])
- else:
- faces = self.get_color_augmented_faces(
- reference_face.identity, w_augmented, labels
- )
-
- return faces
-
- def save_faces(self, faces, h5_file):
- """ Save a list of faces
-
- Parameters
- ----------
- faces :
- List of Faces to save
- h5_file : str
- Path to the hdf5 file where the Face's representations (latents, embeddings) must be stored
- """
- with h5py.File(h5_file) as f:
- for face in faces:
- face.save(
- os.path.join(self.image_dir, self._get_subdir(face.identity)), f
- )
-
- def load_reference(self, identity):
- """ Load a precomputed reference Face from the filesystem
-
- Parameters
- ----------
- identity : int
- Tag for the identity of the reference to load
-
- Returns
- -------
- :synthface.generate.face_wrapper.Face:
- Face of the reference
-
- Raises
- ------
- RuntimeError
- When the reference cannot be found if the filesytem (either if the HDF5 file does not exists,
- or if it exists but does not contain this identity tag)
- """
- try:
- with h5py.File(self.references_hdf5_path, "r") as f:
- reference = Face.load(
- os.path.join(self.image_dir, self._get_subdir(identity)),
- f,
- identity,
- "reference",
- )
- except:
- raise RuntimeError(
- "Reference for identity {} is not yet generated and thus cannot be restored".format(
- identity
- )
- )
-
- return reference
-
- def list_references(self):
- """
- Run through the filesystem to list all identities whose reference Face
- has already been computed.
-
- Returns
- -------
- List of int
- List of tags of all identities whose reference already exists
- """
- if os.path.exists(self.references_hdf5_path):
- with h5py.File(self.references_hdf5_path, "r") as f:
- return [int(k) for k in f["embedding"].keys()]
- else:
- return []
-
- def create_database(self, identities):
- """ Create a full database (reference + augmentations) for the provided identity tags
-
- Parameters
- ----------
- identities : List of int
- List of unique tags for each identity that must be created
- """
- self.create_references(identities)
- self.augment_identities(identities)
-
- def create_references(self, identities):
- """ Create all references for the provided identity tags
-
- Parameters
- ----------
- identities : List of int
- List of unique tags for each identity that must be created
- """
- print("Creating references...")
- # List preexisting identities and load their embeddings
- previous_identities = self.list_references()
- compared_embeddings = [
- self.load_reference(identity).embedding for identity in previous_identities
- ]
-
- # Don't regenerate preexisting identities
- identities = set(identities).difference(set(previous_identities))
-
- num_identities = len(identities)
- for idx, identity in enumerate(identities):
- print("Identity {} ({}/{})".format(identity, idx + 1, num_identities))
-
- # Create reference
- start_time = time.time()
- ref_face, num_candidates = self.create_reference(
- identity, compared_embeddings
- )
- runtime = time.time() - start_time
-
- # Store stats
- with open(
- os.path.join(self.references_metadata_dir, "stats.dat"), "a"
- ) as f:
- f.write("{} {:.2f} {}\n".format(identity, runtime, num_candidates))
-
- # Store face
- self.save_faces([ref_face], self.references_hdf5_path)
- compared_embeddings.append(ref_face.embedding)
-
- def augment_identities(self, identities):
- """ Create all augmentations for the provided identity tags.
- This method assumes the references for those identities have already been
- created.
-
- Parameters
- ----------
- identities : List of str or int
- List of unique tags for each identity that must be created
- """
- print("Augmenting identities...")
- num_identities = len(identities)
- for idx, identity in enumerate(identities):
- print("Identity {} ({}/{})".format(identity, idx + 1, num_identities))
- ref_face = self.load_reference(identity)
-
- # Augment identity
- start_time = time.time()
- faces = self.augment_identity(ref_face)
- runtime = time.time() - start_time
-
- # Store stats
- with open(
- os.path.join(self.augmentations_metadata_dir, "stats.dat"), "a"
- ) as f:
- f.write("{} {:.2f}\n".format(identity, runtime))
-
- self.save_faces(faces, self.augmentations_hdf5_path)
diff --git a/lib/generate/face_wrapper.py b/lib/generate/face_wrapper.py
deleted file mode 100644
index 6d34f1b2f7177a9fd05481fa82022882d6921b87..0000000000000000000000000000000000000000
--- a/lib/generate/face_wrapper.py
+++ /dev/null
@@ -1,110 +0,0 @@
-import os
-import numpy as np
-from scipy.spatial.distance import cosine as cosine_dist
-import bob.io.image
-import bob.io.base
-
-class Face(object):
- def __init__(self,
- z_latent,
- w_latent,
- image,
- embedding,
- identity,
- sample,
- preprocessed=None):
- self.z_latent = z_latent
- self.w_latent = w_latent
- self.image = image
- self.preprocessed = preprocessed
- self.embedding = embedding
-
- self.identity = identity
- self.sample = sample
-
- def show(self):
- bob.io.image.imshow(self.image)
-
- def show_preprocessed(self):
- bob.io.image.imshow(self.preprocessed)
-
- @staticmethod
- def distance(face1, face2, modules=None):
- """
- Compute the identity-distance between two latent vectors from StyleGAN2's
- latent space.
- """
- unwrap = lambda x : x.embedding if isinstance(x, Face) else x
- return 0.5*np.linalg.norm(unwrap(face1) - unwrap(face2))**2
-
- @staticmethod
- def get_all_distances(face_pairs):
- """
- From a list of size n of 2-tuples of Faces with embeddings, compute distances between all pairs
- and returns a list of length n containing those distances
- """
- return [Face.distance(pair[0], pair[1]) for pair in face_pairs]
-
- @staticmethod
- def _build_unique_id(identity, sample):
- return str(identity) + '_' + str(sample)
-
- def get_unique_id(self):
- return Face._build_unique_id(self.identity, self.sample)
-
- @staticmethod
- def get_path(identity, sample, content, extension='', output_dir=None):
- if output_dir is not None:
- return os.path.join(output_dir, content, str(identity), Face._build_unique_id(identity, sample) + extension)
- else:
- return os.path.join(content, str(identity), Face._build_unique_id(identity, sample) + extension)
-
- def save(self, root_dir,
- h5_file,
- save_preprocessed=False):
-
- if self.z_latent is not None:
- h5_file.create_dataset(Face.get_path(self.identity, self.sample, 'z_latent'), data=self.z_latent)
- h5_file.create_dataset(Face.get_path(self.identity, self.sample, 'w_latent'), data=self.w_latent)
-
- bob.io.base.save(self.image, Face.get_path(self.identity, self.sample, '', '.png', root_dir), create_directories=True)
-
- if save_preprocessed and self.preprocessed is not None:
- h5_file.create_dataset(Face.get_path(self.identity, self.sample, 'preprocessed'), data=self.preprocessed)
- h5_file.create_dataset(Face.get_path(self.identity, self.sample, 'embedding'), data=self.embedding)
- h5_file.flush()
-
- @classmethod
- def load(cls, root_dir, h5_file, identity, sample):
- z_latent_path = Face.get_path(identity, sample, 'z_latent')
- if z_latent_path in h5_file:
- z_latent = h5_file[z_latent_path][:]
- else:
- z_latent = None
-
- w_latent = h5_file[Face.get_path(identity, sample, 'w_latent')][:]
- image = bob.io.base.load(Face.get_path(identity, sample, '', '.png', root_dir))
-
- preprocessed_path = Face.get_path(identity, sample, 'preprocessed')
- if preprocessed_path in h5_file:
- preprocessed = h5_file[preprocessed_path][:]
- else:
- preprocessed = None
- embedding = h5_file[Face.get_path(identity, sample, 'embedding')][:]
-
- face = cls(z_latent=z_latent,
- w_latent=w_latent,
- image=image,
- embedding=embedding,
- identity=identity,
- sample=sample,
- preprocessed=preprocessed)
-
- return face
-
-
-
-
-
-
-
\ No newline at end of file
diff --git a/lib/latent/analysis.py b/lib/latent/analysis.py
deleted file mode 100644
index e755e89bec4117c8cd185c0108b14b03f10506d4..0000000000000000000000000000000000000000
--- a/lib/latent/analysis.py
+++ /dev/null
@@ -1,166 +0,0 @@
-from bob.extension import rc
-from bob.db.multipie import Database as Multipie
-import os
-import pandas as pd
-from sklearn.svm import LinearSVC
-from sklearn.preprocessing import LabelEncoder
-import numpy as np
-
-MAIN_DIR = os.path.join(rc["databases"], "multipie_proj")
-LAT_DIR = os.path.join(MAIN_DIR, "w_latents")
-FAILURE_FILE = os.path.join(MAIN_DIR, "failure.dat")
-MULTIPIE_DIR = rc["bob.db.multipie.directory"]
-
-PROTOCOLS = {
- "expression": "E_lit",
- "pose": "P_lit",
- "illumination": "U"
-}
-
-VALID_CAMERAS = ["08_0", "13_0", "14_0", "05_1", "05_0", "04_1", "19_0"]
-
-POS_TO_CAM = {
- "frontal": ["05_1"],
- "left": ["11_0", "12_0", "09_0", "08_0", "13_0", "14_0"],
- "right": ["05_0", "04_1", "19_0", "20_0", "01_0", "24_0"],
-}
-CAM_TO_POS = {cam: pos for pos, cam_list in POS_TO_CAM.items() for cam in cam_list}
-
-POS_TO_FLASH = {
- "no_flash": [0],
- "left": list(range(1, 7)) + [14, 15],
- "frontal": [7, 16],
- "right": list(range(8, 14)) + [17, 18],
-}
-
-FLASH_TO_POS = {
- flash: pos for pos, flash_list in POS_TO_FLASH.items() for flash in flash_list
-}
-
-VPOS_TO_FLASH = {
- "no_flash": [0],
- "middle": list(range(1, 14)),
- "top": list(range(14, 19)),
-}
-
-FLASH_TO_VPOS = {
- flash: vpos for vpos, flash_list in VPOS_TO_FLASH.items() for flash in flash_list
-}
-
-
-def filter_failure_cases(file_list):
- with open(FAILURE_FILE, "r") as f:
- failures = [item.rstrip() for item in f.readlines()]
-
- return [f for f in file_list if f.path not in failures]
-
-
-def get_covariate(file, covariate):
- if covariate == "expression":
- return file.expression.name
- elif covariate == "pose":
- return CAM_TO_POS[file.file_multiview.camera.name]
- elif covariate == "illumination":
- return FLASH_TO_POS[file.file_multiview.shot_id]
- else:
- raise ValueError(
- "Unknown covariate {} not in {}".format(covariate, PROTOCOLS.keys())
- )
-
-
-def load_latents(covariate, group="world"):
- assert covariate in PROTOCOLS.keys(), "Unknown `covariate` {} not in {}".format(
- covariate, PROTOCOLS.keys()
- )
- db = Multipie(original_directory=MULTIPIE_DIR)
- files = filter_failure_cases(
- db.objects(protocol=PROTOCOLS[covariate], groups=[group], cameras=VALID_CAMERAS)
- )
- df = pd.DataFrame()
- df["file"] = [f.path for f in files]
- df["latent"] = [f.load(LAT_DIR, ".h5") for f in files]
-
- df[covariate] = [get_covariate(f, covariate) for f in files]
-
- return df
-
-
-def binary_analysis(train_df, covariate, target_labels, **kwargs):
- train_df = train_df[train_df[covariate].isin(target_labels)]
-
- svm = LinearSVC(fit_intercept=False, **kwargs)
- train_latents = np.stack(train_df["latent"])
- train_labels = np.stack(train_df[covariate])
-
- # Ensure that 0 and 1 classes match the order of the target_labels
- train_labels = np.where(train_labels == target_labels[0], 0, 1)
-
- svm.fit(train_latents, train_labels)
-
- normal = svm.coef_ / np.linalg.norm(svm.coef_)
-
- distances = train_latents.dot(normal.T)
- negatives = distances[distances < 0]
- positives = distances[distances > 0]
-
- def stats(distances):
- return {
- "min": np.min(distances),
- "max": np.max(distances),
- "mean": np.mean(distances),
- "std": np.std(distances),
- }
-
- neg_stats = stats(negatives)
- pos_stats = stats(positives)
-
- return {
- "svm": svm,
- "normal": normal,
- "neg_stats": neg_stats,
- "pos_stats": pos_stats,
- }
-
-
-def multiclass_analysis(train_df, covariate, neutral_label, other_labels, **kwargs):
- return {
- (neutral_label, label): binary_analysis(
- train_df,
- covariate=covariate,
- target_labels=[neutral_label, label],
- **kwargs
- )
- for label in other_labels
- }
-
-
-def pose_analysis(train_df):
- return binary_analysis(train_df, covariate="pose", target_labels=["left", "right"])
-
-
-def illumination_analysis(train_df):
- return binary_analysis(
- train_df, covariate="illumination", target_labels=["left", "right"]
- )
-
-
-def expression_analysis(train_df):
- return multiclass_analysis(
- train_df,
- covariate="expression",
- neutral_label="neutral",
- other_labels=["smile", "scream", "disgust", "squint", "surprise"],
- )
-
-
-def covariate_analysis(covariate, train_df):
- if covariate == "expression":
- return expression_analysis(train_df)
- elif covariate == "pose":
- return pose_analysis(train_df)
- elif covariate == "illumination":
- return illumination_analysis(train_df)
- else:
- raise ValueError(
- "Unknown covariate {} not in {}".format(covariate, PROTOCOLS.keys())
- )
diff --git a/lib/latent/editing.py b/lib/latent/editing.py
deleted file mode 100644
index 9edb81439dc6fd3a55c5df4dc833e27238a7b9c4..0000000000000000000000000000000000000000
--- a/lib/latent/editing.py
+++ /dev/null
@@ -1,95 +0,0 @@
-import numpy as np
-
-LATENT_DIM = 512
-
-
-def latent_neutralisation(w, covariates_analysis):
- # Pose neutralisation : project on binary decision boundary
- pose_normal = covariates_analysis["pose"]["normal"]
- w -= w.dot(pose_normal.T) * pose_normal
-
- # Illumination neutralisation : project on binary decision boundary
- illumination_normal = covariates_analysis["illumination"]["normal"]
- w -= w.dot(illumination_normal.T) * illumination_normal
-
- # Expression neutralisation
- ### Assumption : default sampled latents are always either "neutral" or "smile"
- expression_normal = covariates_analysis["expression"][("neutral", "smile")][
- "normal"
- ]
- # 1. Cancel component
- w -= w.dot(expression_normal.T) * expression_normal
- # 2. Move towards neutral using the mean distance computed on train set
- w += (
- covariates_analysis["expression"][("neutral", "smile")]["neg_stats"]["mean"]
- * expression_normal
- )
-
- return w
-
-
-def binary_augmentation(w, analysis, num_latents, scaling):
- neg_mean = analysis["neg_stats"]["mean"]
- pos_mean = analysis["pos_stats"]["mean"]
-
- extremum = max(-neg_mean, pos_mean)
-
- normal = analysis["normal"]
- scales = scaling * np.linspace(-extremum, extremum, num_latents)[:, None]
- return w + scales * normal, scales
-
-
-def pose_augmentation(w, pose_analysis, num_latents, scaling=3 / 4):
- return binary_augmentation(w, pose_analysis, num_latents, scaling)
-
-
-def illumination_augmentation(w, illumination_analysis, num_latents, scaling=5 / 4):
- return binary_augmentation(w, illumination_analysis, num_latents, scaling)
-
-
-def expression_augmentation(w, expression_analysis, scaling=3 / 4):
- new_covariates = []
- new_latents = []
- for direction, analysis in expression_analysis.items():
- new_covariates.append(direction[1])
- normal = analysis["normal"]
- # 1. Cancel neutral component
- w_augmented = w - w.dot(normal.T) * normal
- # 2. Move towards expression using the mean distance computed on train set to ensure realistic outcome
- w_augmented += scaling * analysis["pos_stats"]["mean"] * normal
- new_latents.append(w_augmented)
-
- return np.concatenate(new_latents), np.stack(new_covariates)
-
-
-def latent_augmentation(w, covariates_analysis, covariates_scaling):
- # Pose augmentation
- w_pose, pose_labels = pose_augmentation(
- w,
- covariates_analysis["pose"],
- num_latents=6,
- scaling=covariates_scaling["pose"],
- )
- pose_labels = ["pose_{}".format(item) for item in range(len(pose_labels))]
-
- # Illumination augmentation
- w_illumination, illumination_labels = illumination_augmentation(
- w,
- covariates_analysis["illumination"],
- num_latents=6,
- scaling=covariates_scaling["illumination"],
- )
- illumination_labels = [
- "illumination_{}".format(item) for item in range(len(illumination_labels))
- ]
-
- # Expression augmentation
- w_expression, expression_labels = expression_augmentation(
- w, covariates_analysis["expression"], scaling=covariates_scaling["expression"]
- )
- expression_labels = expression_labels.flatten().tolist()
-
- latents = np.concatenate([w, w_pose, w_illumination, w_expression])
- labels = ["original"] + pose_labels + illumination_labels + expression_labels
-
- return latents, labels
diff --git a/lib/stylegan2/LICENSE.txt b/lib/stylegan2/LICENSE.txt
deleted file mode 100755
index d7e85075defbd96f9e7a9d756aa7db0e7e30ccf7..0000000000000000000000000000000000000000
--- a/lib/stylegan2/LICENSE.txt
+++ /dev/null
@@ -1,101 +0,0 @@
-Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-
-
-Nvidia Source Code License-NC
-
-=======================================================================
-
-1. Definitions
-
-"Licensor" means any person or entity that distributes its Work.
-
-"Software" means the original work of authorship made available under
-this License.
-
-"Work" means the Software and any additions to or derivative works of
-the Software that are made available under this License.
-
-"Nvidia Processors" means any central processing unit (CPU), graphics
-processing unit (GPU), field-programmable gate array (FPGA),
-application-specific integrated circuit (ASIC) or any combination
-thereof designed, made, sold, or provided by Nvidia or its affiliates.
-
-The terms "reproduce," "reproduction," "derivative works," and
-"distribution" have the meaning as provided under U.S. copyright law;
-provided, however, that for the purposes of this License, derivative
-works shall not include works that remain separable from, or merely
-link (or bind by name) to the interfaces of, the Work.
-
-Works, including the Software, are "made available" under this License
-by including in or with the Work either (a) a copyright notice
-referencing the applicability of this License to the Work, or (b) a
-copy of this License.
-
-2. License Grants
-
- 2.1 Copyright Grant. Subject to the terms and conditions of this
- License, each Licensor grants to you a perpetual, worldwide,
- non-exclusive, royalty-free, copyright license to reproduce,
- prepare derivative works of, publicly display, publicly perform,
- sublicense and distribute its Work and any resulting derivative
- works in any form.
-
-3. Limitations
-
- 3.1 Redistribution. You may reproduce or distribute the Work only
- if (a) you do so under this License, (b) you include a complete
- copy of this License with your distribution, and (c) you retain
- without modification any copyright, patent, trademark, or
- attribution notices that are present in the Work.
-
- 3.2 Derivative Works. You may specify that additional or different
- terms apply to the use, reproduction, and distribution of your
- derivative works of the Work ("Your Terms") only if (a) Your Terms
- provide that the use limitation in Section 3.3 applies to your
- derivative works, and (b) you identify the specific derivative
- works that are subject to Your Terms. Notwithstanding Your Terms,
- this License (including the redistribution requirements in Section
- 3.1) will continue to apply to the Work itself.
-
- 3.3 Use Limitation. The Work and any derivative works thereof only
- may be used or intended for use non-commercially. The Work or
- derivative works thereof may be used or intended for use by Nvidia
- or its affiliates commercially or non-commercially. As used herein,
- "non-commercially" means for research or evaluation purposes only.
-
- 3.4 Patent Claims. If you bring or threaten to bring a patent claim
- against any Licensor (including any claim, cross-claim or
- counterclaim in a lawsuit) to enforce any patents that you allege
- are infringed by any Work, then your rights under this License from
- such Licensor (including the grants in Sections 2.1 and 2.2) will
- terminate immediately.
-
- 3.5 Trademarks. This License does not grant any rights to use any
- Licensor's or its affiliates' names, logos, or trademarks, except
- as necessary to reproduce the notices described in this License.
-
- 3.6 Termination. If you violate any term of this License, then your
- rights under this License (including the grants in Sections 2.1 and
- 2.2) will terminate immediately.
-
-4. Disclaimer of Warranty.
-
-THE WORK IS PROVIDED "AS IS" WITHOUT WARRANTIES OR CONDITIONS OF ANY
-KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WARRANTIES OR CONDITIONS OF
-MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE OR
-NON-INFRINGEMENT. YOU BEAR THE RISK OF UNDERTAKING ANY ACTIVITIES UNDER
-THIS LICENSE.
-
-5. Limitation of Liability.
-
-EXCEPT AS PROHIBITED BY APPLICABLE LAW, IN NO EVENT AND UNDER NO LEGAL
-THEORY, WHETHER IN TORT (INCLUDING NEGLIGENCE), CONTRACT, OR OTHERWISE
-SHALL ANY LICENSOR BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY DIRECT,
-INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT OF
-OR RELATED TO THIS LICENSE, THE USE OR INABILITY TO USE THE WORK
-(INCLUDING BUT NOT LIMITED TO LOSS OF GOODWILL, BUSINESS INTERRUPTION,
-LOST PROFITS OR DATA, COMPUTER FAILURE OR MALFUNCTION, OR ANY OTHER
-COMMERCIAL DAMAGES OR LOSSES), EVEN IF THE LICENSOR HAS BEEN ADVISED OF
-THE POSSIBILITY OF SUCH DAMAGES.
-
-=======================================================================
diff --git a/lib/stylegan2/README.md b/lib/stylegan2/README.md
deleted file mode 100755
index 7f8d46bb85b8476d494e47cd658427a753d92be4..0000000000000000000000000000000000000000
--- a/lib/stylegan2/README.md
+++ /dev/null
@@ -1,220 +0,0 @@
-## StyleGAN2 — Official TensorFlow Implementation
-
-
-
-**Analyzing and Improving the Image Quality of StyleGAN**<br>
-Tero Karras, Samuli Laine, Miika Aittala, Janne Hellsten, Jaakko Lehtinen, Timo Aila<br>
-
-Paper: http://arxiv.org/abs/1912.04958<br>
-Video: https://youtu.be/c-NJtV9Jvp0<br>
-
-Abstract: *The style-based GAN architecture (StyleGAN) yields state-of-the-art results in data-driven unconditional generative image modeling. We expose and analyze several of its characteristic artifacts, and propose changes in both model architecture and training methods to address them. In particular, we redesign generator normalization, revisit progressive growing, and regularize the generator to encourage good conditioning in the mapping from latent vectors to images. In addition to improving image quality, this path length regularizer yields the additional benefit that the generator becomes significantly easier to invert. This makes it possible to reliably detect if an image is generated by a particular network. We furthermore visualize how well the generator utilizes its output resolution, and identify a capacity problem, motivating us to train larger models for additional quality improvements. Overall, our improved model redefines the state of the art in unconditional image modeling, both in terms of existing distribution quality metrics as well as perceived image quality.*
-
-For business inquiries, please contact [researchinquiries@nvidia.com](mailto:researchinquiries@nvidia.com)<br>
-For press and other inquiries, please contact Hector Marinez at [hmarinez@nvidia.com](mailto:hmarinez@nvidia.com)<br>
-
-| Additional material |
-| :--- | :----------
-| [StyleGAN2](https://drive.google.com/open?id=1QHc-yF5C3DChRwSdZKcx1w6K8JvSxQi7) | Main Google Drive folder
-| ├ [stylegan2-paper.pdf](https://drive.google.com/open?id=1fnF-QsiQeKaxF-HbvFiGtzHF_Bf3CzJu) | High-quality version of the paper
-| ├ [stylegan2-video.mp4](https://drive.google.com/open?id=1f_gbKW6FUUHKkUxciJ_lQx29mCq_fSBy) | High-quality version of the video
-| ├ [images](https://drive.google.com/open?id=1Sak157_DLX84ytqHHqZaH_59HoEWzfB7) | Example images produced using our method
-| │ ├ [curated-images](https://drive.google.com/open?id=1ydWb8xCHzDKMTW9kQ7sL-B1R0zATHVHp) | Hand-picked images showcasing our results
-| │ └ [100k-generated-images](https://drive.google.com/open?id=1BA2OZ1GshdfFZGYZPob5QWOGBuJCdu5q) | Random images with and without truncation
-| ├ [videos](https://drive.google.com/open?id=1yXDV96SFXoUiZKU7AyE6DyKgDpIk4wUZ) | Individual clips of the video as high-quality MP4
-| └ [networks](https://drive.google.com/open?id=1yanUI9m4b4PWzR0eurKNq6JR1Bbfbh6L) | Pre-trained networks
-|    ├ [stylegan2-ffhq-config-f.pkl](http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-ffhq-config-f.pkl) | StyleGAN2 for <span style="font-variant:small-caps">FFHQ</span> dataset at 1024×1024
-|    ├ [stylegan2-car-config-f.pkl](http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-car-config-f.pkl) | StyleGAN2 for <span style="font-variant:small-caps">LSUN Car</span> dataset at 512×384
-|    ├ [stylegan2-cat-config-f.pkl](http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-cat-config-f.pkl) | StyleGAN2 for <span style="font-variant:small-caps">LSUN Cat</span> dataset at 256×256
-|    ├ [stylegan2-church-config-f.pkl](http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-church-config-f.pkl) | StyleGAN2 for <span style="font-variant:small-caps">LSUN Church</span> dataset at 256×256
-|    ├ [stylegan2-horse-config-f.pkl](http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-horse-config-f.pkl) | StyleGAN2 for <span style="font-variant:small-caps">LSUN Horse</span> dataset at 256×256
-|    └ ⋯ | Other training configurations used in the paper
-
-## Requirements
-
-* Both Linux and Windows are supported. Linux is recommended for performance and compatibility reasons.
-* 64-bit Python 3.6 installation. We recommend Anaconda3 with numpy 1.14.3 or newer.
-* TensorFlow 1.14 or 1.15 with GPU support. The code does not support TensorFlow 2.0.
-* On Windows, you need to use TensorFlow 1.14 — TensorFlow 1.15 will not work.
-* One or more high-end NVIDIA GPUs, NVIDIA drivers, CUDA 10.0 toolkit and cuDNN 7.5. To reproduce the results reported in the paper, you need an NVIDIA GPU with at least 16 GB of DRAM.
-* Docker users: use the [provided Dockerfile](./Dockerfile) to build an image with the required library dependencies.
-
-StyleGAN2 relies on custom TensorFlow ops that are compiled on the fly using [NVCC](https://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html). To test that your NVCC installation is working correctly, run:
-
-```.bash
-nvcc test_nvcc.cu -o test_nvcc -run
-| CPU says hello.
-| GPU says hello.
-```
-
-On Windows, the compilation requires Microsoft Visual Studio to be in `PATH`. We recommend installing [Visual Studio Community Edition](https://visualstudio.microsoft.com/vs/) and adding into `PATH` using `"C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Auxiliary\Build\vcvars64.bat"`.
-
-## Using pre-trained networks
-
-Pre-trained networks are stored as `*.pkl` files on the [StyleGAN2 Google Drive folder](https://drive.google.com/open?id=1QHc-yF5C3DChRwSdZKcx1w6K8JvSxQi7). Below, you can either reference them directly using the syntax `gdrive:networks/<filename>.pkl`, or download them manually and reference by filename.
-
-```.bash
-# Generate uncurated ffhq images (matches paper Figure 12)
-python run_generator.py generate-images --network=gdrive:networks/stylegan2-ffhq-config-f.pkl \
- --seeds=6600-6625 --truncation-psi=0.5
-
-# Generate curated ffhq images (matches paper Figure 11)
-python run_generator.py generate-images --network=gdrive:networks/stylegan2-ffhq-config-f.pkl \
- --seeds=66,230,389,1518 --truncation-psi=1.0
-
-# Generate uncurated car images
-python run_generator.py generate-images --network=gdrive:networks/stylegan2-car-config-f.pkl \
- --seeds=6000-6025 --truncation-psi=0.5
-
-# Example of style mixing (matches the corresponding video clip)
-python run_generator.py style-mixing-example --network=gdrive:networks/stylegan2-ffhq-config-f.pkl \
- --row-seeds=85,100,75,458,1500 --col-seeds=55,821,1789,293 --truncation-psi=1.0
-```
-
-The results are placed in `results/<RUNNING_ID>/*.png`. You can change the location with `--result-dir`. For example, `--result-dir=~/my-stylegan2-results`.
-
-You can import the networks in your own Python code using `pickle.load()`. For this to work, you need to include the `dnnlib` source directory in `PYTHONPATH` and create a default TensorFlow session by calling `dnnlib.tflib.init_tf()`. See [run_generator.py](./run_generator.py) and [pretrained_networks.py](./pretrained_networks.py) for examples.
-
-## Preparing datasets
-
-Datasets are stored as multi-resolution TFRecords, similar to the [original StyleGAN](https://github.com/NVlabs/stylegan). Each dataset consists of multiple `*.tfrecords` files stored under a common directory, e.g., `~/datasets/ffhq/ffhq-r*.tfrecords`. In the following sections, the datasets are referenced using a combination of `--dataset` and `--data-dir` arguments, e.g., `--dataset=ffhq --data-dir=~/datasets`.
-
-**FFHQ**. To download the [Flickr-Faces-HQ](https://github.com/NVlabs/ffhq-dataset) dataset as multi-resolution TFRecords, run:
-
-```.bash
-pushd ~
-git clone https://github.com/NVlabs/ffhq-dataset.git
-cd ffhq-dataset
-python download_ffhq.py --tfrecords
-popd
-python dataset_tool.py display ~/ffhq-dataset/tfrecords/ffhq
-```
-
-**LSUN**. Download the desired LSUN categories in LMDB format from the [LSUN project page](https://www.yf.io/p/lsun). To convert the data to multi-resolution TFRecords, run:
-
-```.bash
-python dataset_tool.py create_lsun_wide ~/datasets/car ~/lsun/car_lmdb --width=512 --height=384
-python dataset_tool.py create_lsun ~/datasets/cat ~/lsun/cat_lmdb --resolution=256
-python dataset_tool.py create_lsun ~/datasets/church ~/lsun/church_outdoor_train_lmdb --resolution=256
-python dataset_tool.py create_lsun ~/datasets/horse ~/lsun/horse_lmdb --resolution=256
-```
-
-**Custom**. Create custom datasets by placing all training images under a single directory. The images must be square-shaped and they must all have the same power-of-two dimensions. To convert the images to multi-resolution TFRecords, run:
-
-```.bash
-python dataset_tool.py create_from_images ~/datasets/my-custom-dataset ~/my-custom-images
-python dataset_tool.py display ~/datasets/my-custom-dataset
-```
-
-## Projecting images to latent space
-
-To find the matching latent vectors for a set of images, run:
-
-```.bash
-# Project generated images
-python run_projector.py project-generated-images --network=gdrive:networks/stylegan2-car-config-f.pkl \
- --seeds=0,1,5
-
-# Project real images
-python run_projector.py project-real-images --network=gdrive:networks/stylegan2-car-config-f.pkl \
- --dataset=car --data-dir=~/datasets
-```
-
-## Training networks
-
-To reproduce the training runs for config F in Tables 1 and 3, run:
-
-```.bash
-python run_training.py --num-gpus=8 --data-dir=~/datasets --config=config-f \
- --dataset=ffhq --mirror-augment=true
-python run_training.py --num-gpus=8 --data-dir=~/datasets --config=config-f \
- --dataset=car --total-kimg=57000
-python run_training.py --num-gpus=8 --data-dir=~/datasets --config=config-f \
- --dataset=cat --total-kimg=88000
-python run_training.py --num-gpus=8 --data-dir=~/datasets --config=config-f \
- --dataset=church --total-kimg 88000 --gamma=100
-python run_training.py --num-gpus=8 --data-dir=~/datasets --config=config-f \
- --dataset=horse --total-kimg 100000 --gamma=100
-```
-
-For other configurations, see `python run_training.py --help`.
-
-We have verified that the results match the paper when training with 1, 2, 4, or 8 GPUs. Note that training FFHQ at 1024×1024 resolution requires GPU(s) with at least 16 GB of memory. The following table lists typical training times using NVIDIA DGX-1 with 8 Tesla V100 GPUs:
-
-| Configuration | Resolution | Total kimg | 1 GPU | 2 GPUs | 4 GPUs | 8 GPUs | GPU mem |
-| :------------ | :-------------: | :--------: | :-----: | :-----: | :-----: | :----: | :-----: |
-| `config-f` | 1024×1024 | 25000 | 69d 23h | 36d 4h | 18d 14h | 9d 18h | 13.3 GB |
-| `config-f` | 1024×1024 | 10000 | 27d 23h | 14d 11h | 7d 10h | 3d 22h | 13.3 GB |
-| `config-e` | 1024×1024 | 25000 | 35d 11h | 18d 15h | 9d 15h | 5d 6h | 8.6 GB |
-| `config-e` | 1024×1024 | 10000 | 14d 4h | 7d 11h | 3d 20h | 2d 3h | 8.6 GB |
-| `config-f` | 256×256 | 25000 | 32d 13h | 16d 23h | 8d 21h | 4d 18h | 6.4 GB |
-| `config-f` | 256×256 | 10000 | 13d 0h | 6d 19h | 3d 13h | 1d 22h | 6.4 GB |
-
-Training curves for FFHQ config F (StyleGAN2) compared to original StyleGAN using 8 GPUs:
-
-
-
-After training, the resulting networks can be used the same way as the official pre-trained networks:
-
-```.bash
-# Generate 1000 random images without truncation
-python run_generator.py generate-images --seeds=0-999 --truncation-psi=1.0 \
- --network=results/00006-stylegan2-ffhq-8gpu-config-f/networks-final.pkl
-```
-
-## Evaluation metrics
-
-To reproduce the numbers for config F in Tables 1 and 3, run:
-
-```.bash
-python run_metrics.py --data-dir=~/datasets --network=gdrive:networks/stylegan2-ffhq-config-f.pkl \
- --metrics=fid50k,ppl_wend --dataset=ffhq --mirror-augment=true
-python run_metrics.py --data-dir=~/datasets --network=gdrive:networks/stylegan2-car-config-f.pkl \
- --metrics=fid50k,ppl2_wend --dataset=car
-python run_metrics.py --data-dir=~/datasets --network=gdrive:networks/stylegan2-cat-config-f.pkl \
- --metrics=fid50k,ppl2_wend --dataset=cat
-python run_metrics.py --data-dir=~/datasets --network=gdrive:networks/stylegan2-church-config-f.pkl \
- --metrics=fid50k,ppl2_wend --dataset=church
-python run_metrics.py --data-dir=~/datasets --network=gdrive:networks/stylegan2-horse-config-f.pkl \
- --metrics=fid50k,ppl2_wend --dataset=horse
-```
-
-For other configurations, see the [StyleGAN2 Google Drive folder](https://drive.google.com/open?id=1QHc-yF5C3DChRwSdZKcx1w6K8JvSxQi7).
-
-Note that the metrics are evaluated using a different random seed each time, so the results will vary between runs. In the paper, we reported the average result of running each metric 10 times. The following table lists the available metrics along with their expected runtimes and random variation:
-
-| Metric | FFHQ config F | 1 GPU | 2 GPUs | 4 GPUs | Description |
-| :---------- | :------------: | :----: | :-----: | :----: | :---------- |
-| `fid50k` | 2.84 ± 0.03 | 22 min | 14 min | 10 min | [Fréchet Inception Distance](https://arxiv.org/abs/1706.08500)
-| `is50k` | 5.13 ± 0.02 | 23 min | 14 min | 8 min | [Inception Score](https://arxiv.org/abs/1606.03498)
-| `ppl_zfull` | 348.0 ± 3.8 | 41 min | 22 min | 14 min | [Perceptual Path Length](https://arxiv.org/abs/1812.04948) in Z, full paths
-| `ppl_wfull` | 126.9 ± 0.2 | 42 min | 22 min | 13 min | [Perceptual Path Length](https://arxiv.org/abs/1812.04948) in W, full paths
-| `ppl_zend` | 348.6 ± 3.0 | 41 min | 22 min | 14 min | [Perceptual Path Length](https://arxiv.org/abs/1812.04948) in Z, path endpoints
-| `ppl_wend` | 129.4 ± 0.8 | 40 min | 23 min | 13 min | [Perceptual Path Length](https://arxiv.org/abs/1812.04948) in W, path endpoints
-| `ppl2_wend` | 145.0 ± 0.5 | 41 min | 23 min | 14 min | [Perceptual Path Length](https://arxiv.org/abs/1812.04948) without center crop
-| `ls` | 154.2 / 4.27 | 10 hrs | 6 hrs | 4 hrs | [Linear Separability](https://arxiv.org/abs/1812.04948)
-| `pr50k3` | 0.689 / 0.492 | 26 min | 17 min | 12 min | [Precision and Recall](https://arxiv.org/abs/1904.06991)
-
-Note that some of the metrics cache dataset-specific data on the disk, and they will take somewhat longer when run for the first time.
-
-## License
-
-Copyright © 2019, NVIDIA Corporation. All rights reserved.
-
-This work is made available under the Nvidia Source Code License-NC. To view a copy of this license, visit https://nvlabs.github.io/stylegan2/license.html
-
-## Citation
-
-```
-@article{Karras2019stylegan2,
- title = {Analyzing and Improving the Image Quality of {StyleGAN}},
- author = {Tero Karras and Samuli Laine and Miika Aittala and Janne Hellsten and Jaakko Lehtinen and Timo Aila},
- journal = {CoRR},
- volume = {abs/1912.04958},
- year = {2019},
-}
-```
-
-## Acknowledgements
-
-We thank Ming-Yu Liu for an early review, Timo Viitanen for his help with code release, and Tero Kuosmanen for compute infrastructure.
diff --git a/lib/stylegan2/__init__.py b/lib/stylegan2/__init__.py
deleted file mode 100644
index 0bfb5a62b4f6c2c8cb1a6a865bc2d4b36ebfa64e..0000000000000000000000000000000000000000
--- a/lib/stylegan2/__init__.py
+++ /dev/null
@@ -1,2 +0,0 @@
-from pkgutil import extend_path
-__path__ = extend_path(__path__, __name__)
\ No newline at end of file
diff --git a/lib/stylegan2/dnnlib/__init__.py b/lib/stylegan2/dnnlib/__init__.py
deleted file mode 100644
index 31efcc4040d5b9e6857e519c471eb1971fdc78bb..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/__init__.py
+++ /dev/null
@@ -1,23 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-
-from . import submission
-
-from .submission.run_context import RunContext
-
-from .submission.submit import SubmitTarget
-from .submission.submit import PathType
-from .submission.submit import SubmitConfig
-from .submission.submit import submit_run
-from .submission.submit import get_path_from_template
-from .submission.submit import convert_path
-from .submission.submit import make_run_dir_path
-
-from .util import EasyDict
-
-
-submit_config: SubmitConfig = None # Package level variable for SubmitConfig which is only valid when inside the run function.
diff --git a/lib/stylegan2/dnnlib/submission/__init__.py b/lib/stylegan2/dnnlib/submission/__init__.py
deleted file mode 100644
index acf2fbee4b216cb9f2a0b73993fd1c7042e2248d..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/submission/__init__.py
+++ /dev/null
@@ -1,8 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-from . import run_context
-from . import submit
diff --git a/lib/stylegan2/dnnlib/submission/internal/__init__.py b/lib/stylegan2/dnnlib/submission/internal/__init__.py
deleted file mode 100644
index 0f11279893d6056e8cb6f9e04e12aad07a776496..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/submission/internal/__init__.py
+++ /dev/null
@@ -1,7 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-from . import local
diff --git a/lib/stylegan2/dnnlib/submission/internal/local.py b/lib/stylegan2/dnnlib/submission/internal/local.py
deleted file mode 100644
index c03c79e93ca19704157782a0bae556a7752b775c..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/submission/internal/local.py
+++ /dev/null
@@ -1,22 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-class TargetOptions():
- def __init__(self):
- self.do_not_copy_source_files = False
-
-class Target():
- def __init__(self):
- pass
-
- def finalize_submit_config(self, submit_config, host_run_dir):
- print ('Local submit ', end='', flush=True)
- submit_config.run_dir = host_run_dir
-
- def submit(self, submit_config, host_run_dir):
- from ..submit import run_wrapper, convert_path
- print('- run_dir: %s' % convert_path(submit_config.run_dir), flush=True)
- return run_wrapper(submit_config)
diff --git a/lib/stylegan2/dnnlib/submission/run_context.py b/lib/stylegan2/dnnlib/submission/run_context.py
deleted file mode 100644
index 62fbb1afd86be9d5fa963a1958485a2fc6d1152a..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/submission/run_context.py
+++ /dev/null
@@ -1,110 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Helpers for managing the run/training loop."""
-
-import datetime
-import json
-import os
-import pprint
-import time
-import types
-
-from typing import Any
-
-from . import submit
-
-# Singleton RunContext
-_run_context = None
-
-class RunContext(object):
- """Helper class for managing the run/training loop.
-
- The context will hide the implementation details of a basic run/training loop.
- It will set things up properly, tell if run should be stopped, and then cleans up.
- User should call update periodically and use should_stop to determine if run should be stopped.
-
- Args:
- submit_config: The SubmitConfig that is used for the current run.
- config_module: (deprecated) The whole config module that is used for the current run.
- """
-
- def __init__(self, submit_config: submit.SubmitConfig, config_module: types.ModuleType = None):
- global _run_context
- # Only a single RunContext can be alive
- assert _run_context is None
- _run_context = self
- self.submit_config = submit_config
- self.should_stop_flag = False
- self.has_closed = False
- self.start_time = time.time()
- self.last_update_time = time.time()
- self.last_update_interval = 0.0
- self.progress_monitor_file_path = None
-
- # vestigial config_module support just prints a warning
- if config_module is not None:
- print("RunContext.config_module parameter support has been removed.")
-
- # write out details about the run to a text file
- self.run_txt_data = {"task_name": submit_config.task_name, "host_name": submit_config.host_name, "start_time": datetime.datetime.now().isoformat(sep=" ")}
- with open(os.path.join(submit_config.run_dir, "run.txt"), "w") as f:
- pprint.pprint(self.run_txt_data, stream=f, indent=4, width=200, compact=False)
-
- def __enter__(self) -> "RunContext":
- return self
-
- def __exit__(self, exc_type: Any, exc_value: Any, traceback: Any) -> None:
- self.close()
-
- def update(self, loss: Any = 0, cur_epoch: Any = 0, max_epoch: Any = None) -> None:
- """Do general housekeeping and keep the state of the context up-to-date.
- Should be called often enough but not in a tight loop."""
- assert not self.has_closed
-
- self.last_update_interval = time.time() - self.last_update_time
- self.last_update_time = time.time()
-
- if os.path.exists(os.path.join(self.submit_config.run_dir, "abort.txt")):
- self.should_stop_flag = True
-
- def should_stop(self) -> bool:
- """Tell whether a stopping condition has been triggered one way or another."""
- return self.should_stop_flag
-
- def get_time_since_start(self) -> float:
- """How much time has passed since the creation of the context."""
- return time.time() - self.start_time
-
- def get_time_since_last_update(self) -> float:
- """How much time has passed since the last call to update."""
- return time.time() - self.last_update_time
-
- def get_last_update_interval(self) -> float:
- """How much time passed between the previous two calls to update."""
- return self.last_update_interval
-
- def close(self) -> None:
- """Close the context and clean up.
- Should only be called once."""
- if not self.has_closed:
- # update the run.txt with stopping time
- self.run_txt_data["stop_time"] = datetime.datetime.now().isoformat(sep=" ")
- with open(os.path.join(self.submit_config.run_dir, "run.txt"), "w") as f:
- pprint.pprint(self.run_txt_data, stream=f, indent=4, width=200, compact=False)
- self.has_closed = True
-
- # detach the global singleton
- global _run_context
- if _run_context is self:
- _run_context = None
-
- @staticmethod
- def get():
- import dnnlib
- if _run_context is not None:
- return _run_context
- return RunContext(dnnlib.submit_config)
diff --git a/lib/stylegan2/dnnlib/submission/submit.py b/lib/stylegan2/dnnlib/submission/submit.py
deleted file mode 100644
index 514647dd6a0585c7bd6864380a95b8059bcfba42..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/submission/submit.py
+++ /dev/null
@@ -1,343 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Submit a function to be run either locally or in a computing cluster."""
-
-import copy
-import inspect
-import os
-import pathlib
-import pickle
-import platform
-import pprint
-import re
-import shutil
-import sys
-import time
-import traceback
-
-from enum import Enum
-
-from .. import util
-from ..util import EasyDict
-
-from . import internal
-
-class SubmitTarget(Enum):
- """The target where the function should be run.
-
- LOCAL: Run it locally.
- """
- LOCAL = 1
-
-
-class PathType(Enum):
- """Determines in which format should a path be formatted.
-
- WINDOWS: Format with Windows style.
- LINUX: Format with Linux/Posix style.
- AUTO: Use current OS type to select either WINDOWS or LINUX.
- """
- WINDOWS = 1
- LINUX = 2
- AUTO = 3
-
-
-class PlatformExtras:
- """A mixed bag of values used by dnnlib heuristics.
-
- Attributes:
-
- data_reader_buffer_size: Used by DataReader to size internal shared memory buffers.
- data_reader_process_count: Number of worker processes to spawn (zero for single thread operation)
- """
- def __init__(self):
- self.data_reader_buffer_size = 1<<30 # 1 GB
- self.data_reader_process_count = 0 # single threaded default
-
-
-_user_name_override = None
-
-class SubmitConfig(util.EasyDict):
- """Strongly typed config dict needed to submit runs.
-
- Attributes:
- run_dir_root: Path to the run dir root. Can be optionally templated with tags. Needs to always be run through get_path_from_template.
- run_desc: Description of the run. Will be used in the run dir and task name.
- run_dir_ignore: List of file patterns used to ignore files when copying files to the run dir.
- run_dir_extra_files: List of (abs_path, rel_path) tuples of file paths. rel_path root will be the src directory inside the run dir.
- submit_target: Submit target enum value. Used to select where the run is actually launched.
- num_gpus: Number of GPUs used/requested for the run.
- print_info: Whether to print debug information when submitting.
- local.do_not_copy_source_files: Do not copy source files from the working directory to the run dir.
- run_id: Automatically populated value during submit.
- run_name: Automatically populated value during submit.
- run_dir: Automatically populated value during submit.
- run_func_name: Automatically populated value during submit.
- run_func_kwargs: Automatically populated value during submit.
- user_name: Automatically populated value during submit. Can be set by the user which will then override the automatic value.
- task_name: Automatically populated value during submit.
- host_name: Automatically populated value during submit.
- platform_extras: Automatically populated values during submit. Used by various dnnlib libraries such as the DataReader class.
- """
-
- def __init__(self):
- super().__init__()
-
- # run (set these)
- self.run_dir_root = "" # should always be passed through get_path_from_template
- self.run_desc = ""
- self.run_dir_ignore = ["__pycache__", "*.pyproj", "*.sln", "*.suo", ".cache", ".idea", ".vs", ".vscode", "_cudacache"]
- self.run_dir_extra_files = []
-
- # submit (set these)
- self.submit_target = SubmitTarget.LOCAL
- self.num_gpus = 1
- self.print_info = False
- self.nvprof = False
- self.local = internal.local.TargetOptions()
- self.datasets = []
-
- # (automatically populated)
- self.run_id = None
- self.run_name = None
- self.run_dir = None
- self.run_func_name = None
- self.run_func_kwargs = None
- self.user_name = None
- self.task_name = None
- self.host_name = "localhost"
- self.platform_extras = PlatformExtras()
-
-
-def get_path_from_template(path_template: str, path_type: PathType = PathType.AUTO) -> str:
- """Replace tags in the given path template and return either Windows or Linux formatted path."""
- # automatically select path type depending on running OS
- if path_type == PathType.AUTO:
- if platform.system() == "Windows":
- path_type = PathType.WINDOWS
- elif platform.system() == "Linux":
- path_type = PathType.LINUX
- else:
- raise RuntimeError("Unknown platform")
-
- path_template = path_template.replace("<USERNAME>", get_user_name())
-
- # return correctly formatted path
- if path_type == PathType.WINDOWS:
- return str(pathlib.PureWindowsPath(path_template))
- elif path_type == PathType.LINUX:
- return str(pathlib.PurePosixPath(path_template))
- else:
- raise RuntimeError("Unknown platform")
-
-
-def get_template_from_path(path: str) -> str:
- """Convert a normal path back to its template representation."""
- path = path.replace("\\", "/")
- return path
-
-
-def convert_path(path: str, path_type: PathType = PathType.AUTO) -> str:
- """Convert a normal path to template and the convert it back to a normal path with given path type."""
- path_template = get_template_from_path(path)
- path = get_path_from_template(path_template, path_type)
- return path
-
-
-def set_user_name_override(name: str) -> None:
- """Set the global username override value."""
- global _user_name_override
- _user_name_override = name
-
-
-def get_user_name():
- """Get the current user name."""
- if _user_name_override is not None:
- return _user_name_override
- elif platform.system() == "Windows":
- return os.getlogin()
- elif platform.system() == "Linux":
- try:
- import pwd
- return pwd.getpwuid(os.geteuid()).pw_name
- except:
- return "unknown"
- else:
- raise RuntimeError("Unknown platform")
-
-
-def make_run_dir_path(*paths):
- """Make a path/filename that resides under the current submit run_dir.
-
- Args:
- *paths: Path components to be passed to os.path.join
-
- Returns:
- A file/dirname rooted at submit_config.run_dir. If there's no
- submit_config or run_dir, the base directory is the current
- working directory.
-
- E.g., `os.path.join(dnnlib.submit_config.run_dir, "output.txt"))`
- """
- import dnnlib
- if (dnnlib.submit_config is None) or (dnnlib.submit_config.run_dir is None):
- return os.path.join(os.getcwd(), *paths)
- return os.path.join(dnnlib.submit_config.run_dir, *paths)
-
-
-def _create_run_dir_local(submit_config: SubmitConfig) -> str:
- """Create a new run dir with increasing ID number at the start."""
- run_dir_root = get_path_from_template(submit_config.run_dir_root, PathType.AUTO)
-
- if not os.path.exists(run_dir_root):
- os.makedirs(run_dir_root)
-
- submit_config.run_id = _get_next_run_id_local(run_dir_root)
- submit_config.run_name = "{0:05d}-{1}".format(submit_config.run_id, submit_config.run_desc)
- run_dir = os.path.join(run_dir_root, submit_config.run_name)
-
- if os.path.exists(run_dir):
- raise RuntimeError("The run dir already exists! ({0})".format(run_dir))
-
- os.makedirs(run_dir)
-
- return run_dir
-
-
-def _get_next_run_id_local(run_dir_root: str) -> int:
- """Reads all directory names in a given directory (non-recursive) and returns the next (increasing) run id. Assumes IDs are numbers at the start of the directory names."""
- dir_names = [d for d in os.listdir(run_dir_root) if os.path.isdir(os.path.join(run_dir_root, d))]
- r = re.compile("^\\d+") # match one or more digits at the start of the string
- run_id = 0
-
- for dir_name in dir_names:
- m = r.match(dir_name)
-
- if m is not None:
- i = int(m.group())
- run_id = max(run_id, i + 1)
-
- return run_id
-
-
-def _populate_run_dir(submit_config: SubmitConfig, run_dir: str) -> None:
- """Copy all necessary files into the run dir. Assumes that the dir exists, is local, and is writable."""
- pickle.dump(submit_config, open(os.path.join(run_dir, "submit_config.pkl"), "wb"))
- with open(os.path.join(run_dir, "submit_config.txt"), "w") as f:
- pprint.pprint(submit_config, stream=f, indent=4, width=200, compact=False)
-
- if (submit_config.submit_target == SubmitTarget.LOCAL) and submit_config.local.do_not_copy_source_files:
- return
-
- files = []
-
- run_func_module_dir_path = util.get_module_dir_by_obj_name(submit_config.run_func_name)
- assert '.' in submit_config.run_func_name
- for _idx in range(submit_config.run_func_name.count('.') - 1):
- run_func_module_dir_path = os.path.dirname(run_func_module_dir_path)
- files += util.list_dir_recursively_with_ignore(run_func_module_dir_path, ignores=submit_config.run_dir_ignore, add_base_to_relative=False)
-
- dnnlib_module_dir_path = util.get_module_dir_by_obj_name("dnnlib")
- files += util.list_dir_recursively_with_ignore(dnnlib_module_dir_path, ignores=submit_config.run_dir_ignore, add_base_to_relative=True)
-
- files += submit_config.run_dir_extra_files
-
- files = [(f[0], os.path.join(run_dir, "src", f[1])) for f in files]
- files += [(os.path.join(dnnlib_module_dir_path, "submission", "internal", "run.py"), os.path.join(run_dir, "run.py"))]
-
- util.copy_files_and_create_dirs(files)
-
-
-
-def run_wrapper(submit_config: SubmitConfig) -> None:
- """Wrap the actual run function call for handling logging, exceptions, typing, etc."""
- is_local = submit_config.submit_target == SubmitTarget.LOCAL
-
- # when running locally, redirect stderr to stdout, log stdout to a file, and force flushing
- if is_local:
- logger = util.Logger(file_name=os.path.join(submit_config.run_dir, "log.txt"), file_mode="w", should_flush=True)
- else: # when running in a cluster, redirect stderr to stdout, and just force flushing (log writing is handled by run.sh)
- logger = util.Logger(file_name=None, should_flush=True)
-
- import dnnlib
- dnnlib.submit_config = submit_config
-
- exit_with_errcode = False
- try:
- print("dnnlib: Running {0}() on {1}...".format(submit_config.run_func_name, submit_config.host_name))
- start_time = time.time()
-
- run_func_obj = util.get_obj_by_name(submit_config.run_func_name)
- assert callable(run_func_obj)
- sig = inspect.signature(run_func_obj)
- if 'submit_config' in sig.parameters:
- run_func_obj(submit_config=submit_config, **submit_config.run_func_kwargs)
- else:
- run_func_obj(**submit_config.run_func_kwargs)
-
- print("dnnlib: Finished {0}() in {1}.".format(submit_config.run_func_name, util.format_time(time.time() - start_time)))
- except:
- if is_local:
- raise
- else:
- traceback.print_exc()
-
- log_src = os.path.join(submit_config.run_dir, "log.txt")
- log_dst = os.path.join(get_path_from_template(submit_config.run_dir_root), "{0}-error.txt".format(submit_config.run_name))
- shutil.copyfile(log_src, log_dst)
-
- # Defer sys.exit(1) to happen after we close the logs and create a _finished.txt
- exit_with_errcode = True
- finally:
- open(os.path.join(submit_config.run_dir, "_finished.txt"), "w").close()
-
- dnnlib.RunContext.get().close()
- dnnlib.submit_config = None
- logger.close()
-
- # If we hit an error, get out of the script now and signal the error
- # to whatever process that started this script.
- if exit_with_errcode:
- sys.exit(1)
-
- return submit_config
-
-
-def submit_run(submit_config: SubmitConfig, run_func_name: str, **run_func_kwargs) -> None:
- """Create a run dir, gather files related to the run, copy files to the run dir, and launch the run in appropriate place."""
- submit_config = copy.deepcopy(submit_config)
-
- submit_target = submit_config.submit_target
- farm = None
- if submit_target == SubmitTarget.LOCAL:
- farm = internal.local.Target()
- assert farm is not None # unknown target
-
- # Disallow submitting jobs with zero num_gpus.
- if (submit_config.num_gpus is None) or (submit_config.num_gpus == 0):
- raise RuntimeError("submit_config.num_gpus must be set to a non-zero value")
-
- if submit_config.user_name is None:
- submit_config.user_name = get_user_name()
-
- submit_config.run_func_name = run_func_name
- submit_config.run_func_kwargs = run_func_kwargs
-
- #--------------------------------------------------------------------
- # Prepare submission by populating the run dir
- #--------------------------------------------------------------------
- host_run_dir = _create_run_dir_local(submit_config)
-
- submit_config.task_name = "{0}-{1:05d}-{2}".format(submit_config.user_name, submit_config.run_id, submit_config.run_desc)
- docker_valid_name_regex = "^[a-zA-Z0-9][a-zA-Z0-9_.-]+$"
- if not re.match(docker_valid_name_regex, submit_config.task_name):
- raise RuntimeError("Invalid task name. Probable reason: unacceptable characters in your submit_config.run_desc. Task name must be accepted by the following regex: " + docker_valid_name_regex + ", got " + submit_config.task_name)
-
- # Farm specific preparations for a submit
- farm.finalize_submit_config(submit_config, host_run_dir)
- _populate_run_dir(submit_config, host_run_dir)
- return farm.submit(submit_config, host_run_dir)
diff --git a/lib/stylegan2/dnnlib/tflib/__init__.py b/lib/stylegan2/dnnlib/tflib/__init__.py
deleted file mode 100644
index 02c25173d3f2391c88b142cf80af02cd93b0b5a0..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/__init__.py
+++ /dev/null
@@ -1,18 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-from . import autosummary
-from . import network
-from . import optimizer
-from . import tfutil
-from . import custom_ops
-
-from .tfutil import *
-from .network import Network
-
-from .optimizer import Optimizer
-
-from .custom_ops import get_plugin
diff --git a/lib/stylegan2/dnnlib/tflib/autosummary.py b/lib/stylegan2/dnnlib/tflib/autosummary.py
deleted file mode 100644
index 6b0d80b371620bedadf8164772b7d6f87806fc11..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/autosummary.py
+++ /dev/null
@@ -1,191 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Helper for adding automatically tracked values to Tensorboard.
-
-Autosummary creates an identity op that internally keeps track of the input
-values and automatically shows up in TensorBoard. The reported value
-represents an average over input components. The average is accumulated
-constantly over time and flushed when save_summaries() is called.
-
-Notes:
-- The output tensor must be used as an input for something else in the
- graph. Otherwise, the autosummary op will not get executed, and the average
- value will not get accumulated.
-- It is perfectly fine to include autosummaries with the same name in
- several places throughout the graph, even if they are executed concurrently.
-- It is ok to also pass in a python scalar or numpy array. In this case, it
- is added to the average immediately.
-"""
-
-from collections import OrderedDict
-import numpy as np
-import tensorflow as tf
-from tensorboard import summary as summary_lib
-from tensorboard.plugins.custom_scalar import layout_pb2
-
-from . import tfutil
-from .tfutil import TfExpression
-from .tfutil import TfExpressionEx
-
-# Enable "Custom scalars" tab in TensorBoard for advanced formatting.
-# Disabled by default to reduce tfevents file size.
-enable_custom_scalars = False
-
-_dtype = tf.float64
-_vars = OrderedDict() # name => [var, ...]
-_immediate = OrderedDict() # name => update_op, update_value
-_finalized = False
-_merge_op = None
-
-
-def _create_var(name: str, value_expr: TfExpression) -> TfExpression:
- """Internal helper for creating autosummary accumulators."""
- assert not _finalized
- name_id = name.replace("/", "_")
- v = tf.cast(value_expr, _dtype)
-
- if v.shape.is_fully_defined():
- size = np.prod(v.shape.as_list())
- size_expr = tf.constant(size, dtype=_dtype)
- else:
- size = None
- size_expr = tf.reduce_prod(tf.cast(tf.shape(v), _dtype))
-
- if size == 1:
- if v.shape.ndims != 0:
- v = tf.reshape(v, [])
- v = [size_expr, v, tf.square(v)]
- else:
- v = [size_expr, tf.reduce_sum(v), tf.reduce_sum(tf.square(v))]
- v = tf.cond(tf.is_finite(v[1]), lambda: tf.stack(v), lambda: tf.zeros(3, dtype=_dtype))
-
- with tfutil.absolute_name_scope("Autosummary/" + name_id), tf.control_dependencies(None):
- var = tf.Variable(tf.zeros(3, dtype=_dtype), trainable=False) # [sum(1), sum(x), sum(x**2)]
- update_op = tf.cond(tf.is_variable_initialized(var), lambda: tf.assign_add(var, v), lambda: tf.assign(var, v))
-
- if name in _vars:
- _vars[name].append(var)
- else:
- _vars[name] = [var]
- return update_op
-
-
-def autosummary(name: str, value: TfExpressionEx, passthru: TfExpressionEx = None, condition: TfExpressionEx = True) -> TfExpressionEx:
- """Create a new autosummary.
-
- Args:
- name: Name to use in TensorBoard
- value: TensorFlow expression or python value to track
- passthru: Optionally return this TF node without modifications but tack an autosummary update side-effect to this node.
-
- Example use of the passthru mechanism:
-
- n = autosummary('l2loss', loss, passthru=n)
-
- This is a shorthand for the following code:
-
- with tf.control_dependencies([autosummary('l2loss', loss)]):
- n = tf.identity(n)
- """
- tfutil.assert_tf_initialized()
- name_id = name.replace("/", "_")
-
- if tfutil.is_tf_expression(value):
- with tf.name_scope("summary_" + name_id), tf.device(value.device):
- condition = tf.convert_to_tensor(condition, name='condition')
- update_op = tf.cond(condition, lambda: tf.group(_create_var(name, value)), tf.no_op)
- with tf.control_dependencies([update_op]):
- return tf.identity(value if passthru is None else passthru)
-
- else: # python scalar or numpy array
- assert not tfutil.is_tf_expression(passthru)
- assert not tfutil.is_tf_expression(condition)
- if condition:
- if name not in _immediate:
- with tfutil.absolute_name_scope("Autosummary/" + name_id), tf.device(None), tf.control_dependencies(None):
- update_value = tf.placeholder(_dtype)
- update_op = _create_var(name, update_value)
- _immediate[name] = update_op, update_value
- update_op, update_value = _immediate[name]
- tfutil.run(update_op, {update_value: value})
- return value if passthru is None else passthru
-
-
-def finalize_autosummaries() -> None:
- """Create the necessary ops to include autosummaries in TensorBoard report.
- Note: This should be done only once per graph.
- """
- global _finalized
- tfutil.assert_tf_initialized()
-
- if _finalized:
- return None
-
- _finalized = True
- tfutil.init_uninitialized_vars([var for vars_list in _vars.values() for var in vars_list])
-
- # Create summary ops.
- with tf.device(None), tf.control_dependencies(None):
- for name, vars_list in _vars.items():
- name_id = name.replace("/", "_")
- with tfutil.absolute_name_scope("Autosummary/" + name_id):
- moments = tf.add_n(vars_list)
- moments /= moments[0]
- with tf.control_dependencies([moments]): # read before resetting
- reset_ops = [tf.assign(var, tf.zeros(3, dtype=_dtype)) for var in vars_list]
- with tf.name_scope(None), tf.control_dependencies(reset_ops): # reset before reporting
- mean = moments[1]
- std = tf.sqrt(moments[2] - tf.square(moments[1]))
- tf.summary.scalar(name, mean)
- if enable_custom_scalars:
- tf.summary.scalar("xCustomScalars/" + name + "/margin_lo", mean - std)
- tf.summary.scalar("xCustomScalars/" + name + "/margin_hi", mean + std)
-
- # Setup layout for custom scalars.
- layout = None
- if enable_custom_scalars:
- cat_dict = OrderedDict()
- for series_name in sorted(_vars.keys()):
- p = series_name.split("/")
- cat = p[0] if len(p) >= 2 else ""
- chart = "/".join(p[1:-1]) if len(p) >= 3 else p[-1]
- if cat not in cat_dict:
- cat_dict[cat] = OrderedDict()
- if chart not in cat_dict[cat]:
- cat_dict[cat][chart] = []
- cat_dict[cat][chart].append(series_name)
- categories = []
- for cat_name, chart_dict in cat_dict.items():
- charts = []
- for chart_name, series_names in chart_dict.items():
- series = []
- for series_name in series_names:
- series.append(layout_pb2.MarginChartContent.Series(
- value=series_name,
- lower="xCustomScalars/" + series_name + "/margin_lo",
- upper="xCustomScalars/" + series_name + "/margin_hi"))
- margin = layout_pb2.MarginChartContent(series=series)
- charts.append(layout_pb2.Chart(title=chart_name, margin=margin))
- categories.append(layout_pb2.Category(title=cat_name, chart=charts))
- layout = summary_lib.custom_scalar_pb(layout_pb2.Layout(category=categories))
- return layout
-
-def save_summaries(file_writer, global_step=None):
- """Call FileWriter.add_summary() with all summaries in the default graph,
- automatically finalizing and merging them on the first call.
- """
- global _merge_op
- tfutil.assert_tf_initialized()
-
- if _merge_op is None:
- layout = finalize_autosummaries()
- if layout is not None:
- file_writer.add_summary(layout)
- with tf.device(None), tf.control_dependencies(None):
- _merge_op = tf.summary.merge_all()
-
- file_writer.add_summary(_merge_op.eval(), global_step)
diff --git a/lib/stylegan2/dnnlib/tflib/custom_ops.py b/lib/stylegan2/dnnlib/tflib/custom_ops.py
deleted file mode 100644
index bc72a8fa588a5e78cac9f5512aa3213f7a64a26f..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/custom_ops.py
+++ /dev/null
@@ -1,169 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""TensorFlow custom ops builder.
-"""
-
-import os
-import re
-import uuid
-import hashlib
-import tempfile
-import shutil
-import tensorflow as tf
-from tensorflow.python.client import device_lib # pylint: disable=no-name-in-module
-
-#----------------------------------------------------------------------------
-# Global options.
-
-cuda_cache_path = os.path.join(os.path.dirname(__file__), '_cudacache')
-cuda_cache_version_tag = 'v1'
-do_not_hash_included_headers = False # Speed up compilation by assuming that headers included by the CUDA code never change. Unsafe!
-verbose = True # Print status messages to stdout.
-
-compiler_bindir_search_path = [
- 'C:/Program Files (x86)/Microsoft Visual Studio/2017/Community/VC/Tools/MSVC/14.14.26428/bin/Hostx64/x64',
- 'C:/Program Files (x86)/Microsoft Visual Studio/2019/Community/VC/Tools/MSVC/14.23.28105/bin/Hostx64/x64',
- 'C:/Program Files (x86)/Microsoft Visual Studio 14.0/vc/bin',
-]
-
-#----------------------------------------------------------------------------
-# Internal helper funcs.
-
-def _find_compiler_bindir():
- for compiler_path in compiler_bindir_search_path:
- if os.path.isdir(compiler_path):
- return compiler_path
- return None
-
-def _get_compute_cap(device):
- caps_str = device.physical_device_desc
- m = re.search('compute capability: (\\d+).(\\d+)', caps_str)
- major = m.group(1)
- minor = m.group(2)
- return (major, minor)
-
-def _get_cuda_gpu_arch_string():
- gpus = [x for x in device_lib.list_local_devices() if x.device_type == 'GPU']
- if len(gpus) == 0:
- raise RuntimeError('No GPU devices found')
- (major, minor) = _get_compute_cap(gpus[0])
- return 'sm_%s%s' % (major, minor)
-
-def _run_cmd(cmd):
- with os.popen(cmd) as pipe:
- output = pipe.read()
- status = pipe.close()
- if status is not None:
- raise RuntimeError('NVCC returned an error. See below for full command line and output log:\n\n%s\n\n%s' % (cmd, output))
-
-def _prepare_nvcc_cli(opts):
- cmd = 'nvcc ' + opts.strip()
- cmd += ' --disable-warnings'
- cmd += ' --include-path "%s"' % tf.sysconfig.get_include()
- cmd += ' --include-path "%s"' % os.path.join(tf.sysconfig.get_include(), 'external', 'protobuf_archive', 'src')
- cmd += ' --include-path "%s"' % os.path.join(tf.sysconfig.get_include(), 'external', 'com_google_absl')
- cmd += ' --include-path "%s"' % os.path.join(tf.sysconfig.get_include(), 'external', 'eigen_archive')
-
- compiler_bindir = _find_compiler_bindir()
- if compiler_bindir is None:
- # Require that _find_compiler_bindir succeeds on Windows. Allow
- # nvcc to use whatever is the default on Linux.
- if os.name == 'nt':
- raise RuntimeError('Could not find MSVC/GCC/CLANG installation on this computer. Check compiler_bindir_search_path list in "%s".' % __file__)
- else:
- cmd += ' --compiler-bindir "%s"' % compiler_bindir
- cmd += ' 2>&1'
- return cmd
-
-#----------------------------------------------------------------------------
-# Main entry point.
-
-_plugin_cache = dict()
-
-def get_plugin(cuda_file):
- cuda_file_base = os.path.basename(cuda_file)
- cuda_file_name, cuda_file_ext = os.path.splitext(cuda_file_base)
-
- # Already in cache?
- if cuda_file in _plugin_cache:
- return _plugin_cache[cuda_file]
-
- # Setup plugin.
- if verbose:
- print('Setting up TensorFlow plugin "%s": ' % cuda_file_base, end='', flush=True)
- try:
- # Hash CUDA source.
- md5 = hashlib.md5()
- with open(cuda_file, 'rb') as f:
- md5.update(f.read())
- md5.update(b'\n')
-
- # Hash headers included by the CUDA code by running it through the preprocessor.
- if not do_not_hash_included_headers:
- if verbose:
- print('Preprocessing... ', end='', flush=True)
- with tempfile.TemporaryDirectory() as tmp_dir:
- tmp_file = os.path.join(tmp_dir, cuda_file_name + '_tmp' + cuda_file_ext)
- _run_cmd(_prepare_nvcc_cli('"%s" --preprocess -o "%s" --keep --keep-dir "%s"' % (cuda_file, tmp_file, tmp_dir)))
- with open(tmp_file, 'rb') as f:
- bad_file_str = ('"' + cuda_file.replace('\\', '/') + '"').encode('utf-8') # __FILE__ in error check macros
- good_file_str = ('"' + cuda_file_base + '"').encode('utf-8')
- for ln in f:
- if not ln.startswith(b'# ') and not ln.startswith(b'#line '): # ignore line number pragmas
- ln = ln.replace(bad_file_str, good_file_str)
- md5.update(ln)
- md5.update(b'\n')
-
- # Select compiler options.
- compile_opts = ''
- if os.name == 'nt':
- compile_opts += '"%s"' % os.path.join(tf.sysconfig.get_lib(), 'python', '_pywrap_tensorflow_internal.lib')
- elif os.name == 'posix':
- compile_opts += '"%s"' % os.path.join(tf.sysconfig.get_lib(), 'python', '_pywrap_tensorflow_internal.so')
- compile_opts += ' --compiler-options \'-fPIC -D_GLIBCXX_USE_CXX11_ABI=1\''
- else:
- assert False # not Windows or Linux, w00t?
- compile_opts += ' --gpu-architecture=%s' % _get_cuda_gpu_arch_string()
- compile_opts += ' --use_fast_math'
- nvcc_cmd = _prepare_nvcc_cli(compile_opts)
-
- # Hash build configuration.
- md5.update(('nvcc_cmd: ' + nvcc_cmd).encode('utf-8') + b'\n')
- md5.update(('tf.VERSION: ' + tf.VERSION).encode('utf-8') + b'\n')
- md5.update(('cuda_cache_version_tag: ' + cuda_cache_version_tag).encode('utf-8') + b'\n')
-
- # Compile if not already compiled.
- bin_file_ext = '.dll' if os.name == 'nt' else '.so'
- bin_file = os.path.join(cuda_cache_path, cuda_file_name + '_' + md5.hexdigest() + bin_file_ext)
- if not os.path.isfile(bin_file):
- if verbose:
- print('Compiling... ', end='', flush=True)
- with tempfile.TemporaryDirectory() as tmp_dir:
- tmp_file = os.path.join(tmp_dir, cuda_file_name + '_tmp' + bin_file_ext)
- _run_cmd(nvcc_cmd + ' "%s" --shared -o "%s" --keep --keep-dir "%s"' % (cuda_file, tmp_file, tmp_dir))
- os.makedirs(cuda_cache_path, exist_ok=True)
- intermediate_file = os.path.join(cuda_cache_path, cuda_file_name + '_' + uuid.uuid4().hex + '_tmp' + bin_file_ext)
- shutil.copyfile(tmp_file, intermediate_file)
- os.rename(intermediate_file, bin_file) # atomic
-
- # Load.
- if verbose:
- print('Loading... ', end='', flush=True)
- plugin = tf.load_op_library(bin_file)
-
- # Add to cache.
- _plugin_cache[cuda_file] = plugin
- if verbose:
- print('Done.', flush=True)
- return plugin
-
- except:
- if verbose:
- print('Failed!', flush=True)
- raise
-
-#----------------------------------------------------------------------------
diff --git a/lib/stylegan2/dnnlib/tflib/network.py b/lib/stylegan2/dnnlib/tflib/network.py
deleted file mode 100644
index 253fe36584e65d84d195cba71aecab09feaf23d1..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/network.py
+++ /dev/null
@@ -1,590 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Helper for managing networks."""
-
-import types
-import inspect
-import re
-import uuid
-import sys
-import numpy as np
-import tensorflow as tf
-
-from collections import OrderedDict
-from typing import Any, List, Tuple, Union
-
-from . import tfutil
-from .. import util
-
-from .tfutil import TfExpression, TfExpressionEx
-
-_import_handlers = [] # Custom import handlers for dealing with legacy data in pickle import.
-_import_module_src = dict() # Source code for temporary modules created during pickle import.
-
-
-def import_handler(handler_func):
- """Function decorator for declaring custom import handlers."""
- _import_handlers.append(handler_func)
- return handler_func
-
-
-class Network:
- """Generic network abstraction.
-
- Acts as a convenience wrapper for a parameterized network construction
- function, providing several utility methods and convenient access to
- the inputs/outputs/weights.
-
- Network objects can be safely pickled and unpickled for long-term
- archival purposes. The pickling works reliably as long as the underlying
- network construction function is defined in a standalone Python module
- that has no side effects or application-specific imports.
-
- Args:
- name: Network name. Used to select TensorFlow name and variable scopes.
- func_name: Fully qualified name of the underlying network construction function, or a top-level function object.
- static_kwargs: Keyword arguments to be passed in to the network construction function.
-
- Attributes:
- name: User-specified name, defaults to build func name if None.
- scope: Unique TensorFlow scope containing template graph and variables, derived from the user-specified name.
- static_kwargs: Arguments passed to the user-supplied build func.
- components: Container for sub-networks. Passed to the build func, and retained between calls.
- num_inputs: Number of input tensors.
- num_outputs: Number of output tensors.
- input_shapes: Input tensor shapes (NC or NCHW), including minibatch dimension.
- output_shapes: Output tensor shapes (NC or NCHW), including minibatch dimension.
- input_shape: Short-hand for input_shapes[0].
- output_shape: Short-hand for output_shapes[0].
- input_templates: Input placeholders in the template graph.
- output_templates: Output tensors in the template graph.
- input_names: Name string for each input.
- output_names: Name string for each output.
- own_vars: Variables defined by this network (local_name => var), excluding sub-networks.
- vars: All variables (local_name => var).
- trainables: All trainable variables (local_name => var).
- var_global_to_local: Mapping from variable global names to local names.
- """
-
- def __init__(self, name: str = None, func_name: Any = None, **static_kwargs):
- tfutil.assert_tf_initialized()
- assert isinstance(name, str) or name is None
- assert func_name is not None
- assert isinstance(func_name, str) or util.is_top_level_function(func_name)
- assert util.is_pickleable(static_kwargs)
-
- self._init_fields()
- self.name = name
- self.static_kwargs = util.EasyDict(static_kwargs)
-
- # Locate the user-specified network build function.
- if util.is_top_level_function(func_name):
- func_name = util.get_top_level_function_name(func_name)
- module, self._build_func_name = util.get_module_from_obj_name(func_name)
- self._build_func = util.get_obj_from_module(module, self._build_func_name)
- assert callable(self._build_func)
-
- # Dig up source code for the module containing the build function.
- self._build_module_src = _import_module_src.get(module, None)
- if self._build_module_src is None:
- self._build_module_src = inspect.getsource(module)
-
- # Init TensorFlow graph.
- self._init_graph()
- self.reset_own_vars()
-
- def _init_fields(self) -> None:
- self.name = None
- self.scope = None
- self.static_kwargs = util.EasyDict()
- self.components = util.EasyDict()
- self.num_inputs = 0
- self.num_outputs = 0
- self.input_shapes = [[]]
- self.output_shapes = [[]]
- self.input_shape = []
- self.output_shape = []
- self.input_templates = []
- self.output_templates = []
- self.input_names = []
- self.output_names = []
- self.own_vars = OrderedDict()
- self.vars = OrderedDict()
- self.trainables = OrderedDict()
- self.var_global_to_local = OrderedDict()
-
- self._build_func = None # User-supplied build function that constructs the network.
- self._build_func_name = None # Name of the build function.
- self._build_module_src = None # Full source code of the module containing the build function.
- self._run_cache = dict() # Cached graph data for Network.run().
-
- def _init_graph(self) -> None:
- # Collect inputs.
- self.input_names = []
-
- for param in inspect.signature(self._build_func).parameters.values():
- if param.kind == param.POSITIONAL_OR_KEYWORD and param.default is param.empty:
- self.input_names.append(param.name)
-
- self.num_inputs = len(self.input_names)
- assert self.num_inputs >= 1
-
- # Choose name and scope.
- if self.name is None:
- self.name = self._build_func_name
- assert re.match("^[A-Za-z0-9_.\\-]*$", self.name)
- with tf.name_scope(None):
- self.scope = tf.get_default_graph().unique_name(self.name, mark_as_used=True)
-
- # Finalize build func kwargs.
- build_kwargs = dict(self.static_kwargs)
- build_kwargs["is_template_graph"] = True
- build_kwargs["components"] = self.components
-
- # Build template graph.
- with tfutil.absolute_variable_scope(self.scope, reuse=False), tfutil.absolute_name_scope(self.scope): # ignore surrounding scopes
- assert tf.get_variable_scope().name == self.scope
- assert tf.get_default_graph().get_name_scope() == self.scope
- with tf.control_dependencies(None): # ignore surrounding control dependencies
- self.input_templates = [tf.placeholder(tf.float32, name=name) for name in self.input_names]
- out_expr = self._build_func(*self.input_templates, **build_kwargs)
-
- # Collect outputs.
- assert tfutil.is_tf_expression(out_expr) or isinstance(out_expr, tuple)
- self.output_templates = [out_expr] if tfutil.is_tf_expression(out_expr) else list(out_expr)
- self.num_outputs = len(self.output_templates)
- assert self.num_outputs >= 1
- assert all(tfutil.is_tf_expression(t) for t in self.output_templates)
-
- # Perform sanity checks.
- if any(t.shape.ndims is None for t in self.input_templates):
- raise ValueError("Network input shapes not defined. Please call x.set_shape() for each input.")
- if any(t.shape.ndims is None for t in self.output_templates):
- raise ValueError("Network output shapes not defined. Please call x.set_shape() where applicable.")
- if any(not isinstance(comp, Network) for comp in self.components.values()):
- raise ValueError("Components of a Network must be Networks themselves.")
- if len(self.components) != len(set(comp.name for comp in self.components.values())):
- raise ValueError("Components of a Network must have unique names.")
-
- # List inputs and outputs.
- self.input_shapes = [t.shape.as_list() for t in self.input_templates]
- self.output_shapes = [t.shape.as_list() for t in self.output_templates]
- self.input_shape = self.input_shapes[0]
- self.output_shape = self.output_shapes[0]
- self.output_names = [t.name.split("/")[-1].split(":")[0] for t in self.output_templates]
-
- # List variables.
- self.own_vars = OrderedDict((var.name[len(self.scope) + 1:].split(":")[0], var) for var in tf.global_variables(self.scope + "/"))
- self.vars = OrderedDict(self.own_vars)
- self.vars.update((comp.name + "/" + name, var) for comp in self.components.values() for name, var in comp.vars.items())
- self.trainables = OrderedDict((name, var) for name, var in self.vars.items() if var.trainable)
- self.var_global_to_local = OrderedDict((var.name.split(":")[0], name) for name, var in self.vars.items())
-
- def reset_own_vars(self) -> None:
- """Re-initialize all variables of this network, excluding sub-networks."""
- tfutil.run([var.initializer for var in self.own_vars.values()])
-
- def reset_vars(self) -> None:
- """Re-initialize all variables of this network, including sub-networks."""
- tfutil.run([var.initializer for var in self.vars.values()])
-
- def reset_trainables(self) -> None:
- """Re-initialize all trainable variables of this network, including sub-networks."""
- tfutil.run([var.initializer for var in self.trainables.values()])
-
- def get_output_for(self, *in_expr: TfExpression, return_as_list: bool = False, **dynamic_kwargs) -> Union[TfExpression, List[TfExpression]]:
- """Construct TensorFlow expression(s) for the output(s) of this network, given the input expression(s)."""
- assert len(in_expr) == self.num_inputs
- assert not all(expr is None for expr in in_expr)
-
- # Finalize build func kwargs.
- build_kwargs = dict(self.static_kwargs)
- build_kwargs.update(dynamic_kwargs)
- build_kwargs["is_template_graph"] = False
- build_kwargs["components"] = self.components
-
- # Build TensorFlow graph to evaluate the network.
- with tfutil.absolute_variable_scope(self.scope, reuse=True), tf.name_scope(self.name):
- assert tf.get_variable_scope().name == self.scope
- valid_inputs = [expr for expr in in_expr if expr is not None]
- final_inputs = []
- for expr, name, shape in zip(in_expr, self.input_names, self.input_shapes):
- if expr is not None:
- expr = tf.identity(expr, name=name)
- else:
- expr = tf.zeros([tf.shape(valid_inputs[0])[0]] + shape[1:], name=name)
- final_inputs.append(expr)
- out_expr = self._build_func(*final_inputs, **build_kwargs)
-
- # Propagate input shapes back to the user-specified expressions.
- for expr, final in zip(in_expr, final_inputs):
- if isinstance(expr, tf.Tensor):
- expr.set_shape(final.shape)
-
- # Express outputs in the desired format.
- assert tfutil.is_tf_expression(out_expr) or isinstance(out_expr, tuple)
- if return_as_list:
- out_expr = [out_expr] if tfutil.is_tf_expression(out_expr) else list(out_expr)
- return out_expr
-
- def get_var_local_name(self, var_or_global_name: Union[TfExpression, str]) -> str:
- """Get the local name of a given variable, without any surrounding name scopes."""
- assert tfutil.is_tf_expression(var_or_global_name) or isinstance(var_or_global_name, str)
- global_name = var_or_global_name if isinstance(var_or_global_name, str) else var_or_global_name.name
- return self.var_global_to_local[global_name]
-
- def find_var(self, var_or_local_name: Union[TfExpression, str]) -> TfExpression:
- """Find variable by local or global name."""
- assert tfutil.is_tf_expression(var_or_local_name) or isinstance(var_or_local_name, str)
- return self.vars[var_or_local_name] if isinstance(var_or_local_name, str) else var_or_local_name
-
- def get_var(self, var_or_local_name: Union[TfExpression, str]) -> np.ndarray:
- """Get the value of a given variable as NumPy array.
- Note: This method is very inefficient -- prefer to use tflib.run(list_of_vars) whenever possible."""
- return self.find_var(var_or_local_name).eval()
-
- def set_var(self, var_or_local_name: Union[TfExpression, str], new_value: Union[int, float, np.ndarray]) -> None:
- """Set the value of a given variable based on the given NumPy array.
- Note: This method is very inefficient -- prefer to use tflib.set_vars() whenever possible."""
- tfutil.set_vars({self.find_var(var_or_local_name): new_value})
-
- def __getstate__(self) -> dict:
- """Pickle export."""
- state = dict()
- state["version"] = 4
- state["name"] = self.name
- state["static_kwargs"] = dict(self.static_kwargs)
- state["components"] = dict(self.components)
- state["build_module_src"] = self._build_module_src
- state["build_func_name"] = self._build_func_name
- state["variables"] = list(zip(self.own_vars.keys(), tfutil.run(list(self.own_vars.values()))))
- return state
-
- def __setstate__(self, state: dict) -> None:
- """Pickle import."""
- # pylint: disable=attribute-defined-outside-init
- tfutil.assert_tf_initialized()
- self._init_fields()
-
- # Execute custom import handlers.
- for handler in _import_handlers:
- state = handler(state)
-
- # Set basic fields.
- assert state["version"] in [2, 3, 4]
- self.name = state["name"]
- self.static_kwargs = util.EasyDict(state["static_kwargs"])
- self.components = util.EasyDict(state.get("components", {}))
- self._build_module_src = state["build_module_src"]
- self._build_func_name = state["build_func_name"]
-
- # Create temporary module from the imported source code.
- module_name = "_tflib_network_import_" + uuid.uuid4().hex
- module = types.ModuleType(module_name)
- sys.modules[module_name] = module
- _import_module_src[module] = self._build_module_src
- exec(self._build_module_src, module.__dict__) # pylint: disable=exec-used
-
- # Locate network build function in the temporary module.
- self._build_func = util.get_obj_from_module(module, self._build_func_name)
- assert callable(self._build_func)
-
- # Init TensorFlow graph.
- self._init_graph()
- self.reset_own_vars()
- tfutil.set_vars({self.find_var(name): value for name, value in state["variables"]})
-
- def clone(self, name: str = None, **new_static_kwargs) -> "Network":
- """Create a clone of this network with its own copy of the variables."""
- # pylint: disable=protected-access
- net = object.__new__(Network)
- net._init_fields()
- net.name = name if name is not None else self.name
- net.static_kwargs = util.EasyDict(self.static_kwargs)
- net.static_kwargs.update(new_static_kwargs)
- net._build_module_src = self._build_module_src
- net._build_func_name = self._build_func_name
- net._build_func = self._build_func
- net._init_graph()
- net.copy_vars_from(self)
- return net
-
- def copy_own_vars_from(self, src_net: "Network") -> None:
- """Copy the values of all variables from the given network, excluding sub-networks."""
- names = [name for name in self.own_vars.keys() if name in src_net.own_vars]
- tfutil.set_vars(tfutil.run({self.vars[name]: src_net.vars[name] for name in names}))
-
- def copy_vars_from(self, src_net: "Network") -> None:
- """Copy the values of all variables from the given network, including sub-networks."""
- names = [name for name in self.vars.keys() if name in src_net.vars]
- tfutil.set_vars(tfutil.run({self.vars[name]: src_net.vars[name] for name in names}))
-
- def copy_trainables_from(self, src_net: "Network") -> None:
- """Copy the values of all trainable variables from the given network, including sub-networks."""
- names = [name for name in self.trainables.keys() if name in src_net.trainables]
- tfutil.set_vars(tfutil.run({self.vars[name]: src_net.vars[name] for name in names}))
-
- def convert(self, new_func_name: str, new_name: str = None, **new_static_kwargs) -> "Network":
- """Create new network with the given parameters, and copy all variables from this network."""
- if new_name is None:
- new_name = self.name
- static_kwargs = dict(self.static_kwargs)
- static_kwargs.update(new_static_kwargs)
- net = Network(name=new_name, func_name=new_func_name, **static_kwargs)
- net.copy_vars_from(self)
- return net
-
- def setup_as_moving_average_of(self, src_net: "Network", beta: TfExpressionEx = 0.99, beta_nontrainable: TfExpressionEx = 0.0) -> tf.Operation:
- """Construct a TensorFlow op that updates the variables of this network
- to be slightly closer to those of the given network."""
- with tfutil.absolute_name_scope(self.scope + "/_MovingAvg"):
- ops = []
- for name, var in self.vars.items():
- if name in src_net.vars:
- cur_beta = beta if name in self.trainables else beta_nontrainable
- new_value = tfutil.lerp(src_net.vars[name], var, cur_beta)
- ops.append(var.assign(new_value))
- return tf.group(*ops)
-
- def run(self,
- *in_arrays: Tuple[Union[np.ndarray, None], ...],
- input_transform: dict = None,
- output_transform: dict = None,
- return_as_list: bool = False,
- print_progress: bool = False,
- minibatch_size: int = None,
- num_gpus: int = 1,
- assume_frozen: bool = False,
- **dynamic_kwargs) -> Union[np.ndarray, Tuple[np.ndarray, ...], List[np.ndarray]]:
- """Run this network for the given NumPy array(s), and return the output(s) as NumPy array(s).
-
- Args:
- input_transform: A dict specifying a custom transformation to be applied to the input tensor(s) before evaluating the network.
- The dict must contain a 'func' field that points to a top-level function. The function is called with the input
- TensorFlow expression(s) as positional arguments. Any remaining fields of the dict will be passed in as kwargs.
- output_transform: A dict specifying a custom transformation to be applied to the output tensor(s) after evaluating the network.
- The dict must contain a 'func' field that points to a top-level function. The function is called with the output
- TensorFlow expression(s) as positional arguments. Any remaining fields of the dict will be passed in as kwargs.
- return_as_list: True = return a list of NumPy arrays, False = return a single NumPy array, or a tuple if there are multiple outputs.
- print_progress: Print progress to the console? Useful for very large input arrays.
- minibatch_size: Maximum minibatch size to use, None = disable batching.
- num_gpus: Number of GPUs to use.
- assume_frozen: Improve multi-GPU performance by assuming that the trainable parameters will remain changed between calls.
- dynamic_kwargs: Additional keyword arguments to be passed into the network build function.
- """
- assert len(in_arrays) == self.num_inputs
- assert not all(arr is None for arr in in_arrays)
- assert input_transform is None or util.is_top_level_function(input_transform["func"])
- assert output_transform is None or util.is_top_level_function(output_transform["func"])
- output_transform, dynamic_kwargs = _handle_legacy_output_transforms(output_transform, dynamic_kwargs)
- num_items = in_arrays[0].shape[0]
- if minibatch_size is None:
- minibatch_size = num_items
-
- # Construct unique hash key from all arguments that affect the TensorFlow graph.
- key = dict(input_transform=input_transform, output_transform=output_transform, num_gpus=num_gpus, assume_frozen=assume_frozen, dynamic_kwargs=dynamic_kwargs)
- def unwind_key(obj):
- if isinstance(obj, dict):
- return [(key, unwind_key(value)) for key, value in sorted(obj.items())]
- if callable(obj):
- return util.get_top_level_function_name(obj)
- return obj
- key = repr(unwind_key(key))
-
- # Build graph.
- if key not in self._run_cache:
- with tfutil.absolute_name_scope(self.scope + "/_Run"), tf.control_dependencies(None):
- with tf.device("/cpu:0"):
- in_expr = [tf.placeholder(tf.float32, name=name) for name in self.input_names]
- in_split = list(zip(*[tf.split(x, num_gpus) for x in in_expr]))
-
- out_split = []
- for gpu in range(num_gpus):
- with tf.device("/gpu:%d" % gpu):
- net_gpu = self.clone() if assume_frozen else self
- in_gpu = in_split[gpu]
-
- if input_transform is not None:
- in_kwargs = dict(input_transform)
- in_gpu = in_kwargs.pop("func")(*in_gpu, **in_kwargs)
- in_gpu = [in_gpu] if tfutil.is_tf_expression(in_gpu) else list(in_gpu)
-
- assert len(in_gpu) == self.num_inputs
- out_gpu = net_gpu.get_output_for(*in_gpu, return_as_list=True, **dynamic_kwargs)
-
- if output_transform is not None:
- out_kwargs = dict(output_transform)
- out_gpu = out_kwargs.pop("func")(*out_gpu, **out_kwargs)
- out_gpu = [out_gpu] if tfutil.is_tf_expression(out_gpu) else list(out_gpu)
-
- #assert len(out_gpu) == self.num_outputs
- out_split.append(out_gpu)
-
- with tf.device("/cpu:0"):
- out_expr = [tf.concat(outputs, axis=0) for outputs in zip(*out_split)]
- self._run_cache[key] = in_expr, out_expr
-
- # Run minibatches.
- in_expr, out_expr = self._run_cache[key]
- out_arrays = [np.empty([num_items] + expr.shape.as_list()[1:], expr.dtype.name) for expr in out_expr]
-
- for mb_begin in range(0, num_items, minibatch_size):
- if print_progress:
- print("\r%d / %d" % (mb_begin, num_items), end="")
-
- mb_end = min(mb_begin + minibatch_size, num_items)
- mb_num = mb_end - mb_begin
- mb_in = [src[mb_begin : mb_end] if src is not None else np.zeros([mb_num] + shape[1:]) for src, shape in zip(in_arrays, self.input_shapes)]
- mb_out = tf.get_default_session().run(out_expr, dict(zip(in_expr, mb_in)))
-
- for dst, src in zip(out_arrays, mb_out):
- dst[mb_begin: mb_end] = src
-
- # Done.
- if print_progress:
- print("\r%d / %d" % (num_items, num_items))
-
- if not return_as_list:
- out_arrays = out_arrays[0] if len(out_arrays) == 1 else tuple(out_arrays)
- return out_arrays
-
- def list_ops(self) -> List[TfExpression]:
- include_prefix = self.scope + "/"
- exclude_prefix = include_prefix + "_"
- ops = tf.get_default_graph().get_operations()
- ops = [op for op in ops if op.name.startswith(include_prefix)]
- ops = [op for op in ops if not op.name.startswith(exclude_prefix)]
- return ops
-
- def list_layers(self) -> List[Tuple[str, TfExpression, List[TfExpression]]]:
- """Returns a list of (layer_name, output_expr, trainable_vars) tuples corresponding to
- individual layers of the network. Mainly intended to be used for reporting."""
- layers = []
-
- def recurse(scope, parent_ops, parent_vars, level):
- # Ignore specific patterns.
- if any(p in scope for p in ["/Shape", "/strided_slice", "/Cast", "/concat", "/Assign"]):
- return
-
- # Filter ops and vars by scope.
- global_prefix = scope + "/"
- local_prefix = global_prefix[len(self.scope) + 1:]
- cur_ops = [op for op in parent_ops if op.name.startswith(global_prefix) or op.name == global_prefix[:-1]]
- cur_vars = [(name, var) for name, var in parent_vars if name.startswith(local_prefix) or name == local_prefix[:-1]]
- if not cur_ops and not cur_vars:
- return
-
- # Filter out all ops related to variables.
- for var in [op for op in cur_ops if op.type.startswith("Variable")]:
- var_prefix = var.name + "/"
- cur_ops = [op for op in cur_ops if not op.name.startswith(var_prefix)]
-
- # Scope does not contain ops as immediate children => recurse deeper.
- contains_direct_ops = any("/" not in op.name[len(global_prefix):] and op.type not in ["Identity", "Cast", "Transpose"] for op in cur_ops)
- if (level == 0 or not contains_direct_ops) and (len(cur_ops) + len(cur_vars)) > 1:
- visited = set()
- for rel_name in [op.name[len(global_prefix):] for op in cur_ops] + [name[len(local_prefix):] for name, _var in cur_vars]:
- token = rel_name.split("/")[0]
- if token not in visited:
- recurse(global_prefix + token, cur_ops, cur_vars, level + 1)
- visited.add(token)
- return
-
- # Report layer.
- layer_name = scope[len(self.scope) + 1:]
- layer_output = cur_ops[-1].outputs[0] if cur_ops else cur_vars[-1][1]
- layer_trainables = [var for _name, var in cur_vars if var.trainable]
- layers.append((layer_name, layer_output, layer_trainables))
-
- recurse(self.scope, self.list_ops(), list(self.vars.items()), 0)
- return layers
-
- def print_layers(self, title: str = None, hide_layers_with_no_params: bool = False) -> None:
- """Print a summary table of the network structure."""
- rows = [[title if title is not None else self.name, "Params", "OutputShape", "WeightShape"]]
- rows += [["---"] * 4]
- total_params = 0
-
- for layer_name, layer_output, layer_trainables in self.list_layers():
- num_params = sum(int(np.prod(var.shape.as_list())) for var in layer_trainables)
- weights = [var for var in layer_trainables if var.name.endswith("/weight:0")]
- weights.sort(key=lambda x: len(x.name))
- if len(weights) == 0 and len(layer_trainables) == 1:
- weights = layer_trainables
- total_params += num_params
-
- if not hide_layers_with_no_params or num_params != 0:
- num_params_str = str(num_params) if num_params > 0 else "-"
- output_shape_str = str(layer_output.shape)
- weight_shape_str = str(weights[0].shape) if len(weights) >= 1 else "-"
- rows += [[layer_name, num_params_str, output_shape_str, weight_shape_str]]
-
- rows += [["---"] * 4]
- rows += [["Total", str(total_params), "", ""]]
-
- widths = [max(len(cell) for cell in column) for column in zip(*rows)]
- print()
- for row in rows:
- print(" ".join(cell + " " * (width - len(cell)) for cell, width in zip(row, widths)))
- print()
-
- def setup_weight_histograms(self, title: str = None) -> None:
- """Construct summary ops to include histograms of all trainable parameters in TensorBoard."""
- if title is None:
- title = self.name
-
- with tf.name_scope(None), tf.device(None), tf.control_dependencies(None):
- for local_name, var in self.trainables.items():
- if "/" in local_name:
- p = local_name.split("/")
- name = title + "_" + p[-1] + "/" + "_".join(p[:-1])
- else:
- name = title + "_toplevel/" + local_name
-
- tf.summary.histogram(name, var)
-
-#----------------------------------------------------------------------------
-# Backwards-compatible emulation of legacy output transformation in Network.run().
-
-_print_legacy_warning = True
-
-def _handle_legacy_output_transforms(output_transform, dynamic_kwargs):
- global _print_legacy_warning
- legacy_kwargs = ["out_mul", "out_add", "out_shrink", "out_dtype"]
- if not any(kwarg in dynamic_kwargs for kwarg in legacy_kwargs):
- return output_transform, dynamic_kwargs
-
- if _print_legacy_warning:
- _print_legacy_warning = False
- print()
- print("WARNING: Old-style output transformations in Network.run() are deprecated.")
- print("Consider using 'output_transform=dict(func=tflib.convert_images_to_uint8)'")
- print("instead of 'out_mul=127.5, out_add=127.5, out_dtype=np.uint8'.")
- print()
- assert output_transform is None
-
- new_kwargs = dict(dynamic_kwargs)
- new_transform = {kwarg: new_kwargs.pop(kwarg) for kwarg in legacy_kwargs if kwarg in dynamic_kwargs}
- new_transform["func"] = _legacy_output_transform_func
- return new_transform, new_kwargs
-
-def _legacy_output_transform_func(*expr, out_mul=1.0, out_add=0.0, out_shrink=1, out_dtype=None):
- if out_mul != 1.0:
- expr = [x * out_mul for x in expr]
-
- if out_add != 0.0:
- expr = [x + out_add for x in expr]
-
- if out_shrink > 1:
- ksize = [1, 1, out_shrink, out_shrink]
- expr = [tf.nn.avg_pool(x, ksize=ksize, strides=ksize, padding="VALID", data_format="NCHW") for x in expr]
-
- if out_dtype is not None:
- if tf.as_dtype(out_dtype).is_integer:
- expr = [tf.round(x) for x in expr]
- expr = [tf.saturate_cast(x, out_dtype) for x in expr]
- return expr
diff --git a/lib/stylegan2/dnnlib/tflib/ops/__init__.py b/lib/stylegan2/dnnlib/tflib/ops/__init__.py
deleted file mode 100644
index 9ab9908efa3cb38af52e8d5bcaa8acffde5a8875..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/ops/__init__.py
+++ /dev/null
@@ -1,7 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-# empty
diff --git a/lib/stylegan2/dnnlib/tflib/ops/fused_bias_act.cu b/lib/stylegan2/dnnlib/tflib/ops/fused_bias_act.cu
deleted file mode 100644
index 1102f624fadd0b803bdfb99fecfe145d7ec8abc4..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/ops/fused_bias_act.cu
+++ /dev/null
@@ -1,188 +0,0 @@
-// Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-//
-// This work is made available under the Nvidia Source Code License-NC.
-// To view a copy of this license, visit
-// https://nvlabs.github.io/stylegan2/license.html
-
-#define EIGEN_USE_GPU
-#define __CUDA_INCLUDE_COMPILER_INTERNAL_HEADERS__
-#include "tensorflow/core/framework/op.h"
-#include "tensorflow/core/framework/op_kernel.h"
-#include "tensorflow/core/framework/shape_inference.h"
-#include <stdio.h>
-
-using namespace tensorflow;
-using namespace tensorflow::shape_inference;
-
-#define OP_CHECK_CUDA_ERROR(CTX, CUDA_CALL) do { cudaError_t err = CUDA_CALL; OP_REQUIRES(CTX, err == cudaSuccess, errors::Internal(cudaGetErrorName(err))); } while (false)
-
-//------------------------------------------------------------------------
-// CUDA kernel.
-
-template <class T>
-struct FusedBiasActKernelParams
-{
- const T* x; // [sizeX]
- const T* b; // [sizeB] or NULL
- const T* ref; // [sizeX] or NULL
- T* y; // [sizeX]
-
- int grad;
- int axis;
- int act;
- float alpha;
- float gain;
-
- int sizeX;
- int sizeB;
- int stepB;
- int loopX;
-};
-
-template <class T>
-static __global__ void FusedBiasActKernel(const FusedBiasActKernelParams<T> p)
-{
- const float expRange = 80.0f;
- const float halfExpRange = 40.0f;
- const float seluScale = 1.0507009873554804934193349852946f;
- const float seluAlpha = 1.6732632423543772848170429916717f;
-
- // Loop over elements.
- int xi = blockIdx.x * p.loopX * blockDim.x + threadIdx.x;
- for (int loopIdx = 0; loopIdx < p.loopX && xi < p.sizeX; loopIdx++, xi += blockDim.x)
- {
- // Load and apply bias.
- float x = (float)p.x[xi];
- if (p.b)
- x += (float)p.b[(xi / p.stepB) % p.sizeB];
- float ref = (p.ref) ? (float)p.ref[xi] : 0.0f;
- if (p.gain != 0.0f & p.act != 9)
- ref /= p.gain;
-
- // Evaluate activation func.
- float y;
- switch (p.act * 10 + p.grad)
- {
- // linear
- default:
- case 10: y = x; break;
- case 11: y = x; break;
- case 12: y = 0.0f; break;
-
- // relu
- case 20: y = (x > 0.0f) ? x : 0.0f; break;
- case 21: y = (ref > 0.0f) ? x : 0.0f; break;
- case 22: y = 0.0f; break;
-
- // lrelu
- case 30: y = (x > 0.0f) ? x : x * p.alpha; break;
- case 31: y = (ref > 0.0f) ? x : x * p.alpha; break;
- case 32: y = 0.0f; break;
-
- // tanh
- case 40: { float c = expf(x); float d = 1.0f / c; y = (x < -expRange) ? -1.0f : (x > expRange) ? 1.0f : (c - d) / (c + d); } break;
- case 41: y = x * (1.0f - ref * ref); break;
- case 42: y = x * (1.0f - ref * ref) * (-2.0f * ref); break;
-
- // sigmoid
- case 50: y = (x < -expRange) ? 0.0f : 1.0f / (expf(-x) + 1.0f); break;
- case 51: y = x * ref * (1.0f - ref); break;
- case 52: y = x * ref * (1.0f - ref) * (1.0f - 2.0f * ref); break;
-
- // elu
- case 60: y = (x >= 0.0f) ? x : expf(x) - 1.0f; break;
- case 61: y = (ref >= 0.0f) ? x : x * (ref + 1.0f); break;
- case 62: y = (ref >= 0.0f) ? 0.0f : x * (ref + 1.0f); break;
-
- // selu
- case 70: y = (x >= 0.0f) ? seluScale * x : (seluScale * seluAlpha) * (expf(x) - 1.0f); break;
- case 71: y = (ref >= 0.0f) ? x * seluScale : x * (ref + seluScale * seluAlpha); break;
- case 72: y = (ref >= 0.0f) ? 0.0f : x * (ref + seluScale * seluAlpha); break;
-
- // softplus
- case 80: y = (x > expRange) ? x : logf(expf(x) + 1.0f); break;
- case 81: y = x * (1.0f - expf(-ref)); break;
- case 82: { float c = expf(-ref); y = x * c * (1.0f - c); } break;
-
- // swish
- case 90: y = (x < -expRange) ? 0.0f : x / (expf(-x) + 1.0f); break;
- case 91: { float c = expf(ref); float d = c + 1.0f; y = (ref > halfExpRange) ? x : x * c * (ref + d) / (d * d); } break;
- case 92: { float c = expf(ref); float d = c + 1.0f; y = (ref > halfExpRange) ? 0.0f : x * c * (ref * (2.0f - d) + 2.0f * d) / (d * d * d); } break;
- }
-
- // Apply gain and store.
- p.y[xi] = (T)(y * p.gain);
- }
-}
-
-//------------------------------------------------------------------------
-// TensorFlow op.
-
-template <class T>
-struct FusedBiasActOp : public OpKernel
-{
- FusedBiasActKernelParams<T> m_attribs;
-
- FusedBiasActOp(OpKernelConstruction* ctx) : OpKernel(ctx)
- {
- memset(&m_attribs, 0, sizeof(m_attribs));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("grad", &m_attribs.grad));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("axis", &m_attribs.axis));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("act", &m_attribs.act));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("alpha", &m_attribs.alpha));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("gain", &m_attribs.gain));
- OP_REQUIRES(ctx, m_attribs.grad >= 0, errors::InvalidArgument("grad must be non-negative"));
- OP_REQUIRES(ctx, m_attribs.axis >= 0, errors::InvalidArgument("axis must be non-negative"));
- OP_REQUIRES(ctx, m_attribs.act >= 0, errors::InvalidArgument("act must be non-negative"));
- }
-
- void Compute(OpKernelContext* ctx)
- {
- FusedBiasActKernelParams<T> p = m_attribs;
- cudaStream_t stream = ctx->eigen_device<Eigen::GpuDevice>().stream();
-
- const Tensor& x = ctx->input(0); // [...]
- const Tensor& b = ctx->input(1); // [sizeB] or [0]
- const Tensor& ref = ctx->input(2); // x.shape or [0]
- p.x = x.flat<T>().data();
- p.b = (b.NumElements()) ? b.flat<T>().data() : NULL;
- p.ref = (ref.NumElements()) ? ref.flat<T>().data() : NULL;
- OP_REQUIRES(ctx, b.NumElements() == 0 || m_attribs.axis < x.dims(), errors::InvalidArgument("axis out of bounds"));
- OP_REQUIRES(ctx, b.dims() == 1, errors::InvalidArgument("b must have rank 1"));
- OP_REQUIRES(ctx, b.NumElements() == 0 || b.NumElements() == x.dim_size(m_attribs.axis), errors::InvalidArgument("b has wrong number of elements"));
- OP_REQUIRES(ctx, ref.NumElements() == ((p.grad == 0) ? 0 : x.NumElements()), errors::InvalidArgument("ref has wrong number of elements"));
- OP_REQUIRES(ctx, x.NumElements() <= kint32max, errors::InvalidArgument("x is too large"));
-
- p.sizeX = (int)x.NumElements();
- p.sizeB = (int)b.NumElements();
- p.stepB = 1;
- for (int i = m_attribs.axis + 1; i < x.dims(); i++)
- p.stepB *= (int)x.dim_size(i);
-
- Tensor* y = NULL; // x.shape
- OP_REQUIRES_OK(ctx, ctx->allocate_output(0, x.shape(), &y));
- p.y = y->flat<T>().data();
-
- p.loopX = 4;
- int blockSize = 4 * 32;
- int gridSize = (p.sizeX - 1) / (p.loopX * blockSize) + 1;
- void* args[] = {&p};
- OP_CHECK_CUDA_ERROR(ctx, cudaLaunchKernel((void*)FusedBiasActKernel<T>, gridSize, blockSize, args, 0, stream));
- }
-};
-
-REGISTER_OP("FusedBiasAct")
- .Input ("x: T")
- .Input ("b: T")
- .Input ("ref: T")
- .Output ("y: T")
- .Attr ("T: {float, half}")
- .Attr ("grad: int = 0")
- .Attr ("axis: int = 1")
- .Attr ("act: int = 0")
- .Attr ("alpha: float = 0.0")
- .Attr ("gain: float = 1.0");
-REGISTER_KERNEL_BUILDER(Name("FusedBiasAct").Device(DEVICE_GPU).TypeConstraint<float>("T"), FusedBiasActOp<float>);
-REGISTER_KERNEL_BUILDER(Name("FusedBiasAct").Device(DEVICE_GPU).TypeConstraint<Eigen::half>("T"), FusedBiasActOp<Eigen::half>);
-
-//------------------------------------------------------------------------
diff --git a/lib/stylegan2/dnnlib/tflib/ops/fused_bias_act.py b/lib/stylegan2/dnnlib/tflib/ops/fused_bias_act.py
deleted file mode 100644
index 52f6bfd77a4b0151103c1a76fa877e084831f7c4..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/ops/fused_bias_act.py
+++ /dev/null
@@ -1,196 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Custom TensorFlow ops for efficient bias and activation."""
-
-import os
-import numpy as np
-import tensorflow as tf
-from .. import custom_ops
-from ...util import EasyDict
-
-def _get_plugin():
- return custom_ops.get_plugin(os.path.splitext(__file__)[0] + '.cu')
-
-#----------------------------------------------------------------------------
-
-activation_funcs = {
- 'linear': EasyDict(func=lambda x, **_: x, def_alpha=None, def_gain=1.0, cuda_idx=1, ref='y', zero_2nd_grad=True),
- 'relu': EasyDict(func=lambda x, **_: tf.nn.relu(x), def_alpha=None, def_gain=np.sqrt(2), cuda_idx=2, ref='y', zero_2nd_grad=True),
- 'lrelu': EasyDict(func=lambda x, alpha, **_: tf.nn.leaky_relu(x, alpha), def_alpha=0.2, def_gain=np.sqrt(2), cuda_idx=3, ref='y', zero_2nd_grad=True),
- 'tanh': EasyDict(func=lambda x, **_: tf.nn.tanh(x), def_alpha=None, def_gain=1.0, cuda_idx=4, ref='y', zero_2nd_grad=False),
- 'sigmoid': EasyDict(func=lambda x, **_: tf.nn.sigmoid(x), def_alpha=None, def_gain=1.0, cuda_idx=5, ref='y', zero_2nd_grad=False),
- 'elu': EasyDict(func=lambda x, **_: tf.nn.elu(x), def_alpha=None, def_gain=1.0, cuda_idx=6, ref='y', zero_2nd_grad=False),
- 'selu': EasyDict(func=lambda x, **_: tf.nn.selu(x), def_alpha=None, def_gain=1.0, cuda_idx=7, ref='y', zero_2nd_grad=False),
- 'softplus': EasyDict(func=lambda x, **_: tf.nn.softplus(x), def_alpha=None, def_gain=1.0, cuda_idx=8, ref='y', zero_2nd_grad=False),
- 'swish': EasyDict(func=lambda x, **_: tf.nn.sigmoid(x) * x, def_alpha=None, def_gain=np.sqrt(2), cuda_idx=9, ref='x', zero_2nd_grad=False),
-}
-
-#----------------------------------------------------------------------------
-
-def fused_bias_act(x, b=None, axis=1, act='linear', alpha=None, gain=None, impl='cuda'):
- r"""Fused bias and activation function.
-
- Adds bias `b` to activation tensor `x`, evaluates activation function `act`,
- and scales the result by `gain`. Each of the steps is optional. In most cases,
- the fused op is considerably more efficient than performing the same calculation
- using standard TensorFlow ops. It supports first and second order gradients,
- but not third order gradients.
-
- Args:
- x: Input activation tensor. Can have any shape, but if `b` is defined, the
- dimension corresponding to `axis`, as well as the rank, must be known.
- b: Bias vector, or `None` to disable. Must be a 1D tensor of the same type
- as `x`. The shape must be known, and it must match the dimension of `x`
- corresponding to `axis`.
- axis: The dimension in `x` corresponding to the elements of `b`.
- The value of `axis` is ignored if `b` is not specified.
- act: Name of the activation function to evaluate, or `"linear"` to disable.
- Can be e.g. `"relu"`, `"lrelu"`, `"tanh"`, `"sigmoid"`, `"swish"`, etc.
- See `activation_funcs` for a full list. `None` is not allowed.
- alpha: Shape parameter for the activation function, or `None` to use the default.
- gain: Scaling factor for the output tensor, or `None` to use default.
- See `activation_funcs` for the default scaling of each activation function.
- If unsure, consider specifying `1.0`.
- impl: Name of the implementation to use. Can be `"ref"` or `"cuda"` (default).
-
- Returns:
- Tensor of the same shape and datatype as `x`.
- """
-
- impl_dict = {
- 'ref': _fused_bias_act_ref,
- 'cuda': _fused_bias_act_cuda,
- }
- return impl_dict[impl](x=x, b=b, axis=axis, act=act, alpha=alpha, gain=gain)
-
-#----------------------------------------------------------------------------
-
-def _fused_bias_act_ref(x, b, axis, act, alpha, gain):
- """Slow reference implementation of `fused_bias_act()` using standard TensorFlow ops."""
-
- # Validate arguments.
- x = tf.convert_to_tensor(x)
- b = tf.convert_to_tensor(b) if b is not None else tf.constant([], dtype=x.dtype)
- act_spec = activation_funcs[act]
- assert b.shape.rank == 1 and (b.shape[0] == 0 or b.shape[0] == x.shape[axis])
- assert b.shape[0] == 0 or 0 <= axis < x.shape.rank
- if alpha is None:
- alpha = act_spec.def_alpha
- if gain is None:
- gain = act_spec.def_gain
-
- # Add bias.
- if b.shape[0] != 0:
- x += tf.reshape(b, [-1 if i == axis else 1 for i in range(x.shape.rank)])
-
- # Evaluate activation function.
- x = act_spec.func(x, alpha=alpha)
-
- # Scale by gain.
- if gain != 1:
- x *= gain
- return x
-
-#----------------------------------------------------------------------------
-
-def _fused_bias_act_cuda(x, b, axis, act, alpha, gain):
- """Fast CUDA implementation of `fused_bias_act()` using custom ops."""
-
- # Validate arguments.
- x = tf.convert_to_tensor(x)
- empty_tensor = tf.constant([], dtype=x.dtype)
- b = tf.convert_to_tensor(b) if b is not None else empty_tensor
- act_spec = activation_funcs[act]
- assert b.shape.rank == 1 and (b.shape[0] == 0 or b.shape[0] == x.shape[axis])
- assert b.shape[0] == 0 or 0 <= axis < x.shape.rank
- if alpha is None:
- alpha = act_spec.def_alpha
- if gain is None:
- gain = act_spec.def_gain
-
- # Special cases.
- if act == 'linear' and b is None and gain == 1.0:
- return x
- if act_spec.cuda_idx is None:
- return _fused_bias_act_ref(x=x, b=b, axis=axis, act=act, alpha=alpha, gain=gain)
-
- # CUDA kernel.
- cuda_kernel = _get_plugin().fused_bias_act
- cuda_kwargs = dict(axis=axis, act=act_spec.cuda_idx, alpha=alpha, gain=gain)
-
- # Forward pass: y = func(x, b).
- def func_y(x, b):
- y = cuda_kernel(x=x, b=b, ref=empty_tensor, grad=0, **cuda_kwargs)
- y.set_shape(x.shape)
- return y
-
- # Backward pass: dx, db = grad(dy, x, y)
- def grad_dx(dy, x, y):
- ref = {'x': x, 'y': y}[act_spec.ref]
- dx = cuda_kernel(x=dy, b=empty_tensor, ref=ref, grad=1, **cuda_kwargs)
- dx.set_shape(x.shape)
- return dx
- def grad_db(dx):
- if b.shape[0] == 0:
- return empty_tensor
- db = dx
- if axis < x.shape.rank - 1:
- db = tf.reduce_sum(db, list(range(axis + 1, x.shape.rank)))
- if axis > 0:
- db = tf.reduce_sum(db, list(range(axis)))
- db.set_shape(b.shape)
- return db
-
- # Second order gradients: d_dy, d_x = grad2(d_dx, d_db, x, y)
- def grad2_d_dy(d_dx, d_db, x, y):
- ref = {'x': x, 'y': y}[act_spec.ref]
- d_dy = cuda_kernel(x=d_dx, b=d_db, ref=ref, grad=1, **cuda_kwargs)
- d_dy.set_shape(x.shape)
- return d_dy
- def grad2_d_x(d_dx, d_db, x, y):
- ref = {'x': x, 'y': y}[act_spec.ref]
- d_x = cuda_kernel(x=d_dx, b=d_db, ref=ref, grad=2, **cuda_kwargs)
- d_x.set_shape(x.shape)
- return d_x
-
- # Fast version for piecewise-linear activation funcs.
- @tf.custom_gradient
- def func_zero_2nd_grad(x, b):
- y = func_y(x, b)
- @tf.custom_gradient
- def grad(dy):
- dx = grad_dx(dy, x, y)
- db = grad_db(dx)
- def grad2(d_dx, d_db):
- d_dy = grad2_d_dy(d_dx, d_db, x, y)
- return d_dy
- return (dx, db), grad2
- return y, grad
-
- # Slow version for general activation funcs.
- @tf.custom_gradient
- def func_nonzero_2nd_grad(x, b):
- y = func_y(x, b)
- def grad_wrap(dy):
- @tf.custom_gradient
- def grad_impl(dy, x):
- dx = grad_dx(dy, x, y)
- db = grad_db(dx)
- def grad2(d_dx, d_db):
- d_dy = grad2_d_dy(d_dx, d_db, x, y)
- d_x = grad2_d_x(d_dx, d_db, x, y)
- return d_dy, d_x
- return (dx, db), grad2
- return grad_impl(dy, x)
- return y, grad_wrap
-
- # Which version to use?
- if act_spec.zero_2nd_grad:
- return func_zero_2nd_grad(x, b)
- return func_nonzero_2nd_grad(x, b)
-
-#----------------------------------------------------------------------------
diff --git a/lib/stylegan2/dnnlib/tflib/ops/upfirdn_2d.cu b/lib/stylegan2/dnnlib/tflib/ops/upfirdn_2d.cu
deleted file mode 100644
index b97ef36c9e5ba46a92a380dbc687e275235a1ccf..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/ops/upfirdn_2d.cu
+++ /dev/null
@@ -1,326 +0,0 @@
-// Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-//
-// This work is made available under the Nvidia Source Code License-NC.
-// To view a copy of this license, visit
-// https://nvlabs.github.io/stylegan2/license.html
-
-#define EIGEN_USE_GPU
-#define __CUDA_INCLUDE_COMPILER_INTERNAL_HEADERS__
-#include "tensorflow/core/framework/op.h"
-#include "tensorflow/core/framework/op_kernel.h"
-#include "tensorflow/core/framework/shape_inference.h"
-#include <stdio.h>
-
-using namespace tensorflow;
-using namespace tensorflow::shape_inference;
-
-//------------------------------------------------------------------------
-// Helpers.
-
-#define OP_CHECK_CUDA_ERROR(CTX, CUDA_CALL) do { cudaError_t err = CUDA_CALL; OP_REQUIRES(CTX, err == cudaSuccess, errors::Internal(cudaGetErrorName(err))); } while (false)
-
-static __host__ __device__ __forceinline__ int floorDiv(int a, int b)
-{
- int c = a / b;
- if (c * b > a)
- c--;
- return c;
-}
-
-//------------------------------------------------------------------------
-// CUDA kernel params.
-
-template <class T>
-struct UpFirDn2DKernelParams
-{
- const T* x; // [majorDim, inH, inW, minorDim]
- const T* k; // [kernelH, kernelW]
- T* y; // [majorDim, outH, outW, minorDim]
-
- int upx;
- int upy;
- int downx;
- int downy;
- int padx0;
- int padx1;
- int pady0;
- int pady1;
-
- int majorDim;
- int inH;
- int inW;
- int minorDim;
- int kernelH;
- int kernelW;
- int outH;
- int outW;
- int loopMajor;
- int loopX;
-};
-
-//------------------------------------------------------------------------
-// General CUDA implementation for large filter kernels.
-
-template <class T>
-static __global__ void UpFirDn2DKernel_large(const UpFirDn2DKernelParams<T> p)
-{
- // Calculate thread index.
- int minorIdx = blockIdx.x * blockDim.x + threadIdx.x;
- int outY = minorIdx / p.minorDim;
- minorIdx -= outY * p.minorDim;
- int outXBase = blockIdx.y * p.loopX * blockDim.y + threadIdx.y;
- int majorIdxBase = blockIdx.z * p.loopMajor;
- if (outXBase >= p.outW || outY >= p.outH || majorIdxBase >= p.majorDim)
- return;
-
- // Setup Y receptive field.
- int midY = outY * p.downy + p.upy - 1 - p.pady0;
- int inY = min(max(floorDiv(midY, p.upy), 0), p.inH);
- int h = min(max(floorDiv(midY + p.kernelH, p.upy), 0), p.inH) - inY;
- int kernelY = midY + p.kernelH - (inY + 1) * p.upy;
-
- // Loop over majorDim and outX.
- for (int loopMajor = 0, majorIdx = majorIdxBase; loopMajor < p.loopMajor && majorIdx < p.majorDim; loopMajor++, majorIdx++)
- for (int loopX = 0, outX = outXBase; loopX < p.loopX && outX < p.outW; loopX++, outX += blockDim.y)
- {
- // Setup X receptive field.
- int midX = outX * p.downx + p.upx - 1 - p.padx0;
- int inX = min(max(floorDiv(midX, p.upx), 0), p.inW);
- int w = min(max(floorDiv(midX + p.kernelW, p.upx), 0), p.inW) - inX;
- int kernelX = midX + p.kernelW - (inX + 1) * p.upx;
-
- // Initialize pointers.
- const T* xp = &p.x[((majorIdx * p.inH + inY) * p.inW + inX) * p.minorDim + minorIdx];
- const T* kp = &p.k[kernelY * p.kernelW + kernelX];
- int xpx = p.minorDim;
- int kpx = -p.upx;
- int xpy = p.inW * p.minorDim;
- int kpy = -p.upy * p.kernelW;
-
- // Inner loop.
- float v = 0.0f;
- for (int y = 0; y < h; y++)
- {
- for (int x = 0; x < w; x++)
- {
- v += (float)(*xp) * (float)(*kp);
- xp += xpx;
- kp += kpx;
- }
- xp += xpy - w * xpx;
- kp += kpy - w * kpx;
- }
-
- // Store result.
- p.y[((majorIdx * p.outH + outY) * p.outW + outX) * p.minorDim + minorIdx] = (T)v;
- }
-}
-
-//------------------------------------------------------------------------
-// Specialized CUDA implementation for small filter kernels.
-
-template <class T, int upx, int upy, int downx, int downy, int kernelW, int kernelH, int tileOutW, int tileOutH>
-static __global__ void UpFirDn2DKernel_small(const UpFirDn2DKernelParams<T> p)
-{
- //assert(kernelW % upx == 0);
- //assert(kernelH % upy == 0);
- const int tileInW = ((tileOutW - 1) * downx + kernelW - 1) / upx + 1;
- const int tileInH = ((tileOutH - 1) * downy + kernelH - 1) / upy + 1;
- __shared__ volatile float sk[kernelH][kernelW];
- __shared__ volatile float sx[tileInH][tileInW];
-
- // Calculate tile index.
- int minorIdx = blockIdx.x;
- int tileOutY = minorIdx / p.minorDim;
- minorIdx -= tileOutY * p.minorDim;
- tileOutY *= tileOutH;
- int tileOutXBase = blockIdx.y * p.loopX * tileOutW;
- int majorIdxBase = blockIdx.z * p.loopMajor;
- if (tileOutXBase >= p.outW | tileOutY >= p.outH | majorIdxBase >= p.majorDim)
- return;
-
- // Load filter kernel (flipped).
- for (int tapIdx = threadIdx.x; tapIdx < kernelH * kernelW; tapIdx += blockDim.x)
- {
- int ky = tapIdx / kernelW;
- int kx = tapIdx - ky * kernelW;
- float v = 0.0f;
- if (kx < p.kernelW & ky < p.kernelH)
- v = (float)p.k[(p.kernelH - 1 - ky) * p.kernelW + (p.kernelW - 1 - kx)];
- sk[ky][kx] = v;
- }
-
- // Loop over majorDim and outX.
- for (int loopMajor = 0, majorIdx = majorIdxBase; loopMajor < p.loopMajor & majorIdx < p.majorDim; loopMajor++, majorIdx++)
- for (int loopX = 0, tileOutX = tileOutXBase; loopX < p.loopX & tileOutX < p.outW; loopX++, tileOutX += tileOutW)
- {
- // Load input pixels.
- int tileMidX = tileOutX * downx + upx - 1 - p.padx0;
- int tileMidY = tileOutY * downy + upy - 1 - p.pady0;
- int tileInX = floorDiv(tileMidX, upx);
- int tileInY = floorDiv(tileMidY, upy);
- __syncthreads();
- for (int inIdx = threadIdx.x; inIdx < tileInH * tileInW; inIdx += blockDim.x)
- {
- int relInY = inIdx / tileInW;
- int relInX = inIdx - relInY * tileInW;
- int inX = relInX + tileInX;
- int inY = relInY + tileInY;
- float v = 0.0f;
- if (inX >= 0 & inY >= 0 & inX < p.inW & inY < p.inH)
- v = (float)p.x[((majorIdx * p.inH + inY) * p.inW + inX) * p.minorDim + minorIdx];
- sx[relInY][relInX] = v;
- }
-
- // Loop over output pixels.
- __syncthreads();
- for (int outIdx = threadIdx.x; outIdx < tileOutH * tileOutW; outIdx += blockDim.x)
- {
- int relOutY = outIdx / tileOutW;
- int relOutX = outIdx - relOutY * tileOutW;
- int outX = relOutX + tileOutX;
- int outY = relOutY + tileOutY;
-
- // Setup receptive field.
- int midX = tileMidX + relOutX * downx;
- int midY = tileMidY + relOutY * downy;
- int inX = floorDiv(midX, upx);
- int inY = floorDiv(midY, upy);
- int relInX = inX - tileInX;
- int relInY = inY - tileInY;
- int kernelX = (inX + 1) * upx - midX - 1; // flipped
- int kernelY = (inY + 1) * upy - midY - 1; // flipped
-
- // Inner loop.
- float v = 0.0f;
- #pragma unroll
- for (int y = 0; y < kernelH / upy; y++)
- #pragma unroll
- for (int x = 0; x < kernelW / upx; x++)
- v += sx[relInY + y][relInX + x] * sk[kernelY + y * upy][kernelX + x * upx];
-
- // Store result.
- if (outX < p.outW & outY < p.outH)
- p.y[((majorIdx * p.outH + outY) * p.outW + outX) * p.minorDim + minorIdx] = (T)v;
- }
- }
-}
-
-//------------------------------------------------------------------------
-// TensorFlow op.
-
-template <class T>
-struct UpFirDn2DOp : public OpKernel
-{
- UpFirDn2DKernelParams<T> m_attribs;
-
- UpFirDn2DOp(OpKernelConstruction* ctx) : OpKernel(ctx)
- {
- memset(&m_attribs, 0, sizeof(m_attribs));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("upx", &m_attribs.upx));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("upy", &m_attribs.upy));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("downx", &m_attribs.downx));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("downy", &m_attribs.downy));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("padx0", &m_attribs.padx0));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("padx1", &m_attribs.padx1));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("pady0", &m_attribs.pady0));
- OP_REQUIRES_OK(ctx, ctx->GetAttr("pady1", &m_attribs.pady1));
- OP_REQUIRES(ctx, m_attribs.upx >= 1 && m_attribs.upy >= 1, errors::InvalidArgument("upx and upy must be at least 1x1"));
- OP_REQUIRES(ctx, m_attribs.downx >= 1 && m_attribs.downy >= 1, errors::InvalidArgument("downx and downy must be at least 1x1"));
- }
-
- void Compute(OpKernelContext* ctx)
- {
- UpFirDn2DKernelParams<T> p = m_attribs;
- cudaStream_t stream = ctx->eigen_device<Eigen::GpuDevice>().stream();
-
- const Tensor& x = ctx->input(0); // [majorDim, inH, inW, minorDim]
- const Tensor& k = ctx->input(1); // [kernelH, kernelW]
- p.x = x.flat<T>().data();
- p.k = k.flat<T>().data();
- OP_REQUIRES(ctx, x.dims() == 4, errors::InvalidArgument("input must have rank 4"));
- OP_REQUIRES(ctx, k.dims() == 2, errors::InvalidArgument("kernel must have rank 2"));
- OP_REQUIRES(ctx, x.NumElements() <= kint32max, errors::InvalidArgument("input too large"));
- OP_REQUIRES(ctx, k.NumElements() <= kint32max, errors::InvalidArgument("kernel too large"));
-
- p.majorDim = (int)x.dim_size(0);
- p.inH = (int)x.dim_size(1);
- p.inW = (int)x.dim_size(2);
- p.minorDim = (int)x.dim_size(3);
- p.kernelH = (int)k.dim_size(0);
- p.kernelW = (int)k.dim_size(1);
- OP_REQUIRES(ctx, p.kernelW >= 1 && p.kernelH >= 1, errors::InvalidArgument("kernel must be at least 1x1"));
-
- p.outW = (p.inW * p.upx + p.padx0 + p.padx1 - p.kernelW + p.downx) / p.downx;
- p.outH = (p.inH * p.upy + p.pady0 + p.pady1 - p.kernelH + p.downy) / p.downy;
- OP_REQUIRES(ctx, p.outW >= 1 && p.outH >= 1, errors::InvalidArgument("output must be at least 1x1"));
-
- Tensor* y = NULL; // [majorDim, outH, outW, minorDim]
- TensorShape ys;
- ys.AddDim(p.majorDim);
- ys.AddDim(p.outH);
- ys.AddDim(p.outW);
- ys.AddDim(p.minorDim);
- OP_REQUIRES_OK(ctx, ctx->allocate_output(0, ys, &y));
- p.y = y->flat<T>().data();
- OP_REQUIRES(ctx, y->NumElements() <= kint32max, errors::InvalidArgument("output too large"));
-
- // Choose CUDA kernel to use.
- void* cudaKernel = (void*)UpFirDn2DKernel_large<T>;
- int tileOutW = -1;
- int tileOutH = -1;
- if (p.upx == 1 && p.upy == 1 && p.downx == 1 && p.downy == 1 && p.kernelW <= 7 && p.kernelH <= 7) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 1,1, 1,1, 7,7, 64,16>; tileOutW = 64; tileOutH = 16; }
- if (p.upx == 1 && p.upy == 1 && p.downx == 1 && p.downy == 1 && p.kernelW <= 6 && p.kernelH <= 6) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 1,1, 1,1, 6,6, 64,16>; tileOutW = 64; tileOutH = 16; }
- if (p.upx == 1 && p.upy == 1 && p.downx == 1 && p.downy == 1 && p.kernelW <= 5 && p.kernelH <= 5) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 1,1, 1,1, 5,5, 64,16>; tileOutW = 64; tileOutH = 16; }
- if (p.upx == 1 && p.upy == 1 && p.downx == 1 && p.downy == 1 && p.kernelW <= 4 && p.kernelH <= 4) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 1,1, 1,1, 4,4, 64,16>; tileOutW = 64; tileOutH = 16; }
- if (p.upx == 1 && p.upy == 1 && p.downx == 1 && p.downy == 1 && p.kernelW <= 3 && p.kernelH <= 3) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 1,1, 1,1, 3,3, 64,16>; tileOutW = 64; tileOutH = 16; }
- if (p.upx == 2 && p.upy == 2 && p.downx == 1 && p.downy == 1 && p.kernelW <= 8 && p.kernelH <= 8) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 2,2, 1,1, 8,8, 64,16>; tileOutW = 64; tileOutH = 16; }
- if (p.upx == 2 && p.upy == 2 && p.downx == 1 && p.downy == 1 && p.kernelW <= 6 && p.kernelH <= 6) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 2,2, 1,1, 6,6, 64,16>; tileOutW = 64; tileOutH = 16; }
- if (p.upx == 2 && p.upy == 2 && p.downx == 1 && p.downy == 1 && p.kernelW <= 4 && p.kernelH <= 4) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 2,2, 1,1, 4,4, 64,16>; tileOutW = 64; tileOutH = 16; }
- if (p.upx == 2 && p.upy == 2 && p.downx == 1 && p.downy == 1 && p.kernelW <= 2 && p.kernelH <= 2) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 2,2, 1,1, 2,2, 64,16>; tileOutW = 64; tileOutH = 16; }
- if (p.upx == 1 && p.upy == 1 && p.downx == 2 && p.downy == 2 && p.kernelW <= 8 && p.kernelH <= 8) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 1,1, 2,2, 8,8, 32,8>; tileOutW = 32; tileOutH = 8; }
- if (p.upx == 1 && p.upy == 1 && p.downx == 2 && p.downy == 2 && p.kernelW <= 6 && p.kernelH <= 6) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 1,1, 2,2, 6,6, 32,8>; tileOutW = 32; tileOutH = 8; }
- if (p.upx == 1 && p.upy == 1 && p.downx == 2 && p.downy == 2 && p.kernelW <= 4 && p.kernelH <= 4) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 1,1, 2,2, 4,4, 32,8>; tileOutW = 32; tileOutH = 8; }
- if (p.upx == 1 && p.upy == 1 && p.downx == 2 && p.downy == 2 && p.kernelW <= 2 && p.kernelH <= 2) { cudaKernel = (void*)UpFirDn2DKernel_small<T, 1,1, 2,2, 2,2, 32,8>; tileOutW = 32; tileOutH = 8; }
-
- // Choose launch params.
- dim3 blockSize;
- dim3 gridSize;
- if (tileOutW > 0 && tileOutH > 0) // small
- {
- p.loopMajor = (p.majorDim - 1) / 16384 + 1;
- p.loopX = 1;
- blockSize = dim3(32 * 8, 1, 1);
- gridSize = dim3(((p.outH - 1) / tileOutH + 1) * p.minorDim, (p.outW - 1) / (p.loopX * tileOutW) + 1, (p.majorDim - 1) / p.loopMajor + 1);
- }
- else // large
- {
- p.loopMajor = (p.majorDim - 1) / 16384 + 1;
- p.loopX = 4;
- blockSize = dim3(4, 32, 1);
- gridSize = dim3((p.outH * p.minorDim - 1) / blockSize.x + 1, (p.outW - 1) / (p.loopX * blockSize.y) + 1, (p.majorDim - 1) / p.loopMajor + 1);
- }
-
- // Launch CUDA kernel.
- void* args[] = {&p};
- OP_CHECK_CUDA_ERROR(ctx, cudaLaunchKernel(cudaKernel, gridSize, blockSize, args, 0, stream));
- }
-};
-
-REGISTER_OP("UpFirDn2D")
- .Input ("x: T")
- .Input ("k: T")
- .Output ("y: T")
- .Attr ("T: {float, half}")
- .Attr ("upx: int = 1")
- .Attr ("upy: int = 1")
- .Attr ("downx: int = 1")
- .Attr ("downy: int = 1")
- .Attr ("padx0: int = 0")
- .Attr ("padx1: int = 0")
- .Attr ("pady0: int = 0")
- .Attr ("pady1: int = 0");
-REGISTER_KERNEL_BUILDER(Name("UpFirDn2D").Device(DEVICE_GPU).TypeConstraint<float>("T"), UpFirDn2DOp<float>);
-REGISTER_KERNEL_BUILDER(Name("UpFirDn2D").Device(DEVICE_GPU).TypeConstraint<Eigen::half>("T"), UpFirDn2DOp<Eigen::half>);
-
-//------------------------------------------------------------------------
diff --git a/lib/stylegan2/dnnlib/tflib/ops/upfirdn_2d.py b/lib/stylegan2/dnnlib/tflib/ops/upfirdn_2d.py
deleted file mode 100644
index fd23777ebb87bc83e8728d6fe3904fbbfb5c524c..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/ops/upfirdn_2d.py
+++ /dev/null
@@ -1,364 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Custom TensorFlow ops for efficient resampling of 2D images."""
-
-import os
-import numpy as np
-import tensorflow as tf
-from .. import custom_ops
-
-def _get_plugin():
- return custom_ops.get_plugin(os.path.splitext(__file__)[0] + '.cu')
-
-#----------------------------------------------------------------------------
-
-def upfirdn_2d(x, k, upx=1, upy=1, downx=1, downy=1, padx0=0, padx1=0, pady0=0, pady1=0, impl='cuda'):
- r"""Pad, upsample, FIR filter, and downsample a batch of 2D images.
-
- Accepts a batch of 2D images of the shape `[majorDim, inH, inW, minorDim]`
- and performs the following operations for each image, batched across
- `majorDim` and `minorDim`:
-
- 1. Pad the image with zeros by the specified number of pixels on each side
- (`padx0`, `padx1`, `pady0`, `pady1`). Specifying a negative value
- corresponds to cropping the image.
-
- 2. Upsample the image by inserting the zeros after each pixel (`upx`, `upy`).
-
- 3. Convolve the image with the specified 2D FIR filter (`k`), shrinking the
- image so that the footprint of all output pixels lies within the input image.
-
- 4. Downsample the image by throwing away pixels (`downx`, `downy`).
-
- This sequence of operations bears close resemblance to scipy.signal.upfirdn().
- The fused op is considerably more efficient than performing the same calculation
- using standard TensorFlow ops. It supports gradients of arbitrary order.
-
- Args:
- x: Input tensor of the shape `[majorDim, inH, inW, minorDim]`.
- k: 2D FIR filter of the shape `[firH, firW]`.
- upx: Integer upsampling factor along the X-axis (default: 1).
- upy: Integer upsampling factor along the Y-axis (default: 1).
- downx: Integer downsampling factor along the X-axis (default: 1).
- downy: Integer downsampling factor along the Y-axis (default: 1).
- padx0: Number of pixels to pad on the left side (default: 0).
- padx1: Number of pixels to pad on the right side (default: 0).
- pady0: Number of pixels to pad on the top side (default: 0).
- pady1: Number of pixels to pad on the bottom side (default: 0).
- impl: Name of the implementation to use. Can be `"ref"` or `"cuda"` (default).
-
- Returns:
- Tensor of the shape `[majorDim, outH, outW, minorDim]`, and same datatype as `x`.
- """
-
- impl_dict = {
- 'ref': _upfirdn_2d_ref,
- 'cuda': _upfirdn_2d_cuda,
- }
- return impl_dict[impl](x=x, k=k, upx=upx, upy=upy, downx=downx, downy=downy, padx0=padx0, padx1=padx1, pady0=pady0, pady1=pady1)
-
-#----------------------------------------------------------------------------
-
-def _upfirdn_2d_ref(x, k, upx, upy, downx, downy, padx0, padx1, pady0, pady1):
- """Slow reference implementation of `upfirdn_2d()` using standard TensorFlow ops."""
-
- x = tf.convert_to_tensor(x)
- k = np.asarray(k, dtype=np.float32)
- assert x.shape.rank == 4
- inH = x.shape[1].value
- inW = x.shape[2].value
- minorDim = _shape(x, 3)
- kernelH, kernelW = k.shape
- assert inW >= 1 and inH >= 1
- assert kernelW >= 1 and kernelH >= 1
- assert isinstance(upx, int) and isinstance(upy, int)
- assert isinstance(downx, int) and isinstance(downy, int)
- assert isinstance(padx0, int) and isinstance(padx1, int)
- assert isinstance(pady0, int) and isinstance(pady1, int)
-
- # Upsample (insert zeros).
- x = tf.reshape(x, [-1, inH, 1, inW, 1, minorDim])
- x = tf.pad(x, [[0, 0], [0, 0], [0, upy - 1], [0, 0], [0, upx - 1], [0, 0]])
- x = tf.reshape(x, [-1, inH * upy, inW * upx, minorDim])
-
- # Pad (crop if negative).
- x = tf.pad(x, [[0, 0], [max(pady0, 0), max(pady1, 0)], [max(padx0, 0), max(padx1, 0)], [0, 0]])
- x = x[:, max(-pady0, 0) : x.shape[1].value - max(-pady1, 0), max(-padx0, 0) : x.shape[2].value - max(-padx1, 0), :]
-
- # Convolve with filter.
- x = tf.transpose(x, [0, 3, 1, 2])
- x = tf.reshape(x, [-1, 1, inH * upy + pady0 + pady1, inW * upx + padx0 + padx1])
- w = tf.constant(k[::-1, ::-1, np.newaxis, np.newaxis], dtype=x.dtype)
- x = tf.nn.conv2d(x, w, strides=[1,1,1,1], padding='VALID', data_format='NCHW')
- x = tf.reshape(x, [-1, minorDim, inH * upy + pady0 + pady1 - kernelH + 1, inW * upx + padx0 + padx1 - kernelW + 1])
- x = tf.transpose(x, [0, 2, 3, 1])
-
- # Downsample (throw away pixels).
- return x[:, ::downy, ::downx, :]
-
-#----------------------------------------------------------------------------
-
-def _upfirdn_2d_cuda(x, k, upx, upy, downx, downy, padx0, padx1, pady0, pady1):
- """Fast CUDA implementation of `upfirdn_2d()` using custom ops."""
-
- x = tf.convert_to_tensor(x)
- k = np.asarray(k, dtype=np.float32)
- majorDim, inH, inW, minorDim = x.shape.as_list()
- kernelH, kernelW = k.shape
- assert inW >= 1 and inH >= 1
- assert kernelW >= 1 and kernelH >= 1
- assert isinstance(upx, int) and isinstance(upy, int)
- assert isinstance(downx, int) and isinstance(downy, int)
- assert isinstance(padx0, int) and isinstance(padx1, int)
- assert isinstance(pady0, int) and isinstance(pady1, int)
-
- outW = (inW * upx + padx0 + padx1 - kernelW) // downx + 1
- outH = (inH * upy + pady0 + pady1 - kernelH) // downy + 1
- assert outW >= 1 and outH >= 1
-
- kc = tf.constant(k, dtype=x.dtype)
- gkc = tf.constant(k[::-1, ::-1], dtype=x.dtype)
- gpadx0 = kernelW - padx0 - 1
- gpady0 = kernelH - pady0 - 1
- gpadx1 = inW * upx - outW * downx + padx0 - upx + 1
- gpady1 = inH * upy - outH * downy + pady0 - upy + 1
-
- @tf.custom_gradient
- def func(x):
- y = _get_plugin().up_fir_dn2d(x=x, k=kc, upx=upx, upy=upy, downx=downx, downy=downy, padx0=padx0, padx1=padx1, pady0=pady0, pady1=pady1)
- y.set_shape([majorDim, outH, outW, minorDim])
- @tf.custom_gradient
- def grad(dy):
- dx = _get_plugin().up_fir_dn2d(x=dy, k=gkc, upx=downx, upy=downy, downx=upx, downy=upy, padx0=gpadx0, padx1=gpadx1, pady0=gpady0, pady1=gpady1)
- dx.set_shape([majorDim, inH, inW, minorDim])
- return dx, func
- return y, grad
- return func(x)
-
-#----------------------------------------------------------------------------
-
-def filter_2d(x, k, gain=1, data_format='NCHW', impl='cuda'):
- r"""Filter a batch of 2D images with the given FIR filter.
-
- Accepts a batch of 2D images of the shape `[N, C, H, W]` or `[N, H, W, C]`
- and filters each image with the given filter. The filter is normalized so that
- if the input pixels are constant, they will be scaled by the specified `gain`.
- Pixels outside the image are assumed to be zero.
-
- Args:
- x: Input tensor of the shape `[N, C, H, W]` or `[N, H, W, C]`.
- k: FIR filter of the shape `[firH, firW]` or `[firN]` (separable).
- gain: Scaling factor for signal magnitude (default: 1.0).
- data_format: `'NCHW'` or `'NHWC'` (default: `'NCHW'`).
- impl: Name of the implementation to use. Can be `"ref"` or `"cuda"` (default).
-
- Returns:
- Tensor of the same shape and datatype as `x`.
- """
-
- k = _setup_kernel(k) * gain
- p = k.shape[0] - 1
- return _simple_upfirdn_2d(x, k, pad0=(p+1)//2, pad1=p//2, data_format=data_format, impl=impl)
-
-#----------------------------------------------------------------------------
-
-def upsample_2d(x, k=None, factor=2, gain=1, data_format='NCHW', impl='cuda'):
- r"""Upsample a batch of 2D images with the given filter.
-
- Accepts a batch of 2D images of the shape `[N, C, H, W]` or `[N, H, W, C]`
- and upsamples each image with the given filter. The filter is normalized so that
- if the input pixels are constant, they will be scaled by the specified `gain`.
- Pixels outside the image are assumed to be zero, and the filter is padded with
- zeros so that its shape is a multiple of the upsampling factor.
-
- Args:
- x: Input tensor of the shape `[N, C, H, W]` or `[N, H, W, C]`.
- k: FIR filter of the shape `[firH, firW]` or `[firN]` (separable).
- The default is `[1] * factor`, which corresponds to nearest-neighbor
- upsampling.
- factor: Integer upsampling factor (default: 2).
- gain: Scaling factor for signal magnitude (default: 1.0).
- data_format: `'NCHW'` or `'NHWC'` (default: `'NCHW'`).
- impl: Name of the implementation to use. Can be `"ref"` or `"cuda"` (default).
-
- Returns:
- Tensor of the shape `[N, C, H * factor, W * factor]` or
- `[N, H * factor, W * factor, C]`, and same datatype as `x`.
- """
-
- assert isinstance(factor, int) and factor >= 1
- if k is None:
- k = [1] * factor
- k = _setup_kernel(k) * (gain * (factor ** 2))
- p = k.shape[0] - factor
- return _simple_upfirdn_2d(x, k, up=factor, pad0=(p+1)//2+factor-1, pad1=p//2, data_format=data_format, impl=impl)
-
-#----------------------------------------------------------------------------
-
-def downsample_2d(x, k=None, factor=2, gain=1, data_format='NCHW', impl='cuda'):
- r"""Downsample a batch of 2D images with the given filter.
-
- Accepts a batch of 2D images of the shape `[N, C, H, W]` or `[N, H, W, C]`
- and downsamples each image with the given filter. The filter is normalized so that
- if the input pixels are constant, they will be scaled by the specified `gain`.
- Pixels outside the image are assumed to be zero, and the filter is padded with
- zeros so that its shape is a multiple of the downsampling factor.
-
- Args:
- x: Input tensor of the shape `[N, C, H, W]` or `[N, H, W, C]`.
- k: FIR filter of the shape `[firH, firW]` or `[firN]` (separable).
- The default is `[1] * factor`, which corresponds to average pooling.
- factor: Integer downsampling factor (default: 2).
- gain: Scaling factor for signal magnitude (default: 1.0).
- data_format: `'NCHW'` or `'NHWC'` (default: `'NCHW'`).
- impl: Name of the implementation to use. Can be `"ref"` or `"cuda"` (default).
-
- Returns:
- Tensor of the shape `[N, C, H // factor, W // factor]` or
- `[N, H // factor, W // factor, C]`, and same datatype as `x`.
- """
-
- assert isinstance(factor, int) and factor >= 1
- if k is None:
- k = [1] * factor
- k = _setup_kernel(k) * gain
- p = k.shape[0] - factor
- return _simple_upfirdn_2d(x, k, down=factor, pad0=(p+1)//2, pad1=p//2, data_format=data_format, impl=impl)
-
-#----------------------------------------------------------------------------
-
-def upsample_conv_2d(x, w, k=None, factor=2, gain=1, data_format='NCHW', impl='cuda'):
- r"""Fused `upsample_2d()` followed by `tf.nn.conv2d()`.
-
- Padding is performed only once at the beginning, not between the operations.
- The fused op is considerably more efficient than performing the same calculation
- using standard TensorFlow ops. It supports gradients of arbitrary order.
-
- Args:
- x: Input tensor of the shape `[N, C, H, W]` or `[N, H, W, C]`.
- w: Weight tensor of the shape `[filterH, filterW, inChannels, outChannels]`.
- Grouped convolution can be performed by `inChannels = x.shape[0] // numGroups`.
- k: FIR filter of the shape `[firH, firW]` or `[firN]` (separable).
- The default is `[1] * factor`, which corresponds to nearest-neighbor
- upsampling.
- factor: Integer upsampling factor (default: 2).
- gain: Scaling factor for signal magnitude (default: 1.0).
- data_format: `'NCHW'` or `'NHWC'` (default: `'NCHW'`).
- impl: Name of the implementation to use. Can be `"ref"` or `"cuda"` (default).
-
- Returns:
- Tensor of the shape `[N, C, H * factor, W * factor]` or
- `[N, H * factor, W * factor, C]`, and same datatype as `x`.
- """
-
- assert isinstance(factor, int) and factor >= 1
-
- # Check weight shape.
- w = tf.convert_to_tensor(w)
- assert w.shape.rank == 4
- convH = w.shape[0].value
- convW = w.shape[1].value
- inC = _shape(w, 2)
- outC = _shape(w, 3)
- assert convW == convH
-
- # Setup filter kernel.
- if k is None:
- k = [1] * factor
- k = _setup_kernel(k) * (gain * (factor ** 2))
- p = (k.shape[0] - factor) - (convW - 1)
-
- # Determine data dimensions.
- if data_format == 'NCHW':
- stride = [1, 1, factor, factor]
- output_shape = [_shape(x, 0), outC, (_shape(x, 2) - 1) * factor + convH, (_shape(x, 3) - 1) * factor + convW]
- num_groups = _shape(x, 1) // inC
- else:
- stride = [1, factor, factor, 1]
- output_shape = [_shape(x, 0), (_shape(x, 1) - 1) * factor + convH, (_shape(x, 2) - 1) * factor + convW, outC]
- num_groups = _shape(x, 3) // inC
-
- # Transpose weights.
- w = tf.reshape(w, [convH, convW, inC, num_groups, -1])
- w = tf.transpose(w[::-1, ::-1], [0, 1, 4, 3, 2])
- w = tf.reshape(w, [convH, convW, -1, num_groups * inC])
-
- # Execute.
- x = tf.nn.conv2d_transpose(x, w, output_shape=output_shape, strides=stride, padding='VALID', data_format=data_format)
- return _simple_upfirdn_2d(x, k, pad0=(p+1)//2+factor-1, pad1=p//2+1, data_format=data_format, impl=impl)
-
-#----------------------------------------------------------------------------
-
-def conv_downsample_2d(x, w, k=None, factor=2, gain=1, data_format='NCHW', impl='cuda'):
- r"""Fused `tf.nn.conv2d()` followed by `downsample_2d()`.
-
- Padding is performed only once at the beginning, not between the operations.
- The fused op is considerably more efficient than performing the same calculation
- using standard TensorFlow ops. It supports gradients of arbitrary order.
-
- Args:
- x: Input tensor of the shape `[N, C, H, W]` or `[N, H, W, C]`.
- w: Weight tensor of the shape `[filterH, filterW, inChannels, outChannels]`.
- Grouped convolution can be performed by `inChannels = x.shape[0] // numGroups`.
- k: FIR filter of the shape `[firH, firW]` or `[firN]` (separable).
- The default is `[1] * factor`, which corresponds to average pooling.
- factor: Integer downsampling factor (default: 2).
- gain: Scaling factor for signal magnitude (default: 1.0).
- data_format: `'NCHW'` or `'NHWC'` (default: `'NCHW'`).
- impl: Name of the implementation to use. Can be `"ref"` or `"cuda"` (default).
-
- Returns:
- Tensor of the shape `[N, C, H // factor, W // factor]` or
- `[N, H // factor, W // factor, C]`, and same datatype as `x`.
- """
-
- assert isinstance(factor, int) and factor >= 1
- w = tf.convert_to_tensor(w)
- convH, convW, _inC, _outC = w.shape.as_list()
- assert convW == convH
- if k is None:
- k = [1] * factor
- k = _setup_kernel(k) * gain
- p = (k.shape[0] - factor) + (convW - 1)
- if data_format == 'NCHW':
- s = [1, 1, factor, factor]
- else:
- s = [1, factor, factor, 1]
- x = _simple_upfirdn_2d(x, k, pad0=(p+1)//2, pad1=p//2, data_format=data_format, impl=impl)
- return tf.nn.conv2d(x, w, strides=s, padding='VALID', data_format=data_format)
-
-#----------------------------------------------------------------------------
-# Internal helper funcs.
-
-def _shape(tf_expr, dim_idx):
- if tf_expr.shape.rank is not None:
- dim = tf_expr.shape[dim_idx].value
- if dim is not None:
- return dim
- return tf.shape(tf_expr)[dim_idx]
-
-def _setup_kernel(k):
- k = np.asarray(k, dtype=np.float32)
- if k.ndim == 1:
- k = np.outer(k, k)
- k /= np.sum(k)
- assert k.ndim == 2
- assert k.shape[0] == k.shape[1]
- return k
-
-def _simple_upfirdn_2d(x, k, up=1, down=1, pad0=0, pad1=0, data_format='NCHW', impl='cuda'):
- assert data_format in ['NCHW', 'NHWC']
- assert x.shape.rank == 4
- y = x
- if data_format == 'NCHW':
- y = tf.reshape(y, [-1, _shape(y, 2), _shape(y, 3), 1])
- y = upfirdn_2d(y, k, upx=up, upy=up, downx=down, downy=down, padx0=pad0, padx1=pad1, pady0=pad0, pady1=pad1, impl=impl)
- if data_format == 'NCHW':
- y = tf.reshape(y, [-1, _shape(x, 1), _shape(y, 1), _shape(y, 2)])
- return y
-
-#----------------------------------------------------------------------------
diff --git a/lib/stylegan2/dnnlib/tflib/optimizer.py b/lib/stylegan2/dnnlib/tflib/optimizer.py
deleted file mode 100644
index 9a1b1b833e218902ef145c59a03128e2fba73baf..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/optimizer.py
+++ /dev/null
@@ -1,336 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Helper wrapper for a Tensorflow optimizer."""
-
-import numpy as np
-import tensorflow as tf
-
-from collections import OrderedDict
-from typing import List, Union
-
-from . import autosummary
-from . import tfutil
-from .. import util
-
-from .tfutil import TfExpression, TfExpressionEx
-
-try:
- # TensorFlow 1.13
- from tensorflow.python.ops import nccl_ops
-except:
- # Older TensorFlow versions
- import tensorflow.contrib.nccl as nccl_ops
-
-class Optimizer:
- """A Wrapper for tf.train.Optimizer.
-
- Automatically takes care of:
- - Gradient averaging for multi-GPU training.
- - Gradient accumulation for arbitrarily large minibatches.
- - Dynamic loss scaling and typecasts for FP16 training.
- - Ignoring corrupted gradients that contain NaNs/Infs.
- - Reporting statistics.
- - Well-chosen default settings.
- """
-
- def __init__(self,
- name: str = "Train", # Name string that will appear in TensorFlow graph.
- tf_optimizer: str = "tf.train.AdamOptimizer", # Underlying optimizer class.
- learning_rate: TfExpressionEx = 0.001, # Learning rate. Can vary over time.
- minibatch_multiplier: TfExpressionEx = None, # Treat N consecutive minibatches as one by accumulating gradients.
- share: "Optimizer" = None, # Share internal state with a previously created optimizer?
- use_loss_scaling: bool = False, # Enable dynamic loss scaling for robust mixed-precision training?
- loss_scaling_init: float = 64.0, # Log2 of initial loss scaling factor.
- loss_scaling_inc: float = 0.0005, # Log2 of per-minibatch loss scaling increment when there is no overflow.
- loss_scaling_dec: float = 1.0, # Log2 of per-minibatch loss scaling decrement when there is an overflow.
- report_mem_usage: bool = False, # Report fine-grained memory usage statistics in TensorBoard?
- **kwargs):
-
- # Public fields.
- self.name = name
- self.learning_rate = learning_rate
- self.minibatch_multiplier = minibatch_multiplier
- self.id = self.name.replace("/", ".")
- self.scope = tf.get_default_graph().unique_name(self.id)
- self.optimizer_class = util.get_obj_by_name(tf_optimizer)
- self.optimizer_kwargs = dict(kwargs)
- self.use_loss_scaling = use_loss_scaling
- self.loss_scaling_init = loss_scaling_init
- self.loss_scaling_inc = loss_scaling_inc
- self.loss_scaling_dec = loss_scaling_dec
-
- # Private fields.
- self._updates_applied = False
- self._devices = OrderedDict() # device_name => EasyDict()
- self._shared_optimizers = OrderedDict() # device_name => optimizer_class
- self._gradient_shapes = None # [shape, ...]
- self._report_mem_usage = report_mem_usage
-
- # Validate arguments.
- assert callable(self.optimizer_class)
-
- # Share internal state if requested.
- if share is not None:
- assert isinstance(share, Optimizer)
- assert self.optimizer_class is share.optimizer_class
- assert self.learning_rate is share.learning_rate
- assert self.optimizer_kwargs == share.optimizer_kwargs
- self._shared_optimizers = share._shared_optimizers # pylint: disable=protected-access
-
- def _get_device(self, device_name: str):
- """Get internal state for the given TensorFlow device."""
- tfutil.assert_tf_initialized()
- if device_name in self._devices:
- return self._devices[device_name]
-
- # Initialize fields.
- device = util.EasyDict()
- device.name = device_name
- device.optimizer = None # Underlying optimizer: optimizer_class
- device.loss_scaling_var = None # Log2 of loss scaling: tf.Variable
- device.grad_raw = OrderedDict() # Raw gradients: var => [grad, ...]
- device.grad_clean = OrderedDict() # Clean gradients: var => grad
- device.grad_acc_vars = OrderedDict() # Accumulation sums: var => tf.Variable
- device.grad_acc_count = None # Accumulation counter: tf.Variable
- device.grad_acc = OrderedDict() # Accumulated gradients: var => grad
-
- # Setup TensorFlow objects.
- with tfutil.absolute_name_scope(self.scope + "/Devices"), tf.device(device_name), tf.control_dependencies(None):
- if device_name not in self._shared_optimizers:
- optimizer_name = self.scope.replace("/", "_") + "_opt%d" % len(self._shared_optimizers)
- self._shared_optimizers[device_name] = self.optimizer_class(name=optimizer_name, learning_rate=self.learning_rate, **self.optimizer_kwargs)
- device.optimizer = self._shared_optimizers[device_name]
- if self.use_loss_scaling:
- device.loss_scaling_var = tf.Variable(np.float32(self.loss_scaling_init), trainable=False, name="loss_scaling_var")
-
- # Register device.
- self._devices[device_name] = device
- return device
-
- def register_gradients(self, loss: TfExpression, trainable_vars: Union[List, dict]) -> None:
- """Register the gradients of the given loss function with respect to the given variables.
- Intended to be called once per GPU."""
- tfutil.assert_tf_initialized()
- assert not self._updates_applied
- device = self._get_device(loss.device)
-
- # Validate trainables.
- if isinstance(trainable_vars, dict):
- trainable_vars = list(trainable_vars.values()) # allow passing in Network.trainables as vars
- assert isinstance(trainable_vars, list) and len(trainable_vars) >= 1
- assert all(tfutil.is_tf_expression(expr) for expr in trainable_vars + [loss])
- assert all(var.device == device.name for var in trainable_vars)
-
- # Validate shapes.
- if self._gradient_shapes is None:
- self._gradient_shapes = [var.shape.as_list() for var in trainable_vars]
- assert len(trainable_vars) == len(self._gradient_shapes)
- assert all(var.shape.as_list() == var_shape for var, var_shape in zip(trainable_vars, self._gradient_shapes))
-
- # Report memory usage if requested.
- deps = []
- if self._report_mem_usage:
- self._report_mem_usage = False
- try:
- with tf.name_scope(self.id + '_mem'), tf.device(device.name), tf.control_dependencies([loss]):
- deps.append(autosummary.autosummary(self.id + "/mem_usage_gb", tf.contrib.memory_stats.BytesInUse() / 2**30))
- except tf.errors.NotFoundError:
- pass
-
- # Compute gradients.
- with tf.name_scope(self.id + "_grad"), tf.device(device.name), tf.control_dependencies(deps):
- loss = self.apply_loss_scaling(tf.cast(loss, tf.float32))
- gate = tf.train.Optimizer.GATE_NONE # disable gating to reduce memory usage
- grad_list = device.optimizer.compute_gradients(loss=loss, var_list=trainable_vars, gate_gradients=gate)
-
- # Register gradients.
- for grad, var in grad_list:
- if var not in device.grad_raw:
- device.grad_raw[var] = []
- device.grad_raw[var].append(grad)
-
- def apply_updates(self, allow_no_op: bool = False) -> tf.Operation:
- """Construct training op to update the registered variables based on their gradients."""
- tfutil.assert_tf_initialized()
- assert not self._updates_applied
- self._updates_applied = True
- all_ops = []
-
- # Check for no-op.
- if allow_no_op and len(self._devices) == 0:
- with tfutil.absolute_name_scope(self.scope):
- return tf.no_op(name='TrainingOp')
-
- # Clean up gradients.
- for device_idx, device in enumerate(self._devices.values()):
- with tfutil.absolute_name_scope(self.scope + "/Clean%d" % device_idx), tf.device(device.name):
- for var, grad in device.grad_raw.items():
-
- # Filter out disconnected gradients and convert to float32.
- grad = [g for g in grad if g is not None]
- grad = [tf.cast(g, tf.float32) for g in grad]
-
- # Sum within the device.
- if len(grad) == 0:
- grad = tf.zeros(var.shape) # No gradients => zero.
- elif len(grad) == 1:
- grad = grad[0] # Single gradient => use as is.
- else:
- grad = tf.add_n(grad) # Multiple gradients => sum.
-
- # Scale as needed.
- scale = 1.0 / len(device.grad_raw[var]) / len(self._devices)
- scale = tf.constant(scale, dtype=tf.float32, name="scale")
- if self.minibatch_multiplier is not None:
- scale /= tf.cast(self.minibatch_multiplier, tf.float32)
- scale = self.undo_loss_scaling(scale)
- device.grad_clean[var] = grad * scale
-
- # Sum gradients across devices.
- if len(self._devices) > 1:
- with tfutil.absolute_name_scope(self.scope + "/Broadcast"), tf.device(None):
- for all_vars in zip(*[device.grad_clean.keys() for device in self._devices.values()]):
- if len(all_vars) > 0 and all(dim > 0 for dim in all_vars[0].shape.as_list()): # NCCL does not support zero-sized tensors.
- all_grads = [device.grad_clean[var] for device, var in zip(self._devices.values(), all_vars)]
- all_grads = nccl_ops.all_sum(all_grads)
- for device, var, grad in zip(self._devices.values(), all_vars, all_grads):
- device.grad_clean[var] = grad
-
- # Apply updates separately on each device.
- for device_idx, device in enumerate(self._devices.values()):
- with tfutil.absolute_name_scope(self.scope + "/Apply%d" % device_idx), tf.device(device.name):
- # pylint: disable=cell-var-from-loop
-
- # Accumulate gradients over time.
- if self.minibatch_multiplier is None:
- acc_ok = tf.constant(True, name='acc_ok')
- device.grad_acc = OrderedDict(device.grad_clean)
- else:
- # Create variables.
- with tf.control_dependencies(None):
- for var in device.grad_clean.keys():
- device.grad_acc_vars[var] = tf.Variable(tf.zeros(var.shape), trainable=False, name="grad_acc_var")
- device.grad_acc_count = tf.Variable(tf.zeros([]), trainable=False, name="grad_acc_count")
-
- # Track counter.
- count_cur = device.grad_acc_count + 1.0
- count_inc_op = lambda: tf.assign(device.grad_acc_count, count_cur)
- count_reset_op = lambda: tf.assign(device.grad_acc_count, tf.zeros([]))
- acc_ok = (count_cur >= tf.cast(self.minibatch_multiplier, tf.float32))
- all_ops.append(tf.cond(acc_ok, count_reset_op, count_inc_op))
-
- # Track gradients.
- for var, grad in device.grad_clean.items():
- acc_var = device.grad_acc_vars[var]
- acc_cur = acc_var + grad
- device.grad_acc[var] = acc_cur
- with tf.control_dependencies([acc_cur]):
- acc_inc_op = lambda: tf.assign(acc_var, acc_cur)
- acc_reset_op = lambda: tf.assign(acc_var, tf.zeros(var.shape))
- all_ops.append(tf.cond(acc_ok, acc_reset_op, acc_inc_op))
-
- # No overflow => apply gradients.
- all_ok = tf.reduce_all(tf.stack([acc_ok] + [tf.reduce_all(tf.is_finite(g)) for g in device.grad_acc.values()]))
- apply_op = lambda: device.optimizer.apply_gradients([(tf.cast(grad, var.dtype), var) for var, grad in device.grad_acc.items()])
- all_ops.append(tf.cond(all_ok, apply_op, tf.no_op))
-
- # Adjust loss scaling.
- if self.use_loss_scaling:
- ls_inc_op = lambda: tf.assign_add(device.loss_scaling_var, self.loss_scaling_inc)
- ls_dec_op = lambda: tf.assign_sub(device.loss_scaling_var, self.loss_scaling_dec)
- ls_update_op = lambda: tf.group(tf.cond(all_ok, ls_inc_op, ls_dec_op))
- all_ops.append(tf.cond(acc_ok, ls_update_op, tf.no_op))
-
- # Last device => report statistics.
- if device_idx == len(self._devices) - 1:
- all_ops.append(autosummary.autosummary(self.id + "/learning_rate", self.learning_rate))
- all_ops.append(autosummary.autosummary(self.id + "/overflow_frequency", tf.where(all_ok, 0, 1), condition=acc_ok))
- if self.use_loss_scaling:
- all_ops.append(autosummary.autosummary(self.id + "/loss_scaling_log2", device.loss_scaling_var))
-
- # Initialize variables.
- self.reset_optimizer_state()
- if self.use_loss_scaling:
- tfutil.init_uninitialized_vars([device.loss_scaling_var for device in self._devices.values()])
- if self.minibatch_multiplier is not None:
- tfutil.run([var.initializer for device in self._devices.values() for var in list(device.grad_acc_vars.values()) + [device.grad_acc_count]])
-
- # Group everything into a single op.
- with tfutil.absolute_name_scope(self.scope):
- return tf.group(*all_ops, name="TrainingOp")
-
- def reset_optimizer_state(self) -> None:
- """Reset internal state of the underlying optimizer."""
- tfutil.assert_tf_initialized()
- tfutil.run([var.initializer for device in self._devices.values() for var in device.optimizer.variables()])
-
- def get_loss_scaling_var(self, device: str) -> Union[tf.Variable, None]:
- """Get or create variable representing log2 of the current dynamic loss scaling factor."""
- return self._get_device(device).loss_scaling_var
-
- def apply_loss_scaling(self, value: TfExpression) -> TfExpression:
- """Apply dynamic loss scaling for the given expression."""
- assert tfutil.is_tf_expression(value)
- if not self.use_loss_scaling:
- return value
- return value * tfutil.exp2(self.get_loss_scaling_var(value.device))
-
- def undo_loss_scaling(self, value: TfExpression) -> TfExpression:
- """Undo the effect of dynamic loss scaling for the given expression."""
- assert tfutil.is_tf_expression(value)
- if not self.use_loss_scaling:
- return value
- return value * tfutil.exp2(-self.get_loss_scaling_var(value.device)) # pylint: disable=invalid-unary-operand-type
-
-
-class SimpleAdam:
- """Simplified version of tf.train.AdamOptimizer that behaves identically when used with dnnlib.tflib.Optimizer."""
-
- def __init__(self, name="Adam", learning_rate=0.001, beta1=0.9, beta2=0.999, epsilon=1e-8):
- self.name = name
- self.learning_rate = learning_rate
- self.beta1 = beta1
- self.beta2 = beta2
- self.epsilon = epsilon
- self.all_state_vars = []
-
- def variables(self):
- return self.all_state_vars
-
- def compute_gradients(self, loss, var_list, gate_gradients=tf.train.Optimizer.GATE_NONE):
- assert gate_gradients == tf.train.Optimizer.GATE_NONE
- return list(zip(tf.gradients(loss, var_list), var_list))
-
- def apply_gradients(self, grads_and_vars):
- with tf.name_scope(self.name):
- state_vars = []
- update_ops = []
-
- # Adjust learning rate to deal with startup bias.
- with tf.control_dependencies(None):
- b1pow_var = tf.Variable(dtype=tf.float32, initial_value=1, trainable=False)
- b2pow_var = tf.Variable(dtype=tf.float32, initial_value=1, trainable=False)
- state_vars += [b1pow_var, b2pow_var]
- b1pow_new = b1pow_var * self.beta1
- b2pow_new = b2pow_var * self.beta2
- update_ops += [tf.assign(b1pow_var, b1pow_new), tf.assign(b2pow_var, b2pow_new)]
- lr_new = self.learning_rate * tf.sqrt(1 - b2pow_new) / (1 - b1pow_new)
-
- # Construct ops to update each variable.
- for grad, var in grads_and_vars:
- with tf.control_dependencies(None):
- m_var = tf.Variable(dtype=tf.float32, initial_value=tf.zeros_like(var), trainable=False)
- v_var = tf.Variable(dtype=tf.float32, initial_value=tf.zeros_like(var), trainable=False)
- state_vars += [m_var, v_var]
- m_new = self.beta1 * m_var + (1 - self.beta1) * grad
- v_new = self.beta2 * v_var + (1 - self.beta2) * tf.square(grad)
- var_delta = lr_new * m_new / (tf.sqrt(v_new) + self.epsilon)
- update_ops += [tf.assign(m_var, m_new), tf.assign(v_var, v_new), tf.assign_sub(var, var_delta)]
-
- # Group everything together.
- self.all_state_vars += state_vars
- return tf.group(*update_ops)
diff --git a/lib/stylegan2/dnnlib/tflib/tfutil.py b/lib/stylegan2/dnnlib/tflib/tfutil.py
deleted file mode 100644
index 1127c7beecfe526b459b3b99ee34e1c431e19e1c..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/tflib/tfutil.py
+++ /dev/null
@@ -1,252 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Miscellaneous helper utils for Tensorflow."""
-
-import os
-import numpy as np
-import tensorflow as tf
-
-# Silence deprecation warnings from TensorFlow 1.13 onwards
-import logging
-logging.getLogger('tensorflow').setLevel(logging.ERROR)
-import tensorflow.contrib # requires TensorFlow 1.x!
-tf.contrib = tensorflow.contrib
-
-from typing import Any, Iterable, List, Union
-
-TfExpression = Union[tf.Tensor, tf.Variable, tf.Operation]
-"""A type that represents a valid Tensorflow expression."""
-
-TfExpressionEx = Union[TfExpression, int, float, np.ndarray]
-"""A type that can be converted to a valid Tensorflow expression."""
-
-
-def run(*args, **kwargs) -> Any:
- """Run the specified ops in the default session."""
- assert_tf_initialized()
- return tf.get_default_session().run(*args, **kwargs)
-
-
-def is_tf_expression(x: Any) -> bool:
- """Check whether the input is a valid Tensorflow expression, i.e., Tensorflow Tensor, Variable, or Operation."""
- return isinstance(x, (tf.Tensor, tf.Variable, tf.Operation))
-
-
-def shape_to_list(shape: Iterable[tf.Dimension]) -> List[Union[int, None]]:
- """Convert a Tensorflow shape to a list of ints. Retained for backwards compatibility -- use TensorShape.as_list() in new code."""
- return [dim.value for dim in shape]
-
-
-def flatten(x: TfExpressionEx) -> TfExpression:
- """Shortcut function for flattening a tensor."""
- with tf.name_scope("Flatten"):
- return tf.reshape(x, [-1])
-
-
-def log2(x: TfExpressionEx) -> TfExpression:
- """Logarithm in base 2."""
- with tf.name_scope("Log2"):
- return tf.log(x) * np.float32(1.0 / np.log(2.0))
-
-
-def exp2(x: TfExpressionEx) -> TfExpression:
- """Exponent in base 2."""
- with tf.name_scope("Exp2"):
- return tf.exp(x * np.float32(np.log(2.0)))
-
-
-def lerp(a: TfExpressionEx, b: TfExpressionEx, t: TfExpressionEx) -> TfExpressionEx:
- """Linear interpolation."""
- with tf.name_scope("Lerp"):
- return a + (b - a) * t
-
-
-def lerp_clip(a: TfExpressionEx, b: TfExpressionEx, t: TfExpressionEx) -> TfExpression:
- """Linear interpolation with clip."""
- with tf.name_scope("LerpClip"):
- return a + (b - a) * tf.clip_by_value(t, 0.0, 1.0)
-
-
-def absolute_name_scope(scope: str) -> tf.name_scope:
- """Forcefully enter the specified name scope, ignoring any surrounding scopes."""
- return tf.name_scope(scope + "/")
-
-
-def absolute_variable_scope(scope: str, **kwargs) -> tf.variable_scope:
- """Forcefully enter the specified variable scope, ignoring any surrounding scopes."""
- return tf.variable_scope(tf.VariableScope(name=scope, **kwargs), auxiliary_name_scope=False)
-
-
-def _sanitize_tf_config(config_dict: dict = None) -> dict:
- # Defaults.
- cfg = dict()
- cfg["rnd.np_random_seed"] = None # Random seed for NumPy. None = keep as is.
- cfg["rnd.tf_random_seed"] = "auto" # Random seed for TensorFlow. 'auto' = derive from NumPy random state. None = keep as is.
- cfg["env.TF_CPP_MIN_LOG_LEVEL"] = "1" # 0 = Print all available debug info from TensorFlow. 1 = Print warnings and errors, but disable debug info.
- cfg["graph_options.place_pruned_graph"] = True # False = Check that all ops are available on the designated device. True = Skip the check for ops that are not used.
- cfg["gpu_options.allow_growth"] = True # False = Allocate all GPU memory at the beginning. True = Allocate only as much GPU memory as needed.
-
- # Remove defaults for environment variables that are already set.
- for key in list(cfg):
- fields = key.split(".")
- if fields[0] == "env":
- assert len(fields) == 2
- if fields[1] in os.environ:
- del cfg[key]
-
- # User overrides.
- if config_dict is not None:
- cfg.update(config_dict)
- return cfg
-
-
-def init_tf(config_dict: dict = None) -> None:
- """Initialize TensorFlow session using good default settings."""
- # Skip if already initialized.
- if tf.get_default_session() is not None:
- return
-
- # Setup config dict and random seeds.
- cfg = _sanitize_tf_config(config_dict)
- np_random_seed = cfg["rnd.np_random_seed"]
- if np_random_seed is not None:
- np.random.seed(np_random_seed)
- tf_random_seed = cfg["rnd.tf_random_seed"]
- if tf_random_seed == "auto":
- tf_random_seed = np.random.randint(1 << 31)
- if tf_random_seed is not None:
- tf.set_random_seed(tf_random_seed)
-
- # Setup environment variables.
- for key, value in cfg.items():
- fields = key.split(".")
- if fields[0] == "env":
- assert len(fields) == 2
- os.environ[fields[1]] = str(value)
-
- # Create default TensorFlow session.
- create_session(cfg, force_as_default=True)
-
-
-def assert_tf_initialized():
- """Check that TensorFlow session has been initialized."""
- if tf.get_default_session() is None:
- raise RuntimeError("No default TensorFlow session found. Please call dnnlib.tflib.init_tf().")
-
-
-def create_session(config_dict: dict = None, force_as_default: bool = False) -> tf.Session:
- """Create tf.Session based on config dict."""
- # Setup TensorFlow config proto.
- cfg = _sanitize_tf_config(config_dict)
- config_proto = tf.ConfigProto()
- for key, value in cfg.items():
- fields = key.split(".")
- if fields[0] not in ["rnd", "env"]:
- obj = config_proto
- for field in fields[:-1]:
- obj = getattr(obj, field)
- setattr(obj, fields[-1], value)
-
- # Create session.
- session = tf.Session(config=config_proto)
- if force_as_default:
- # pylint: disable=protected-access
- session._default_session = session.as_default()
- session._default_session.enforce_nesting = False
- session._default_session.__enter__()
- return session
-
-
-def init_uninitialized_vars(target_vars: List[tf.Variable] = None) -> None:
- """Initialize all tf.Variables that have not already been initialized.
-
- Equivalent to the following, but more efficient and does not bloat the tf graph:
- tf.variables_initializer(tf.report_uninitialized_variables()).run()
- """
- assert_tf_initialized()
- if target_vars is None:
- target_vars = tf.global_variables()
-
- test_vars = []
- test_ops = []
-
- with tf.control_dependencies(None): # ignore surrounding control_dependencies
- for var in target_vars:
- assert is_tf_expression(var)
-
- try:
- tf.get_default_graph().get_tensor_by_name(var.name.replace(":0", "/IsVariableInitialized:0"))
- except KeyError:
- # Op does not exist => variable may be uninitialized.
- test_vars.append(var)
-
- with absolute_name_scope(var.name.split(":")[0]):
- test_ops.append(tf.is_variable_initialized(var))
-
- init_vars = [var for var, inited in zip(test_vars, run(test_ops)) if not inited]
- run([var.initializer for var in init_vars])
-
-
-def set_vars(var_to_value_dict: dict) -> None:
- """Set the values of given tf.Variables.
-
- Equivalent to the following, but more efficient and does not bloat the tf graph:
- tflib.run([tf.assign(var, value) for var, value in var_to_value_dict.items()]
- """
- assert_tf_initialized()
- ops = []
- feed_dict = {}
-
- for var, value in var_to_value_dict.items():
- assert is_tf_expression(var)
-
- try:
- setter = tf.get_default_graph().get_tensor_by_name(var.name.replace(":0", "/setter:0")) # look for existing op
- except KeyError:
- with absolute_name_scope(var.name.split(":")[0]):
- with tf.control_dependencies(None): # ignore surrounding control_dependencies
- setter = tf.assign(var, tf.placeholder(var.dtype, var.shape, "new_value"), name="setter") # create new setter
-
- ops.append(setter)
- feed_dict[setter.op.inputs[1]] = value
-
- run(ops, feed_dict)
-
-
-def create_var_with_large_initial_value(initial_value: np.ndarray, *args, **kwargs):
- """Create tf.Variable with large initial value without bloating the tf graph."""
- assert_tf_initialized()
- assert isinstance(initial_value, np.ndarray)
- zeros = tf.zeros(initial_value.shape, initial_value.dtype)
- var = tf.Variable(zeros, *args, **kwargs)
- set_vars({var: initial_value})
- return var
-
-
-def convert_images_from_uint8(images, drange=[-1,1], nhwc_to_nchw=False):
- """Convert a minibatch of images from uint8 to float32 with configurable dynamic range.
- Can be used as an input transformation for Network.run().
- """
- images = tf.cast(images, tf.float32)
- if nhwc_to_nchw:
- images = tf.transpose(images, [0, 3, 1, 2])
- return images * ((drange[1] - drange[0]) / 255) + drange[0]
-
-
-def convert_images_to_uint8(images, drange=[-1,1], nchw_to_nhwc=False, shrink=1):
- """Convert a minibatch of images from float32 to uint8 with configurable dynamic range.
- Can be used as an output transformation for Network.run().
- """
- images = tf.cast(images, tf.float32)
- if shrink > 1:
- ksize = [1, 1, shrink, shrink]
- images = tf.nn.avg_pool(images, ksize=ksize, strides=ksize, padding="VALID", data_format="NCHW")
- if nchw_to_nhwc:
- images = tf.transpose(images, [0, 2, 3, 1])
- scale = 255 / (drange[1] - drange[0])
- images = images * scale + (0.5 - drange[0] * scale)
- return tf.saturate_cast(images, tf.uint8)
diff --git a/lib/stylegan2/dnnlib/util.py b/lib/stylegan2/dnnlib/util.py
deleted file mode 100644
index 73c98d73c26e03800a39c62386c322bb518bbd18..0000000000000000000000000000000000000000
--- a/lib/stylegan2/dnnlib/util.py
+++ /dev/null
@@ -1,410 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Miscellaneous utility classes and functions."""
-
-import ctypes
-import fnmatch
-import importlib
-import inspect
-import numpy as np
-import os
-import shutil
-import sys
-import types
-import io
-import pickle
-import re
-import requests
-import html
-import hashlib
-import glob
-import uuid
-
-from distutils.util import strtobool
-from typing import Any, List, Tuple, Union
-
-
-# Util classes
-# ------------------------------------------------------------------------------------------
-
-
-class EasyDict(dict):
- """Convenience class that behaves like a dict but allows access with the attribute syntax."""
-
- def __getattr__(self, name: str) -> Any:
- try:
- return self[name]
- except KeyError:
- raise AttributeError(name)
-
- def __setattr__(self, name: str, value: Any) -> None:
- self[name] = value
-
- def __delattr__(self, name: str) -> None:
- del self[name]
-
-
-class Logger(object):
- """Redirect stderr to stdout, optionally print stdout to a file, and optionally force flushing on both stdout and the file."""
-
- def __init__(self, file_name: str = None, file_mode: str = "w", should_flush: bool = True):
- self.file = None
-
- if file_name is not None:
- self.file = open(file_name, file_mode)
-
- self.should_flush = should_flush
- self.stdout = sys.stdout
- self.stderr = sys.stderr
-
- sys.stdout = self
- sys.stderr = self
-
- def __enter__(self) -> "Logger":
- return self
-
- def __exit__(self, exc_type: Any, exc_value: Any, traceback: Any) -> None:
- self.close()
-
- def write(self, text: str) -> None:
- """Write text to stdout (and a file) and optionally flush."""
- if len(text) == 0: # workaround for a bug in VSCode debugger: sys.stdout.write(''); sys.stdout.flush() => crash
- return
-
- if self.file is not None:
- self.file.write(text)
-
- self.stdout.write(text)
-
- if self.should_flush:
- self.flush()
-
- def flush(self) -> None:
- """Flush written text to both stdout and a file, if open."""
- if self.file is not None:
- self.file.flush()
-
- self.stdout.flush()
-
- def close(self) -> None:
- """Flush, close possible files, and remove stdout/stderr mirroring."""
- self.flush()
-
- # if using multiple loggers, prevent closing in wrong order
- if sys.stdout is self:
- sys.stdout = self.stdout
- if sys.stderr is self:
- sys.stderr = self.stderr
-
- if self.file is not None:
- self.file.close()
-
-
-# Small util functions
-# ------------------------------------------------------------------------------------------
-
-
-def format_time(seconds: Union[int, float]) -> str:
- """Convert the seconds to human readable string with days, hours, minutes and seconds."""
- s = int(np.rint(seconds))
-
- if s < 60:
- return "{0}s".format(s)
- elif s < 60 * 60:
- return "{0}m {1:02}s".format(s // 60, s % 60)
- elif s < 24 * 60 * 60:
- return "{0}h {1:02}m {2:02}s".format(s // (60 * 60), (s // 60) % 60, s % 60)
- else:
- return "{0}d {1:02}h {2:02}m".format(s // (24 * 60 * 60), (s // (60 * 60)) % 24, (s // 60) % 60)
-
-
-def ask_yes_no(question: str) -> bool:
- """Ask the user the question until the user inputs a valid answer."""
- while True:
- try:
- print("{0} [y/n]".format(question))
- return strtobool(input().lower())
- except ValueError:
- pass
-
-
-def tuple_product(t: Tuple) -> Any:
- """Calculate the product of the tuple elements."""
- result = 1
-
- for v in t:
- result *= v
-
- return result
-
-
-_str_to_ctype = {
- "uint8": ctypes.c_ubyte,
- "uint16": ctypes.c_uint16,
- "uint32": ctypes.c_uint32,
- "uint64": ctypes.c_uint64,
- "int8": ctypes.c_byte,
- "int16": ctypes.c_int16,
- "int32": ctypes.c_int32,
- "int64": ctypes.c_int64,
- "float32": ctypes.c_float,
- "float64": ctypes.c_double
-}
-
-
-def get_dtype_and_ctype(type_obj: Any) -> Tuple[np.dtype, Any]:
- """Given a type name string (or an object having a __name__ attribute), return matching Numpy and ctypes types that have the same size in bytes."""
- type_str = None
-
- if isinstance(type_obj, str):
- type_str = type_obj
- elif hasattr(type_obj, "__name__"):
- type_str = type_obj.__name__
- elif hasattr(type_obj, "name"):
- type_str = type_obj.name
- else:
- raise RuntimeError("Cannot infer type name from input")
-
- assert type_str in _str_to_ctype.keys()
-
- my_dtype = np.dtype(type_str)
- my_ctype = _str_to_ctype[type_str]
-
- assert my_dtype.itemsize == ctypes.sizeof(my_ctype)
-
- return my_dtype, my_ctype
-
-
-def is_pickleable(obj: Any) -> bool:
- try:
- with io.BytesIO() as stream:
- pickle.dump(obj, stream)
- return True
- except:
- return False
-
-
-# Functionality to import modules/objects by name, and call functions by name
-# ------------------------------------------------------------------------------------------
-
-def get_module_from_obj_name(obj_name: str) -> Tuple[types.ModuleType, str]:
- """Searches for the underlying module behind the name to some python object.
- Returns the module and the object name (original name with module part removed)."""
-
- # allow convenience shorthands, substitute them by full names
- obj_name = re.sub("^np.", "numpy.", obj_name)
- obj_name = re.sub("^tf.", "tensorflow.", obj_name)
-
- # list alternatives for (module_name, local_obj_name)
- parts = obj_name.split(".")
- name_pairs = [(".".join(parts[:i]), ".".join(parts[i:])) for i in range(len(parts), 0, -1)]
-
- # try each alternative in turn
- for module_name, local_obj_name in name_pairs:
- try:
- module = importlib.import_module(module_name) # may raise ImportError
- get_obj_from_module(module, local_obj_name) # may raise AttributeError
- return module, local_obj_name
- except:
- pass
-
- # maybe some of the modules themselves contain errors?
- for module_name, _local_obj_name in name_pairs:
- try:
- importlib.import_module(module_name) # may raise ImportError
- except ImportError:
- if not str(sys.exc_info()[1]).startswith("No module named '" + module_name + "'"):
- raise
-
- # maybe the requested attribute is missing?
- for module_name, local_obj_name in name_pairs:
- try:
- module = importlib.import_module(module_name) # may raise ImportError
- get_obj_from_module(module, local_obj_name) # may raise AttributeError
- except ImportError:
- pass
-
- # we are out of luck, but we have no idea why
- raise ImportError(obj_name)
-
-
-def get_obj_from_module(module: types.ModuleType, obj_name: str) -> Any:
- """Traverses the object name and returns the last (rightmost) python object."""
- if obj_name == '':
- return module
- obj = module
- for part in obj_name.split("."):
- obj = getattr(obj, part)
- return obj
-
-
-def get_obj_by_name(name: str) -> Any:
- """Finds the python object with the given name."""
- module, obj_name = get_module_from_obj_name(name)
- return get_obj_from_module(module, obj_name)
-
-
-def call_func_by_name(*args, func_name: str = None, **kwargs) -> Any:
- """Finds the python object with the given name and calls it as a function."""
- assert func_name is not None
- func_obj = get_obj_by_name(func_name)
- assert callable(func_obj)
- return func_obj(*args, **kwargs)
-
-
-def get_module_dir_by_obj_name(obj_name: str) -> str:
- """Get the directory path of the module containing the given object name."""
- module, _ = get_module_from_obj_name(obj_name)
- return os.path.dirname(inspect.getfile(module))
-
-
-def is_top_level_function(obj: Any) -> bool:
- """Determine whether the given object is a top-level function, i.e., defined at module scope using 'def'."""
- return callable(obj) and obj.__name__ in sys.modules[obj.__module__].__dict__
-
-
-def get_top_level_function_name(obj: Any) -> str:
- """Return the fully-qualified name of a top-level function."""
- assert is_top_level_function(obj)
- return obj.__module__ + "." + obj.__name__
-
-
-# File system helpers
-# ------------------------------------------------------------------------------------------
-
-def list_dir_recursively_with_ignore(dir_path: str, ignores: List[str] = None, add_base_to_relative: bool = False) -> List[Tuple[str, str]]:
- """List all files recursively in a given directory while ignoring given file and directory names.
- Returns list of tuples containing both absolute and relative paths."""
- assert os.path.isdir(dir_path)
- base_name = os.path.basename(os.path.normpath(dir_path))
-
- if ignores is None:
- ignores = []
-
- result = []
-
- for root, dirs, files in os.walk(dir_path, topdown=True):
- for ignore_ in ignores:
- dirs_to_remove = [d for d in dirs if fnmatch.fnmatch(d, ignore_)]
-
- # dirs need to be edited in-place
- for d in dirs_to_remove:
- dirs.remove(d)
-
- files = [f for f in files if not fnmatch.fnmatch(f, ignore_)]
-
- absolute_paths = [os.path.join(root, f) for f in files]
- relative_paths = [os.path.relpath(p, dir_path) for p in absolute_paths]
-
- if add_base_to_relative:
- relative_paths = [os.path.join(base_name, p) for p in relative_paths]
-
- assert len(absolute_paths) == len(relative_paths)
- result += zip(absolute_paths, relative_paths)
-
- return result
-
-
-def copy_files_and_create_dirs(files: List[Tuple[str, str]]) -> None:
- """Takes in a list of tuples of (src, dst) paths and copies files.
- Will create all necessary directories."""
- for file in files:
- target_dir_name = os.path.dirname(file[1])
-
- # will create all intermediate-level directories
- if not os.path.exists(target_dir_name):
- os.makedirs(target_dir_name)
-
- shutil.copyfile(file[0], file[1])
-
-
-# URL helpers
-# ------------------------------------------------------------------------------------------
-
-def is_url(obj: Any, allow_file_urls: bool = False) -> bool:
- """Determine whether the given object is a valid URL string."""
- if not isinstance(obj, str) or not "://" in obj:
- return False
- if allow_file_urls and obj.startswith('file:///'):
- return True
- try:
- res = requests.compat.urlparse(obj)
- if not res.scheme or not res.netloc or not "." in res.netloc:
- return False
- res = requests.compat.urlparse(requests.compat.urljoin(obj, "/"))
- if not res.scheme or not res.netloc or not "." in res.netloc:
- return False
- except:
- return False
- return True
-
-
-def open_url(url: str, cache_dir: str = None, num_attempts: int = 10, verbose: bool = True) -> Any:
- """Download the given URL and return a binary-mode file object to access the data."""
- assert is_url(url, allow_file_urls=True)
- assert num_attempts >= 1
-
- # Handle file URLs.
- if url.startswith('file:///'):
- return open(url[len('file:///'):], "rb")
-
- # Lookup from cache.
- url_md5 = hashlib.md5(url.encode("utf-8")).hexdigest()
- if cache_dir is not None:
- cache_files = glob.glob(os.path.join(cache_dir, url_md5 + "_*"))
- if len(cache_files) == 1:
- return open(cache_files[0], "rb")
-
- # Download.
- url_name = None
- url_data = None
- with requests.Session() as session:
- if verbose:
- print("Downloading %s ..." % url, end="", flush=True)
- for attempts_left in reversed(range(num_attempts)):
- try:
- with session.get(url) as res:
- res.raise_for_status()
- if len(res.content) == 0:
- raise IOError("No data received")
-
- if len(res.content) < 8192:
- content_str = res.content.decode("utf-8")
- if "download_warning" in res.headers.get("Set-Cookie", ""):
- links = [html.unescape(link) for link in content_str.split('"') if "export=download" in link]
- if len(links) == 1:
- url = requests.compat.urljoin(url, links[0])
- raise IOError("Google Drive virus checker nag")
- if "Google Drive - Quota exceeded" in content_str:
- raise IOError("Google Drive download quota exceeded -- please try again later")
-
- match = re.search(r'filename="([^"]*)"', res.headers.get("Content-Disposition", ""))
- url_name = match[1] if match else url
- url_data = res.content
- if verbose:
- print(" done")
- break
- except:
- if not attempts_left:
- if verbose:
- print(" failed")
- raise
- if verbose:
- print(".", end="", flush=True)
-
- # Save to cache.
- if cache_dir is not None:
- safe_name = re.sub(r"[^0-9a-zA-Z-._]", "_", url_name)
- cache_file = os.path.join(cache_dir, url_md5 + "_" + safe_name)
- temp_file = os.path.join(cache_dir, "tmp_" + uuid.uuid4().hex + "_" + url_md5 + "_" + safe_name)
- os.makedirs(cache_dir, exist_ok=True)
- with open(temp_file, "wb") as f:
- f.write(url_data)
- os.replace(temp_file, cache_file) # atomic
-
- # Return data as file object.
- return io.BytesIO(url_data)
diff --git a/lib/stylegan2/generator.py b/lib/stylegan2/generator.py
deleted file mode 100644
index 31b2877b9e2f2a50329572589a13cbd747a50eaf..0000000000000000000000000000000000000000
--- a/lib/stylegan2/generator.py
+++ /dev/null
@@ -1,170 +0,0 @@
-from . import dnnlib, pretrained_networks
-from .dnnlib import tflib
-from .. import utils
-from bob.extension import rc
-import numpy as np
-import bob.io.image
-import tensorflow as tf
-
-class StyleGAN2Generator(object):
- def __init__(self,
- sg2_path=rc['sg2_morph.sg2_path'],
- randomize_noise=False,
- truncation_psi=0.5,
- batch_size=None):
- """
- Instanciate the StyleGAN2 generator network. Cf.
- T. Karras, *et al.*, \"Analyzing and improving the quality of StyleGAN\", in *Proc.
- IEEE/CVF Conference on Computer Vision and Patter Recognition,
- 2020.*
- Repository : https://github.com/NVlabs/stylegan2
-
- :param sg2_path: Path to the pickled pretrained network
- :param randomize_noise: Whether to randomly sample the noise inputs of the synthetizer, or fix them once
- and for all at initialization.
- :param image_postprocessing_fn : Additional function to apply to the generated images before saving.
- """
- _G, _D, Gs = pretrained_networks.load_networks(sg2_path)
- self.network = Gs
- self.latent_dim = self.network.input_shape[-1]
- self.run_kwargs = {'randomize_noise': randomize_noise,
- 'minibatch_size': batch_size}
-
- if not randomize_noise:
- noise_vars = [var for name, var in self.network.components.synthesis.vars.items() if name.startswith('noise')]
- tflib.set_vars({var: np.random.randn(*var.shape.as_list()) for var in noise_vars}) # [height, width]
-
-
- def run_from_Z(self, latents, truncation_psi=0.5, return_w_latents=False, **kwargs):
- """
- Run the generator from the input latent space Z
-
- Inputs:
- :param latents: batch of latent vectors, of the shape [batch_size, latent_dim].
- :param truncation_psi: (float in [0,1]) value of psi when applying the truncation trick (cf. T. Karras, *et al.*, \"A style-based
- generator architecture for generative adversarial networks\", in *Proc. CVPR*, 2018).
- A value of 0 applies complete truncation (meaning one can only generate the mean face),
- while a value of 1 applies no truncation.
- :param return_w_latents: whether to return the W-space latents as additional output
- :param **kwargs: other parameters that will be feeded to the Network.run method
-
- Outputs:
-
- :return images: Batch of generated images in bob format : tensor with shape [batch_size, 3, 1024, 1024] of uint8
- values in the range [0, 255]
- :return w_latents: Batch of W-space latents vector, of the shape [batch_size, latent_dim].
- Only returned if `return_w_latents` is True.
- """
- run_kwargs = self.run_kwargs
- run_kwargs['truncation_psi'] = truncation_psi
- run_kwargs['return_dlatents'] = return_w_latents
- run_kwargs.update(kwargs)
-
- result = self.network.run(latents, None, **run_kwargs)
- if return_w_latents:
- images, dlatents = result
- return self._fix_range(images), dlatents[:, 0, :]
- else:
- images = result
- return self._fix_range(images)
-
- def run_from_W(self, w_latents, **kwargs):
- """
- Run the generator from the intermediate latent space W
-
- Inputs:
- :param latents: batch of W-space latent vectors, of the shape [batch_size, latent_dim].
- :param **kwargs: other parameters that will be feeded to the Network.run method of the synthesis network
-
- Outputs:
-
- :return images: Batch of generated images in bob format : tensor with shape [batch_size, 3, 1024, 1024] of uint8
- values in the range [0, 255]
- """
- # Repeat the input latent for each style input
- dlatents = np.tile(w_latents[:, np.newaxis, :], [1, 18, 1])
- run_kwargs = self.run_kwargs
- run_kwargs.update(kwargs)
- images = self.network.components['synthesis'].run(dlatents, **run_kwargs)
- return self._fix_range(images)
-
- def run_from_Wplus(self, w_latents, **kwargs):
- """
- Run the generator from the intermediate latent space W
-
- Inputs:
- :param latents: batch of W+-space latent vectors, of the shape [batch_size, 18, latent_dim].
- :param **kwargs: other parameters that will be feeded to the Network.run method of the synthesis network
-
- Outputs:
-
- :return images: Batch of generated images in bob format : tensor with shape [batch_size, 3, 1024, 1024] of uint8
- values in the range [0, 255]
- """
- # Repeat the input latent for each style input
- run_kwargs = self.run_kwargs
- run_kwargs.update(kwargs)
- images = self.network.components['synthesis'].run(w_latents, **run_kwargs)
- return self._fix_range(images)
-
- def _fix_range(self, images):
- return np.stack([utils.adjust_dynamic_range(img, [0,255], 'uint8') for img in images])
-
- def run_with_style_mixing_augmentation(self, w_latents, num_variations, style_mixing_depth, return_w_latents=False, **kwargs):
- z_augmentations = np.random.randn(w_latents.shape[0] * num_variations, self.latent_dim)
- _, w_augmentations = self.run_from_Z(z_augmentations, return_w_latents=True, **kwargs)
-
- repeated_w_latents = np.tile(w_latents, [num_variations, 1])
-
- dlatents = np.concatenate(
- [
- np.tile(repeated_w_latents[:, None, :], [1, style_mixing_depth, 1]),
- np.tile(w_augmentations[:, None, :], [1, 18 - style_mixing_depth, 1])
- ],
- axis=1
- )
- if return_w_latents:
- return self.run_from_Wplus(dlatents, **kwargs), repeated_w_latents, w_augmentations
- else:
- return self.run_from_Wplus(dlatents, **kwargs)
-
-
-
-"""
-def output_transform(images):
- return tflib.convert_images_to_uint8(images, nchw_to_nhwc=False)
-
-def w_output_transform(images, w_latent):
- return [tflib.convert_images_to_uint8(images, nchw_to_nhwc=False), w_latent]
-
-class StyleGAN2Generator(object):
- def __init__(self,
- network_path="/idiap/temp/lcolbois/networks/stylegan2-ffhq-config-f.pkl",
- Gs_kwargs={},
- image_postprocessing_fn=None,
- return_w_latent=True):
-
- self.return_w_latent = return_w_latent
- self.image_postprocessing_fn=image_postprocessing_fn
- if self.image_postprocessing_fn is None:
- self.image_postprocessing_fn = (lambda x: x)
-
- _G, _D, Gs = pretrained_networks.load_networks(network_path)
- self.Gs_kwargs = Gs_kwargs
- if self.return_w_latent:
- self.Gs = Gs.clone(return_dlatents=True)
- self.Gs_kwargs.update({'return_dlatents': True, 'output_transform': dict(func=w_output_transform)})
- else:
- self.Gs= Gs
- self.Gs_kwargs.update({'output_transform': dict(func=output_transform)})
-
- self.latent_dim = self.Gs.input_shape[1]
-
- def __call__(self, z_latent):
- if self.return_w_latent:
- images, w_latent = self.Gs.run(np.stack([z_latent]), None, **self.Gs_kwargs)
- return self.image_postprocessing_fn(images[0]), w_latent[0][0]
- else:
- images = self.Gs.run(np.stack([z_latent]), None, **self.Gs_kwargs)
- return self.image_postprocessing_fn(images[0])
-"""
\ No newline at end of file
diff --git a/lib/stylegan2/preprocessor.py b/lib/stylegan2/preprocessor.py
deleted file mode 100644
index fc22323330b886ca1e3d8159244ac17792dcf971..0000000000000000000000000000000000000000
--- a/lib/stylegan2/preprocessor.py
+++ /dev/null
@@ -1,126 +0,0 @@
-import dlib
-import PIL.Image
-import bob.io.image
-import scipy.ndimage
-import numpy as np
-from bob.extension import rc
-
-class FFHQCropper(object):
- def __init__(self, dlib_lmd_path=rc['sg2_morph.dlib_lmd_path']):
- """
- Instanciate a face cropper that behaves similarly to the one used to preprocess the FFHQ database.
-
- :param dlib_lmd_path: Path to the dlib landmark detector model
- """
- self.detector = dlib.get_frontal_face_detector()
- self.predictor = dlib.shape_predictor(dlib_lmd_path)
-
- def __call__(self, image):
- """
- Run the cropper on the input image.
-
- :param image: input image in bob format (channels first)
- :return : cropped image in bob format, with shape [3, 1024, 1024]
- """
- # Assuming input image in bob format
- channels_last_img = bob.io.image.to_matplotlib(image)
- landmarks = self.detect_landmarks(channels_last_img)
- cropped = self.crop_and_resize(channels_last_img, lm=landmarks)
- return bob.io.image.to_bob(cropped)
-
- def detect_landmarks(self, image):
- """
- Run the dlib landmark detector on the input image to build a list of landmarks.
-
- :param image: input image in matplotlib format (channels last)
- :return: list of 68 landmarks in tuple format : [(x1, y1), (x2, y2), ...]
- """
- detection = self.detector(image, 1)[0]
- shape = self.predictor(image, detection)
- return [(item.x, item.y) for item in shape.parts()]
-
- def crop_and_resize(self, img, lm,
- output_size=1024,
- transform_size=4096,
- enable_padding=True):
- """
- Crop, resize and align the image based on the provided landmarks (lm),
- in the same way FFHQ has been preprocessed for training StyleGAN2
-
- This code was entirely borrowed from
- https://github.com/NVlabs/ffhq-dataset/blob/master/download_ffhq.py (recreate_aligned_images() function),
- with a few adaptations.
-
- """
- lm = np.array(lm)
- lm_chin = lm[0 : 17] # left-right
- lm_eyebrow_left = lm[17 : 22] # left-right
- lm_eyebrow_right = lm[22 : 27] # left-right
- lm_nose = lm[27 : 31] # top-down
- lm_nostrils = lm[31 : 36] # top-down
- lm_eye_left = lm[36 : 42] # left-clockwise
- lm_eye_right = lm[42 : 48] # left-clockwise
- lm_mouth_outer = lm[48 : 60] # left-clockwise
- lm_mouth_inner = lm[60 : 68] # left-clockwise
-
- # Calculate auxiliary vectors.
- eye_left = np.mean(lm_eye_left, axis=0)
- eye_right = np.mean(lm_eye_right, axis=0)
- eye_avg = (eye_left + eye_right) * 0.5
- eye_to_eye = eye_right - eye_left
- mouth_left = lm_mouth_outer[0]
- mouth_right = lm_mouth_outer[6]
- mouth_avg = (mouth_left + mouth_right) * 0.5
- eye_to_mouth = mouth_avg - eye_avg
-
- # Choose oriented crop rectangle.
- x = eye_to_eye - np.flipud(eye_to_mouth) * [-1, 1]
- x /= np.hypot(*x)
- x *= max(np.hypot(*eye_to_eye) * 2.0, np.hypot(*eye_to_mouth) * 1.8)
- y = np.flipud(x) * [-1, 1]
- c = eye_avg + eye_to_mouth * 0.1
- quad = np.stack([c - x - y, c - x + y, c + x + y, c + x - y])
- qsize = np.hypot(*x) * 2
-
- # Convert to PIL (original code used PIL all the way through)
- img = PIL.Image.fromarray(img)
-
- # Shrink.
- shrink = int(np.floor(qsize / output_size * 0.5))
- if shrink > 1:
- rsize = (int(np.rint(float(img.size[0]) / shrink)), int(np.rint(float(img.size[1]) / shrink)))
- img = img.resize(rsize, PIL.Image.ANTIALIAS)
- quad /= shrink
- qsize /= shrink
-
- # Crop.
- border = max(int(np.rint(qsize * 0.1)), 3)
- crop = (int(np.floor(min(quad[:,0]))), int(np.floor(min(quad[:,1]))), int(np.ceil(max(quad[:,0]))), int(np.ceil(max(quad[:,1]))))
- crop = (max(crop[0] - border, 0), max(crop[1] - border, 0), min(crop[2] + border, img.size[0]), min(crop[3] + border, img.size[1]))
- if crop[2] - crop[0] < img.size[0] or crop[3] - crop[1] < img.size[1]:
- img = img.crop(crop)
- quad -= crop[0:2]
-
- # Pad.
- pad = (int(np.floor(min(quad[:,0]))), int(np.floor(min(quad[:,1]))), int(np.ceil(max(quad[:,0]))), int(np.ceil(max(quad[:,1]))))
- pad = (max(-pad[0] + border, 0), max(-pad[1] + border, 0), max(pad[2] - img.size[0] + border, 0), max(pad[3] - img.size[1] + border, 0))
- if enable_padding and max(pad) > border - 4:
- pad = np.maximum(pad, int(np.rint(qsize * 0.3)))
- img = np.pad(np.float32(img), ((pad[1], pad[3]), (pad[0], pad[2]), (0, 0)), 'reflect')
- h, w, _ = img.shape
- y, x, _ = np.ogrid[:h, :w, :1]
- mask = np.maximum(1.0 - np.minimum(np.float32(x) / pad[0], np.float32(w-1-x) / pad[2]), 1.0 - np.minimum(np.float32(y) / pad[1], np.float32(h-1-y) / pad[3]))
- blur = qsize * 0.02
- img += (scipy.ndimage.gaussian_filter(img, [blur, blur, 0]) - img) * np.clip(mask * 3.0 + 1.0, 0.0, 1.0)
- img += (np.median(img, axis=(0,1)) - img) * np.clip(mask, 0.0, 1.0)
- img = PIL.Image.fromarray(np.uint8(np.clip(np.rint(img), 0, 255)), 'RGB')
- quad += pad[:2]
-
- # Transform.
- img = img.transform((transform_size, transform_size), PIL.Image.QUAD, (quad + 0.5).flatten(), PIL.Image.BILINEAR)
- if output_size < transform_size:
- img = img.resize((output_size, output_size), PIL.Image.ANTIALIAS)
-
- # Done ! Back to numpy format
- img = np.array(img)
- return img
\ No newline at end of file
diff --git a/lib/stylegan2/pretrained_networks.py b/lib/stylegan2/pretrained_networks.py
deleted file mode 100644
index 9ad3e5e78641638e38dc613110176fc1bf1d8ca4..0000000000000000000000000000000000000000
--- a/lib/stylegan2/pretrained_networks.py
+++ /dev/null
@@ -1,83 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""List of pre-trained StyleGAN2 networks located on Google Drive."""
-
-import pickle
-import sys
-import os
-sys.path.append('/idiap/home/lcolbois/colbois_reporting/synthface/synthface/stylegan2')
-from . import dnnlib
-from dnnlib import tflib
-
-#----------------------------------------------------------------------------
-# StyleGAN2 Google Drive root: https://drive.google.com/open?id=1QHc-yF5C3DChRwSdZKcx1w6K8JvSxQi7
-
-gdrive_urls = {
- 'gdrive:networks/stylegan2-car-config-a.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-car-config-a.pkl',
- 'gdrive:networks/stylegan2-car-config-b.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-car-config-b.pkl',
- 'gdrive:networks/stylegan2-car-config-c.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-car-config-c.pkl',
- 'gdrive:networks/stylegan2-car-config-d.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-car-config-d.pkl',
- 'gdrive:networks/stylegan2-car-config-e.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-car-config-e.pkl',
- 'gdrive:networks/stylegan2-car-config-f.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-car-config-f.pkl',
- 'gdrive:networks/stylegan2-cat-config-a.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-cat-config-a.pkl',
- 'gdrive:networks/stylegan2-cat-config-f.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-cat-config-f.pkl',
- 'gdrive:networks/stylegan2-church-config-a.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-church-config-a.pkl',
- 'gdrive:networks/stylegan2-church-config-f.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-church-config-f.pkl',
- 'gdrive:networks/stylegan2-ffhq-config-a.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-ffhq-config-a.pkl',
- 'gdrive:networks/stylegan2-ffhq-config-b.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-ffhq-config-b.pkl',
- 'gdrive:networks/stylegan2-ffhq-config-c.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-ffhq-config-c.pkl',
- 'gdrive:networks/stylegan2-ffhq-config-d.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-ffhq-config-d.pkl',
- 'gdrive:networks/stylegan2-ffhq-config-e.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-ffhq-config-e.pkl',
- 'gdrive:networks/stylegan2-ffhq-config-f.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-ffhq-config-f.pkl',
- 'gdrive:networks/stylegan2-horse-config-a.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-horse-config-a.pkl',
- 'gdrive:networks/stylegan2-horse-config-f.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/stylegan2-horse-config-f.pkl',
- 'gdrive:networks/table2/stylegan2-car-config-e-Gorig-Dorig.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-car-config-e-Gorig-Dorig.pkl',
- 'gdrive:networks/table2/stylegan2-car-config-e-Gorig-Dresnet.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-car-config-e-Gorig-Dresnet.pkl',
- 'gdrive:networks/table2/stylegan2-car-config-e-Gorig-Dskip.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-car-config-e-Gorig-Dskip.pkl',
- 'gdrive:networks/table2/stylegan2-car-config-e-Gresnet-Dorig.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-car-config-e-Gresnet-Dorig.pkl',
- 'gdrive:networks/table2/stylegan2-car-config-e-Gresnet-Dresnet.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-car-config-e-Gresnet-Dresnet.pkl',
- 'gdrive:networks/table2/stylegan2-car-config-e-Gresnet-Dskip.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-car-config-e-Gresnet-Dskip.pkl',
- 'gdrive:networks/table2/stylegan2-car-config-e-Gskip-Dorig.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-car-config-e-Gskip-Dorig.pkl',
- 'gdrive:networks/table2/stylegan2-car-config-e-Gskip-Dresnet.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-car-config-e-Gskip-Dresnet.pkl',
- 'gdrive:networks/table2/stylegan2-car-config-e-Gskip-Dskip.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-car-config-e-Gskip-Dskip.pkl',
- 'gdrive:networks/table2/stylegan2-ffhq-config-e-Gorig-Dorig.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-ffhq-config-e-Gorig-Dorig.pkl',
- 'gdrive:networks/table2/stylegan2-ffhq-config-e-Gorig-Dresnet.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-ffhq-config-e-Gorig-Dresnet.pkl',
- 'gdrive:networks/table2/stylegan2-ffhq-config-e-Gorig-Dskip.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-ffhq-config-e-Gorig-Dskip.pkl',
- 'gdrive:networks/table2/stylegan2-ffhq-config-e-Gresnet-Dorig.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-ffhq-config-e-Gresnet-Dorig.pkl',
- 'gdrive:networks/table2/stylegan2-ffhq-config-e-Gresnet-Dresnet.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-ffhq-config-e-Gresnet-Dresnet.pkl',
- 'gdrive:networks/table2/stylegan2-ffhq-config-e-Gresnet-Dskip.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-ffhq-config-e-Gresnet-Dskip.pkl',
- 'gdrive:networks/table2/stylegan2-ffhq-config-e-Gskip-Dorig.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-ffhq-config-e-Gskip-Dorig.pkl',
- 'gdrive:networks/table2/stylegan2-ffhq-config-e-Gskip-Dresnet.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-ffhq-config-e-Gskip-Dresnet.pkl',
- 'gdrive:networks/table2/stylegan2-ffhq-config-e-Gskip-Dskip.pkl': 'http://d36zk2xti64re0.cloudfront.net/stylegan2/networks/table2/stylegan2-ffhq-config-e-Gskip-Dskip.pkl',
-}
-
-#----------------------------------------------------------------------------
-
-def get_path_or_url(path_or_gdrive_path):
- return gdrive_urls.get(path_or_gdrive_path, path_or_gdrive_path)
-
-#----------------------------------------------------------------------------
-
-_cached_networks = dict()
-
-def load_networks(path_or_gdrive_path):
- path_or_url = get_path_or_url(path_or_gdrive_path)
- if path_or_url in _cached_networks:
- return _cached_networks[path_or_url]
-
- if dnnlib.util.is_url(path_or_url):
- stream = dnnlib.util.open_url(path_or_url, cache_dir='.stylegan2-cache')
- else:
- stream = open(path_or_url, 'rb')
-
- tflib.init_tf()
- with stream:
- G, D, Gs = pickle.load(stream, encoding='latin1')
- _cached_networks[path_or_url] = G, D, Gs
- return G, D, Gs
-
-#----------------------------------------------------------------------------
diff --git a/lib/stylegan2/projector.py b/lib/stylegan2/projector.py
deleted file mode 100755
index 66edca8caef59c2731d7d5460be6f2943d591728..0000000000000000000000000000000000000000
--- a/lib/stylegan2/projector.py
+++ /dev/null
@@ -1,263 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-# 14.09.2020 : Adjustements by Laurent Colbois (Idiap, Biometrics Security and Privacy)
-# -> Adding two parameters to the __init__ method.
-# -> Adding verbosity option to the run method
-# -> Image postprocessing (adjust dynamic range)
-from bob.extension import rc
-import numpy as np
-import tensorflow as tf
-from .. import utils
-from . import dnnlib
-from .dnnlib import tflib
-
-import pickle
-
-#----------------------------------------------------------------------------
-
-class Projector(object):
- def __init__(self, vgg16_path=rc['sg2_morph.vgg16_path'], num_steps=1000):
- """
- Instanciate an image projector into the W space of a StyleGAN network.
- This class's code was borrowed from
- https://github.com/NVlabs/stylegan2/blob/master/projector.py,
- with a few adaptations.
-
- :param vgg16_path: path to the pickle file of the pretrained VGG16 network used for
- computing the perceptual loss between images.
- :param num_steps: Number of optimization steps to project a single image.
- """
- self.num_steps = num_steps
- with open(vgg16_path, 'rb') as pkl_file:
- self._lpips = pickle.load(pkl_file)
-
- self.dlatent_avg_samples = 10000
- self.initial_learning_rate = 0.1
- self.initial_noise_factor = 0.05
- self.lr_rampdown_length = 0.25
- self.lr_rampup_length = 0.05
- self.noise_ramp_length = 0.75
- self.regularize_noise_weight = 1e5
- self.verbose = False
- self.clone_net = True
- self.proc_images_size = 256
-
- self._Gs = None
- self._minibatch_size = None
- self._dlatent_avg = None
- self._dlatent_std = None
- self._noise_vars = None
- self._noise_init_op = None
- self._noise_normalize_op = None
- self._dlatents_var = None
- self._noise_in = None
- self._dlatents_expr = None
- self._images_expr = None
- self._target_images_var = None
- self._dist = None
- self._loss = None
- self._reg_sizes = None
- self._lrate_in = None
- self._opt = None
- self._opt_step = None
- self._cur_step = None
-
- def _info(self, *args):
- if self.verbose:
- print('Projector:', *args)
-
- def set_network(self, Gs):
- """
- Assign the generator network Gs to the projector.
- """
- self._Gs = Gs
- self._minibatch_size = 1
- if self._Gs is None:
- return
- if self.clone_net:
- self._Gs = self._Gs.clone()
-
- # Find dlatent stats.
- self._info('Finding W midpoint and stddev using %d samples...' % self.dlatent_avg_samples)
- latent_samples = np.random.RandomState(123).randn(self.dlatent_avg_samples, *self._Gs.input_shapes[0][1:])
- dlatent_samples = self._Gs.components.mapping.run(latent_samples, None)[:, :1, :] # [N, 1, 512]
- self._dlatent_avg = np.mean(dlatent_samples, axis=0, keepdims=True) # [1, 1, 512]
- self._dlatent_std = (np.sum((dlatent_samples - self._dlatent_avg) ** 2) / self.dlatent_avg_samples) ** 0.5
- self._info('std = %g' % self._dlatent_std)
-
- # Find noise inputs.
- self._info('Setting up noise inputs...')
- self._noise_vars = []
- noise_init_ops = []
- noise_normalize_ops = []
- while True:
- n = 'G_synthesis/noise%d' % len(self._noise_vars)
- if not n in self._Gs.vars:
- break
- v = self._Gs.vars[n]
- self._noise_vars.append(v)
- noise_init_ops.append(tf.assign(v, tf.random_normal(tf.shape(v), dtype=tf.float32)))
- noise_mean = tf.reduce_mean(v)
- noise_std = tf.reduce_mean((v - noise_mean)**2)**0.5
- noise_normalize_ops.append(tf.assign(v, (v - noise_mean) / noise_std))
- self._info(n, v)
- self._noise_init_op = tf.group(*noise_init_ops)
- self._noise_normalize_op = tf.group(*noise_normalize_ops)
-
- # Image output graph.
- self._info('Building image output graph...')
- self._dlatents_var = tf.Variable(tf.zeros([self._minibatch_size] + list(self._dlatent_avg.shape[1:])), name='dlatents_var')
- self._noise_in = tf.placeholder(tf.float32, [], name='noise_in')
- dlatents_noise = tf.random.normal(shape=self._dlatents_var.shape) * self._noise_in
- self._dlatents_expr = tf.tile(self._dlatents_var + dlatents_noise, [1, self._Gs.components.synthesis.input_shape[1], 1])
- self._images_expr = self._Gs.components.synthesis.get_output_for(self._dlatents_expr, randomize_noise=False)
-
- # Downsample image to 256x256 if it's larger than that. VGG was built for 224x224 images.
- proc_images_expr = (self._images_expr + 1) * (255 / 2)
- sh = proc_images_expr.shape.as_list()
- if sh[2] > self.proc_images_size:
- factor = sh[2] // self.proc_images_size
- proc_images_expr = tf.reduce_mean(tf.reshape(proc_images_expr, [-1, sh[1], sh[2] // factor, factor, sh[2] // factor, factor]), axis=[3,5])
-
- # Loss graph.
- self._info('Building loss graph...')
- self._target_images_var = tf.Variable(tf.zeros(proc_images_expr.shape), name='target_images_var')
- if self._lpips is None:
- self._lpips = misc.load_pkl('http://d36zk2xti64re0.cloudfront.net/stylegan1/networks/metrics/vgg16_zhang_perceptual.pkl')
- self._dist = self._lpips.get_output_for(proc_images_expr, self._target_images_var)
- self._loss = tf.reduce_sum(self._dist)
-
- # Noise regularization graph.
- self._info('Building noise regularization graph...')
- reg_loss = 0.0
- for v in self._noise_vars:
- sz = v.shape[2]
- while True:
- reg_loss += tf.reduce_mean(v * tf.roll(v, shift=1, axis=3))**2 + tf.reduce_mean(v * tf.roll(v, shift=1, axis=2))**2
- if sz <= 8:
- break # Small enough already
- v = tf.reshape(v, [1, 1, sz//2, 2, sz//2, 2]) # Downscale
- v = tf.reduce_mean(v, axis=[3, 5])
- sz = sz // 2
- self._loss += reg_loss * self.regularize_noise_weight
-
- # Optimizer.
- self._info('Setting up optimizer...')
- self._lrate_in = tf.placeholder(tf.float32, [], name='lrate_in')
- self._opt = dnnlib.tflib.Optimizer(learning_rate=self._lrate_in)
- self._opt.register_gradients(self._loss, [self._dlatents_var] + self._noise_vars)
- self._opt_step = self._opt.apply_updates()
-
- def __call__(self, target_image, verbose=False):
- """
- Project the provided image
-
- :param target_image: Image to project in bob format. Tensor of shape [3, 1024, 1024]
- of uint8 values in [0, 255]
- :param verbose: Display progression
-
- :return: Dictionary containing
- 'image': projection of the target image, in bob format. Tensor of shape [3, 1024, 1024]
- of uint8 values in [0, 255]
- 'w_latent': W-space latent vector corresponding to the projection. Shape [latent_dim]
- 'd_latents': repetition of the w_latent as many time as the synthesis network has style inputs. Shape [num_style_inputs, latent_dim]
- 'noises': optimized noise inputs for the projected image
-
- """
- return self.run(target_image, verbose)
-
- def run(self, target_image, verbose=False):
- """
- Project the provided image
-
- :param target_image: Image to project in bob format. Tensor of shape [3, 1024, 1024]
- of uint8 values in [0, 255]
- :param verbose: Display progression
-
- :return: Dictionary containing
- 'image': projection of the target image, in bob format. Tensor of shape [3, 1024, 1024]
- of uint8 values in [0, 255]
- 'w_latent': W-space latent vector corresponding to the projection. Shape [latent_dim]
- 'd_latents': repetition of the w_latent as many time as the synthesis network has style inputs. Shape [num_style_inputs, latent_dim]
- 'noises': optimized noise inputs for the projected image
-
- """
- # Run to completion.
- target_images = np.stack([target_image])
- self.start(target_images)
- while self._cur_step < self.num_steps:
- if verbose:
- print('Step {}/{}'.format(self._cur_step + 1, self.num_steps))
- self.step()
-
- # Collect results.
- pres = dnnlib.EasyDict()
- pres.dlatents = self.get_dlatents()[0]
- pres.w_latent = pres.dlatents[0, :]
- pres.noises = self.get_noises()
- pres.image = utils.adjust_dynamic_range(self.get_images()[0], [0, 255], 'uint8')
- return pres
-
- def start(self, target_images):
- assert self._Gs is not None
-
- # Prepare target images.
- self._info('Preparing target images...')
- target_images = np.asarray(target_images, dtype='float32')
- sh = target_images.shape
- assert sh[0] == self._minibatch_size
- if sh[2] > self._target_images_var.shape[2]:
- factor = sh[2] // self._target_images_var.shape[2]
- target_images = np.reshape(target_images, [-1, sh[1], sh[2] // factor, factor, sh[3] // factor, factor]).mean((3, 5))
-
- # Initialize optimization state.
- self._info('Initializing optimization state...')
- tflib.set_vars({self._target_images_var: target_images, self._dlatents_var: np.tile(self._dlatent_avg, [self._minibatch_size, 1, 1])})
- tflib.run(self._noise_init_op)
- self._opt.reset_optimizer_state()
- self._cur_step = 0
-
- def step(self):
- assert self._cur_step is not None
- if self._cur_step >= self.num_steps:
- return
- if self._cur_step == 0:
- self._info('Running...')
-
- # Hyperparameters.
- t = self._cur_step / self.num_steps
- noise_strength = self._dlatent_std * self.initial_noise_factor * max(0.0, 1.0 - t / self.noise_ramp_length) ** 2
- lr_ramp = min(1.0, (1.0 - t) / self.lr_rampdown_length)
- lr_ramp = 0.5 - 0.5 * np.cos(lr_ramp * np.pi)
- lr_ramp = lr_ramp * min(1.0, t / self.lr_rampup_length)
- learning_rate = self.initial_learning_rate * lr_ramp
-
- # Train.
- feed_dict = {self._noise_in: noise_strength, self._lrate_in: learning_rate}
- _, dist_value, loss_value = tflib.run([self._opt_step, self._dist, self._loss], feed_dict)
- tflib.run(self._noise_normalize_op)
-
- # Print status.
- self._cur_step += 1
- if self._cur_step == self.num_steps or self._cur_step % 10 == 0:
- self._info('%-8d%-12g%-12g' % (self._cur_step, dist_value, loss_value))
- if self._cur_step == self.num_steps:
- self._info('Done.')
-
- def get_cur_step(self):
- return self._cur_step
-
- def get_dlatents(self):
- return tflib.run(self._dlatents_expr, {self._noise_in: 0})
-
- def get_noises(self):
- return tflib.run(self._noise_vars)
-
- def get_images(self):
- return tflib.run(self._images_expr, {self._noise_in: 0})
-
-#----------------------------------------------------------------------------
diff --git a/lib/stylegan2/training/__init__.py b/lib/stylegan2/training/__init__.py
deleted file mode 100755
index 9ab9908efa3cb38af52e8d5bcaa8acffde5a8875..0000000000000000000000000000000000000000
--- a/lib/stylegan2/training/__init__.py
+++ /dev/null
@@ -1,7 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-# empty
diff --git a/lib/stylegan2/training/dataset.py b/lib/stylegan2/training/dataset.py
deleted file mode 100755
index 2d1059838ab5ec95d5acce62abac6dc93313ba4a..0000000000000000000000000000000000000000
--- a/lib/stylegan2/training/dataset.py
+++ /dev/null
@@ -1,199 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Multi-resolution input data pipeline."""
-
-import os
-import glob
-import numpy as np
-import tensorflow as tf
-import dnnlib
-import dnnlib.tflib as tflib
-
-#----------------------------------------------------------------------------
-# Dataset class that loads data from tfrecords files.
-
-class TFRecordDataset:
- def __init__(self,
- tfrecord_dir, # Directory containing a collection of tfrecords files.
- resolution = None, # Dataset resolution, None = autodetect.
- label_file = None, # Relative path of the labels file, None = autodetect.
- max_label_size = 0, # 0 = no labels, 'full' = full labels, <int> = N first label components.
- max_images = None, # Maximum number of images to use, None = use all images.
- repeat = True, # Repeat dataset indefinitely?
- shuffle_mb = 4096, # Shuffle data within specified window (megabytes), 0 = disable shuffling.
- prefetch_mb = 2048, # Amount of data to prefetch (megabytes), 0 = disable prefetching.
- buffer_mb = 256, # Read buffer size (megabytes).
- num_threads = 2): # Number of concurrent threads.
-
- self.tfrecord_dir = tfrecord_dir
- self.resolution = None
- self.resolution_log2 = None
- self.shape = [] # [channels, height, width]
- self.dtype = 'uint8'
- self.dynamic_range = [0, 255]
- self.label_file = label_file
- self.label_size = None # components
- self.label_dtype = None
- self._np_labels = None
- self._tf_minibatch_in = None
- self._tf_labels_var = None
- self._tf_labels_dataset = None
- self._tf_datasets = dict()
- self._tf_iterator = None
- self._tf_init_ops = dict()
- self._tf_minibatch_np = None
- self._cur_minibatch = -1
- self._cur_lod = -1
-
- # List tfrecords files and inspect their shapes.
- assert os.path.isdir(self.tfrecord_dir)
- tfr_files = sorted(glob.glob(os.path.join(self.tfrecord_dir, '*.tfrecords')))
- assert len(tfr_files) >= 1
- tfr_shapes = []
- for tfr_file in tfr_files:
- tfr_opt = tf.python_io.TFRecordOptions(tf.python_io.TFRecordCompressionType.NONE)
- for record in tf.python_io.tf_record_iterator(tfr_file, tfr_opt):
- tfr_shapes.append(self.parse_tfrecord_np(record).shape)
- break
-
- # Autodetect label filename.
- if self.label_file is None:
- guess = sorted(glob.glob(os.path.join(self.tfrecord_dir, '*.labels')))
- if len(guess):
- self.label_file = guess[0]
- elif not os.path.isfile(self.label_file):
- guess = os.path.join(self.tfrecord_dir, self.label_file)
- if os.path.isfile(guess):
- self.label_file = guess
-
- # Determine shape and resolution.
- max_shape = max(tfr_shapes, key=np.prod)
- self.resolution = resolution if resolution is not None else max_shape[1]
- self.resolution_log2 = int(np.log2(self.resolution))
- self.shape = [max_shape[0], self.resolution, self.resolution]
- tfr_lods = [self.resolution_log2 - int(np.log2(shape[1])) for shape in tfr_shapes]
- assert all(shape[0] == max_shape[0] for shape in tfr_shapes)
- assert all(shape[1] == shape[2] for shape in tfr_shapes)
- assert all(shape[1] == self.resolution // (2**lod) for shape, lod in zip(tfr_shapes, tfr_lods))
- assert all(lod in tfr_lods for lod in range(self.resolution_log2 - 1))
-
- # Load labels.
- assert max_label_size == 'full' or max_label_size >= 0
- self._np_labels = np.zeros([1<<30, 0], dtype=np.float32)
- if self.label_file is not None and max_label_size != 0:
- self._np_labels = np.load(self.label_file)
- assert self._np_labels.ndim == 2
- if max_label_size != 'full' and self._np_labels.shape[1] > max_label_size:
- self._np_labels = self._np_labels[:, :max_label_size]
- if max_images is not None and self._np_labels.shape[0] > max_images:
- self._np_labels = self._np_labels[:max_images]
- self.label_size = self._np_labels.shape[1]
- self.label_dtype = self._np_labels.dtype.name
-
- # Build TF expressions.
- with tf.name_scope('Dataset'), tf.device('/cpu:0'):
- self._tf_minibatch_in = tf.placeholder(tf.int64, name='minibatch_in', shape=[])
- self._tf_labels_var = tflib.create_var_with_large_initial_value(self._np_labels, name='labels_var')
- self._tf_labels_dataset = tf.data.Dataset.from_tensor_slices(self._tf_labels_var)
- for tfr_file, tfr_shape, tfr_lod in zip(tfr_files, tfr_shapes, tfr_lods):
- if tfr_lod < 0:
- continue
- dset = tf.data.TFRecordDataset(tfr_file, compression_type='', buffer_size=buffer_mb<<20)
- if max_images is not None:
- dset = dset.take(max_images)
- dset = dset.map(self.parse_tfrecord_tf, num_parallel_calls=num_threads)
- dset = tf.data.Dataset.zip((dset, self._tf_labels_dataset))
- bytes_per_item = np.prod(tfr_shape) * np.dtype(self.dtype).itemsize
- if shuffle_mb > 0:
- dset = dset.shuffle(((shuffle_mb << 20) - 1) // bytes_per_item + 1)
- if repeat:
- dset = dset.repeat()
- if prefetch_mb > 0:
- dset = dset.prefetch(((prefetch_mb << 20) - 1) // bytes_per_item + 1)
- dset = dset.batch(self._tf_minibatch_in)
- self._tf_datasets[tfr_lod] = dset
- self._tf_iterator = tf.data.Iterator.from_structure(self._tf_datasets[0].output_types, self._tf_datasets[0].output_shapes)
- self._tf_init_ops = {lod: self._tf_iterator.make_initializer(dset) for lod, dset in self._tf_datasets.items()}
-
- def close(self):
- pass
-
- # Use the given minibatch size and level-of-detail for the data returned by get_minibatch_tf().
- def configure(self, minibatch_size, lod=0):
- lod = int(np.floor(lod))
- assert minibatch_size >= 1 and lod in self._tf_datasets
- if self._cur_minibatch != minibatch_size or self._cur_lod != lod:
- self._tf_init_ops[lod].run({self._tf_minibatch_in: minibatch_size})
- self._cur_minibatch = minibatch_size
- self._cur_lod = lod
-
- # Get next minibatch as TensorFlow expressions.
- def get_minibatch_tf(self): # => images, labels
- return self._tf_iterator.get_next()
-
- # Get next minibatch as NumPy arrays.
- def get_minibatch_np(self, minibatch_size, lod=0): # => images, labels
- self.configure(minibatch_size, lod)
- with tf.name_scope('Dataset'):
- if self._tf_minibatch_np is None:
- self._tf_minibatch_np = self.get_minibatch_tf()
- return tflib.run(self._tf_minibatch_np)
-
- # Get random labels as TensorFlow expression.
- def get_random_labels_tf(self, minibatch_size): # => labels
- with tf.name_scope('Dataset'):
- if self.label_size > 0:
- with tf.device('/cpu:0'):
- return tf.gather(self._tf_labels_var, tf.random_uniform([minibatch_size], 0, self._np_labels.shape[0], dtype=tf.int32))
- return tf.zeros([minibatch_size, 0], self.label_dtype)
-
- # Get random labels as NumPy array.
- def get_random_labels_np(self, minibatch_size): # => labels
- if self.label_size > 0:
- return self._np_labels[np.random.randint(self._np_labels.shape[0], size=[minibatch_size])]
- return np.zeros([minibatch_size, 0], self.label_dtype)
-
- # Parse individual image from a tfrecords file into TensorFlow expression.
- @staticmethod
- def parse_tfrecord_tf(record):
- features = tf.parse_single_example(record, features={
- 'shape': tf.FixedLenFeature([3], tf.int64),
- 'data': tf.FixedLenFeature([], tf.string)})
- data = tf.decode_raw(features['data'], tf.uint8)
- return tf.reshape(data, features['shape'])
-
- # Parse individual image from a tfrecords file into NumPy array.
- @staticmethod
- def parse_tfrecord_np(record):
- ex = tf.train.Example()
- ex.ParseFromString(record)
- shape = ex.features.feature['shape'].int64_list.value # pylint: disable=no-member
- data = ex.features.feature['data'].bytes_list.value[0] # pylint: disable=no-member
- return np.fromstring(data, np.uint8).reshape(shape)
-
-#----------------------------------------------------------------------------
-# Helper func for constructing a dataset object using the given options.
-
-def load_dataset(class_name=None, data_dir=None, verbose=False, **kwargs):
- kwargs = dict(kwargs)
- if 'tfrecord_dir' in kwargs:
- if class_name is None:
- class_name = __name__ + '.TFRecordDataset'
- if data_dir is not None:
- kwargs['tfrecord_dir'] = os.path.join(data_dir, kwargs['tfrecord_dir'])
-
- assert class_name is not None
- if verbose:
- print('Streaming data using %s...' % class_name)
- dataset = dnnlib.util.get_obj_by_name(class_name)(**kwargs)
- if verbose:
- print('Dataset shape =', np.int32(dataset.shape).tolist())
- print('Dynamic range =', dataset.dynamic_range)
- print('Label size =', dataset.label_size)
- return dataset
-
-#----------------------------------------------------------------------------
diff --git a/lib/stylegan2/training/loss.py b/lib/stylegan2/training/loss.py
deleted file mode 100755
index 7ad2fe16fb657627bfaa547f28653a615ec395c2..0000000000000000000000000000000000000000
--- a/lib/stylegan2/training/loss.py
+++ /dev/null
@@ -1,197 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Loss functions."""
-
-import numpy as np
-import tensorflow as tf
-import dnnlib.tflib as tflib
-from dnnlib.tflib.autosummary import autosummary
-
-#----------------------------------------------------------------------------
-# Logistic loss from the paper
-# "Generative Adversarial Nets", Goodfellow et al. 2014
-
-def G_logistic(G, D, opt, training_set, minibatch_size):
- _ = opt
- latents = tf.random_normal([minibatch_size] + G.input_shapes[0][1:])
- labels = training_set.get_random_labels_tf(minibatch_size)
- fake_images_out = G.get_output_for(latents, labels, is_training=True)
- fake_scores_out = D.get_output_for(fake_images_out, labels, is_training=True)
- loss = -tf.nn.softplus(fake_scores_out) # log(1-sigmoid(fake_scores_out)) # pylint: disable=invalid-unary-operand-type
- return loss, None
-
-def G_logistic_ns(G, D, opt, training_set, minibatch_size):
- _ = opt
- latents = tf.random_normal([minibatch_size] + G.input_shapes[0][1:])
- labels = training_set.get_random_labels_tf(minibatch_size)
- fake_images_out = G.get_output_for(latents, labels, is_training=True)
- fake_scores_out = D.get_output_for(fake_images_out, labels, is_training=True)
- loss = tf.nn.softplus(-fake_scores_out) # -log(sigmoid(fake_scores_out))
- return loss, None
-
-def D_logistic(G, D, opt, training_set, minibatch_size, reals, labels):
- _ = opt, training_set
- latents = tf.random_normal([minibatch_size] + G.input_shapes[0][1:])
- fake_images_out = G.get_output_for(latents, labels, is_training=True)
- real_scores_out = D.get_output_for(reals, labels, is_training=True)
- fake_scores_out = D.get_output_for(fake_images_out, labels, is_training=True)
- real_scores_out = autosummary('Loss/scores/real', real_scores_out)
- fake_scores_out = autosummary('Loss/scores/fake', fake_scores_out)
- loss = tf.nn.softplus(fake_scores_out) # -log(1-sigmoid(fake_scores_out))
- loss += tf.nn.softplus(-real_scores_out) # -log(sigmoid(real_scores_out)) # pylint: disable=invalid-unary-operand-type
- return loss, None
-
-#----------------------------------------------------------------------------
-# R1 and R2 regularizers from the paper
-# "Which Training Methods for GANs do actually Converge?", Mescheder et al. 2018
-
-def D_logistic_r1(G, D, opt, training_set, minibatch_size, reals, labels, gamma=10.0):
- _ = opt, training_set
- latents = tf.random_normal([minibatch_size] + G.input_shapes[0][1:])
- fake_images_out = G.get_output_for(latents, labels, is_training=True)
- real_scores_out = D.get_output_for(reals, labels, is_training=True)
- fake_scores_out = D.get_output_for(fake_images_out, labels, is_training=True)
- real_scores_out = autosummary('Loss/scores/real', real_scores_out)
- fake_scores_out = autosummary('Loss/scores/fake', fake_scores_out)
- loss = tf.nn.softplus(fake_scores_out) # -log(1-sigmoid(fake_scores_out))
- loss += tf.nn.softplus(-real_scores_out) # -log(sigmoid(real_scores_out)) # pylint: disable=invalid-unary-operand-type
-
- with tf.name_scope('GradientPenalty'):
- real_grads = tf.gradients(tf.reduce_sum(real_scores_out), [reals])[0]
- gradient_penalty = tf.reduce_sum(tf.square(real_grads), axis=[1,2,3])
- gradient_penalty = autosummary('Loss/gradient_penalty', gradient_penalty)
- reg = gradient_penalty * (gamma * 0.5)
- return loss, reg
-
-def D_logistic_r2(G, D, opt, training_set, minibatch_size, reals, labels, gamma=10.0):
- _ = opt, training_set
- latents = tf.random_normal([minibatch_size] + G.input_shapes[0][1:])
- fake_images_out = G.get_output_for(latents, labels, is_training=True)
- real_scores_out = D.get_output_for(reals, labels, is_training=True)
- fake_scores_out = D.get_output_for(fake_images_out, labels, is_training=True)
- real_scores_out = autosummary('Loss/scores/real', real_scores_out)
- fake_scores_out = autosummary('Loss/scores/fake', fake_scores_out)
- loss = tf.nn.softplus(fake_scores_out) # -log(1-sigmoid(fake_scores_out))
- loss += tf.nn.softplus(-real_scores_out) # -log(sigmoid(real_scores_out)) # pylint: disable=invalid-unary-operand-type
-
- with tf.name_scope('GradientPenalty'):
- fake_grads = tf.gradients(tf.reduce_sum(fake_scores_out), [fake_images_out])[0]
- gradient_penalty = tf.reduce_sum(tf.square(fake_grads), axis=[1,2,3])
- gradient_penalty = autosummary('Loss/gradient_penalty', gradient_penalty)
- reg = gradient_penalty * (gamma * 0.5)
- return loss, reg
-
-#----------------------------------------------------------------------------
-# WGAN loss from the paper
-# "Wasserstein Generative Adversarial Networks", Arjovsky et al. 2017
-
-def G_wgan(G, D, opt, training_set, minibatch_size):
- _ = opt
- latents = tf.random_normal([minibatch_size] + G.input_shapes[0][1:])
- labels = training_set.get_random_labels_tf(minibatch_size)
- fake_images_out = G.get_output_for(latents, labels, is_training=True)
- fake_scores_out = D.get_output_for(fake_images_out, labels, is_training=True)
- loss = -fake_scores_out
- return loss, None
-
-def D_wgan(G, D, opt, training_set, minibatch_size, reals, labels, wgan_epsilon=0.001):
- _ = opt, training_set
- latents = tf.random_normal([minibatch_size] + G.input_shapes[0][1:])
- fake_images_out = G.get_output_for(latents, labels, is_training=True)
- real_scores_out = D.get_output_for(reals, labels, is_training=True)
- fake_scores_out = D.get_output_for(fake_images_out, labels, is_training=True)
- real_scores_out = autosummary('Loss/scores/real', real_scores_out)
- fake_scores_out = autosummary('Loss/scores/fake', fake_scores_out)
- loss = fake_scores_out - real_scores_out
- with tf.name_scope('EpsilonPenalty'):
- epsilon_penalty = autosummary('Loss/epsilon_penalty', tf.square(real_scores_out))
- loss += epsilon_penalty * wgan_epsilon
- return loss, None
-
-#----------------------------------------------------------------------------
-# WGAN-GP loss from the paper
-# "Improved Training of Wasserstein GANs", Gulrajani et al. 2017
-
-def D_wgan_gp(G, D, opt, training_set, minibatch_size, reals, labels, wgan_lambda=10.0, wgan_epsilon=0.001, wgan_target=1.0):
- _ = opt, training_set
- latents = tf.random_normal([minibatch_size] + G.input_shapes[0][1:])
- fake_images_out = G.get_output_for(latents, labels, is_training=True)
- real_scores_out = D.get_output_for(reals, labels, is_training=True)
- fake_scores_out = D.get_output_for(fake_images_out, labels, is_training=True)
- real_scores_out = autosummary('Loss/scores/real', real_scores_out)
- fake_scores_out = autosummary('Loss/scores/fake', fake_scores_out)
- loss = fake_scores_out - real_scores_out
- with tf.name_scope('EpsilonPenalty'):
- epsilon_penalty = autosummary('Loss/epsilon_penalty', tf.square(real_scores_out))
- loss += epsilon_penalty * wgan_epsilon
-
- with tf.name_scope('GradientPenalty'):
- mixing_factors = tf.random_uniform([minibatch_size, 1, 1, 1], 0.0, 1.0, dtype=fake_images_out.dtype)
- mixed_images_out = tflib.lerp(tf.cast(reals, fake_images_out.dtype), fake_images_out, mixing_factors)
- mixed_scores_out = D.get_output_for(mixed_images_out, labels, is_training=True)
- mixed_scores_out = autosummary('Loss/scores/mixed', mixed_scores_out)
- mixed_grads = tf.gradients(tf.reduce_sum(mixed_scores_out), [mixed_images_out])[0]
- mixed_norms = tf.sqrt(tf.reduce_sum(tf.square(mixed_grads), axis=[1,2,3]))
- mixed_norms = autosummary('Loss/mixed_norms', mixed_norms)
- gradient_penalty = tf.square(mixed_norms - wgan_target)
- reg = gradient_penalty * (wgan_lambda / (wgan_target**2))
- return loss, reg
-
-#----------------------------------------------------------------------------
-# Non-saturating logistic loss with path length regularizer from the paper
-# "Analyzing and Improving the Image Quality of StyleGAN", Karras et al. 2019
-
-def G_logistic_ns_pathreg(G, D, opt, training_set, minibatch_size, pl_minibatch_shrink=2, pl_decay=0.01, pl_weight=2.0):
- _ = opt
- latents = tf.random_normal([minibatch_size] + G.input_shapes[0][1:])
- labels = training_set.get_random_labels_tf(minibatch_size)
- fake_images_out, fake_dlatents_out = G.get_output_for(latents, labels, is_training=True, return_dlatents=True)
- fake_scores_out = D.get_output_for(fake_images_out, labels, is_training=True)
- loss = tf.nn.softplus(-fake_scores_out) # -log(sigmoid(fake_scores_out))
-
- # Path length regularization.
- with tf.name_scope('PathReg'):
-
- # Evaluate the regularization term using a smaller minibatch to conserve memory.
- if pl_minibatch_shrink > 1:
- pl_minibatch = minibatch_size // pl_minibatch_shrink
- pl_latents = tf.random_normal([pl_minibatch] + G.input_shapes[0][1:])
- pl_labels = training_set.get_random_labels_tf(pl_minibatch)
- fake_images_out, fake_dlatents_out = G.get_output_for(pl_latents, pl_labels, is_training=True, return_dlatents=True)
-
- # Compute |J*y|.
- pl_noise = tf.random_normal(tf.shape(fake_images_out)) / np.sqrt(np.prod(G.output_shape[2:]))
- pl_grads = tf.gradients(tf.reduce_sum(fake_images_out * pl_noise), [fake_dlatents_out])[0]
- pl_lengths = tf.sqrt(tf.reduce_mean(tf.reduce_sum(tf.square(pl_grads), axis=2), axis=1))
- pl_lengths = autosummary('Loss/pl_lengths', pl_lengths)
-
- # Track exponential moving average of |J*y|.
- with tf.control_dependencies(None):
- pl_mean_var = tf.Variable(name='pl_mean', trainable=False, initial_value=0.0, dtype=tf.float32)
- pl_mean = pl_mean_var + pl_decay * (tf.reduce_mean(pl_lengths) - pl_mean_var)
- pl_update = tf.assign(pl_mean_var, pl_mean)
-
- # Calculate (|J*y|-a)^2.
- with tf.control_dependencies([pl_update]):
- pl_penalty = tf.square(pl_lengths - pl_mean)
- pl_penalty = autosummary('Loss/pl_penalty', pl_penalty)
-
- # Apply weight.
- #
- # Note: The division in pl_noise decreases the weight by num_pixels, and the reduce_mean
- # in pl_lengths decreases it by num_affine_layers. The effective weight then becomes:
- #
- # gamma_pl = pl_weight / num_pixels / num_affine_layers
- # = 2 / (r^2) / (log2(r) * 2 - 2)
- # = 1 / (r^2 * (log2(r) - 1))
- # = ln(2) / (r^2 * (ln(r) - ln(2))
- #
- reg = pl_penalty * pl_weight
-
- return loss, reg
-
-#----------------------------------------------------------------------------
diff --git a/lib/stylegan2/training/misc.py b/lib/stylegan2/training/misc.py
deleted file mode 100755
index fb559c417056ad86140bc7fe55efc397d67a0945..0000000000000000000000000000000000000000
--- a/lib/stylegan2/training/misc.py
+++ /dev/null
@@ -1,145 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Miscellaneous utility functions."""
-
-import os
-import pickle
-import numpy as np
-import PIL.Image
-import PIL.ImageFont
-import dnnlib
-
-#----------------------------------------------------------------------------
-# Convenience wrappers for pickle that are able to load data produced by
-# older versions of the code, and from external URLs.
-
-def open_file_or_url(file_or_url):
- if dnnlib.util.is_url(file_or_url):
- return dnnlib.util.open_url(file_or_url, cache_dir='/idiap/temp/lcolbois/cache/stylegan2/')
- return open(file_or_url, 'rb')
-
-def load_pkl(file_or_url):
- with open_file_or_url(file_or_url) as file:
- return pickle.load(file, encoding='latin1')
-
-def save_pkl(obj, filename):
- with open(filename, 'wb') as file:
- pickle.dump(obj, file, protocol=pickle.HIGHEST_PROTOCOL)
-
-#----------------------------------------------------------------------------
-# Image utils.
-
-def adjust_dynamic_range(data, drange_in, drange_out):
- if drange_in != drange_out:
- scale = (np.float32(drange_out[1]) - np.float32(drange_out[0])) / (np.float32(drange_in[1]) - np.float32(drange_in[0]))
- bias = (np.float32(drange_out[0]) - np.float32(drange_in[0]) * scale)
- data = data * scale + bias
- return data
-
-def create_image_grid(images, grid_size=None):
- assert images.ndim == 3 or images.ndim == 4
- num, img_w, img_h = images.shape[0], images.shape[-1], images.shape[-2]
-
- if grid_size is not None:
- grid_w, grid_h = tuple(grid_size)
- else:
- grid_w = max(int(np.ceil(np.sqrt(num))), 1)
- grid_h = max((num - 1) // grid_w + 1, 1)
-
- grid = np.zeros(list(images.shape[1:-2]) + [grid_h * img_h, grid_w * img_w], dtype=images.dtype)
- for idx in range(num):
- x = (idx % grid_w) * img_w
- y = (idx // grid_w) * img_h
- grid[..., y : y + img_h, x : x + img_w] = images[idx]
- return grid
-
-def convert_to_pil_image(image, drange=[0,1]):
- assert image.ndim == 2 or image.ndim == 3
- if image.ndim == 3:
- if image.shape[0] == 1:
- image = image[0] # grayscale CHW => HW
- else:
- image = image.transpose(1, 2, 0) # CHW -> HWC
-
- image = adjust_dynamic_range(image, drange, [0,255])
- image = np.rint(image).clip(0, 255).astype(np.uint8)
- fmt = 'RGB' if image.ndim == 3 else 'L'
- return PIL.Image.fromarray(image, fmt)
-
-def save_image_grid(images, filename, drange=[0,1], grid_size=None):
- convert_to_pil_image(create_image_grid(images, grid_size), drange).save(filename)
-
-def apply_mirror_augment(minibatch):
- mask = np.random.rand(minibatch.shape[0]) < 0.5
- minibatch = np.array(minibatch)
- minibatch[mask] = minibatch[mask, :, :, ::-1]
- return minibatch
-
-#----------------------------------------------------------------------------
-# Loading data from previous training runs.
-
-def parse_config_for_previous_run(run_dir):
- with open(os.path.join(run_dir, 'submit_config.pkl'), 'rb') as f:
- data = pickle.load(f)
- data = data.get('run_func_kwargs', {})
- return dict(train=data, dataset=data.get('dataset_args', {}))
-
-#----------------------------------------------------------------------------
-# Size and contents of the image snapshot grids that are exported
-# periodically during training.
-
-def setup_snapshot_image_grid(training_set,
- size = '1080p', # '1080p' = to be viewed on 1080p display, '4k' = to be viewed on 4k display.
- layout = 'random'): # 'random' = grid contents are selected randomly, 'row_per_class' = each row corresponds to one class label.
-
- # Select size.
- gw = 1; gh = 1
- if size == '1080p':
- gw = np.clip(1920 // training_set.shape[2], 3, 32)
- gh = np.clip(1080 // training_set.shape[1], 2, 32)
- if size == '4k':
- gw = np.clip(3840 // training_set.shape[2], 7, 32)
- gh = np.clip(2160 // training_set.shape[1], 4, 32)
- if size == '8k':
- gw = np.clip(7680 // training_set.shape[2], 7, 32)
- gh = np.clip(4320 // training_set.shape[1], 4, 32)
-
- # Initialize data arrays.
- reals = np.zeros([gw * gh] + training_set.shape, dtype=training_set.dtype)
- labels = np.zeros([gw * gh, training_set.label_size], dtype=training_set.label_dtype)
-
- # Random layout.
- if layout == 'random':
- reals[:], labels[:] = training_set.get_minibatch_np(gw * gh)
-
- # Class-conditional layouts.
- class_layouts = dict(row_per_class=[gw,1], col_per_class=[1,gh], class4x4=[4,4])
- if layout in class_layouts:
- bw, bh = class_layouts[layout]
- nw = (gw - 1) // bw + 1
- nh = (gh - 1) // bh + 1
- blocks = [[] for _i in range(nw * nh)]
- for _iter in range(1000000):
- real, label = training_set.get_minibatch_np(1)
- idx = np.argmax(label[0])
- while idx < len(blocks) and len(blocks[idx]) >= bw * bh:
- idx += training_set.label_size
- if idx < len(blocks):
- blocks[idx].append((real, label))
- if all(len(block) >= bw * bh for block in blocks):
- break
- for i, block in enumerate(blocks):
- for j, (real, label) in enumerate(block):
- x = (i % nw) * bw + j % bw
- y = (i // nw) * bh + j // bw
- if x < gw and y < gh:
- reals[x + y * gw] = real[0]
- labels[x + y * gw] = label[0]
-
- return (gw, gh), reals, labels
-
-#----------------------------------------------------------------------------
diff --git a/lib/stylegan2/training/networks_stylegan.py b/lib/stylegan2/training/networks_stylegan.py
deleted file mode 100755
index 76ce31caa0890becebdfc481d92ec81d0023f999..0000000000000000000000000000000000000000
--- a/lib/stylegan2/training/networks_stylegan.py
+++ /dev/null
@@ -1,660 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Network architectures used in the StyleGAN paper."""
-
-import numpy as np
-import tensorflow as tf
-import dnnlib
-import dnnlib.tflib as tflib
-
-# NOTE: Do not import any application-specific modules here!
-# Specify all network parameters as kwargs.
-
-#----------------------------------------------------------------------------
-# Primitive ops for manipulating 4D activation tensors.
-# The gradients of these are not necessary efficient or even meaningful.
-
-def _blur2d(x, f=[1,2,1], normalize=True, flip=False, stride=1):
- assert x.shape.ndims == 4 and all(dim.value is not None for dim in x.shape[1:])
- assert isinstance(stride, int) and stride >= 1
-
- # Finalize filter kernel.
- f = np.array(f, dtype=np.float32)
- if f.ndim == 1:
- f = f[:, np.newaxis] * f[np.newaxis, :]
- assert f.ndim == 2
- if normalize:
- f /= np.sum(f)
- if flip:
- f = f[::-1, ::-1]
- f = f[:, :, np.newaxis, np.newaxis]
- f = np.tile(f, [1, 1, int(x.shape[1]), 1])
-
- # No-op => early exit.
- if f.shape == (1, 1) and f[0,0] == 1:
- return x
-
- # Convolve using depthwise_conv2d.
- orig_dtype = x.dtype
- x = tf.cast(x, tf.float32) # tf.nn.depthwise_conv2d() doesn't support fp16
- f = tf.constant(f, dtype=x.dtype, name='filter')
- strides = [1, 1, stride, stride]
- x = tf.nn.depthwise_conv2d(x, f, strides=strides, padding='SAME', data_format='NCHW')
- x = tf.cast(x, orig_dtype)
- return x
-
-def _upscale2d(x, factor=2, gain=1):
- assert x.shape.ndims == 4 and all(dim.value is not None for dim in x.shape[1:])
- assert isinstance(factor, int) and factor >= 1
-
- # Apply gain.
- if gain != 1:
- x *= gain
-
- # No-op => early exit.
- if factor == 1:
- return x
-
- # Upscale using tf.tile().
- s = x.shape
- x = tf.reshape(x, [-1, s[1], s[2], 1, s[3], 1])
- x = tf.tile(x, [1, 1, 1, factor, 1, factor])
- x = tf.reshape(x, [-1, s[1], s[2] * factor, s[3] * factor])
- return x
-
-def _downscale2d(x, factor=2, gain=1):
- assert x.shape.ndims == 4 and all(dim.value is not None for dim in x.shape[1:])
- assert isinstance(factor, int) and factor >= 1
-
- # 2x2, float32 => downscale using _blur2d().
- if factor == 2 and x.dtype == tf.float32:
- f = [np.sqrt(gain) / factor] * factor
- return _blur2d(x, f=f, normalize=False, stride=factor)
-
- # Apply gain.
- if gain != 1:
- x *= gain
-
- # No-op => early exit.
- if factor == 1:
- return x
-
- # Large factor => downscale using tf.nn.avg_pool().
- # NOTE: Requires tf_config['graph_options.place_pruned_graph']=True to work.
- ksize = [1, 1, factor, factor]
- return tf.nn.avg_pool(x, ksize=ksize, strides=ksize, padding='VALID', data_format='NCHW')
-
-#----------------------------------------------------------------------------
-# High-level ops for manipulating 4D activation tensors.
-# The gradients of these are meant to be as efficient as possible.
-
-def blur2d(x, f=[1,2,1], normalize=True):
- with tf.variable_scope('Blur2D'):
- @tf.custom_gradient
- def func(x):
- y = _blur2d(x, f, normalize)
- @tf.custom_gradient
- def grad(dy):
- dx = _blur2d(dy, f, normalize, flip=True)
- return dx, lambda ddx: _blur2d(ddx, f, normalize)
- return y, grad
- return func(x)
-
-def upscale2d(x, factor=2):
- with tf.variable_scope('Upscale2D'):
- @tf.custom_gradient
- def func(x):
- y = _upscale2d(x, factor)
- @tf.custom_gradient
- def grad(dy):
- dx = _downscale2d(dy, factor, gain=factor**2)
- return dx, lambda ddx: _upscale2d(ddx, factor)
- return y, grad
- return func(x)
-
-def downscale2d(x, factor=2):
- with tf.variable_scope('Downscale2D'):
- @tf.custom_gradient
- def func(x):
- y = _downscale2d(x, factor)
- @tf.custom_gradient
- def grad(dy):
- dx = _upscale2d(dy, factor, gain=1/factor**2)
- return dx, lambda ddx: _downscale2d(ddx, factor)
- return y, grad
- return func(x)
-
-#----------------------------------------------------------------------------
-# Get/create weight tensor for a convolutional or fully-connected layer.
-
-def get_weight(shape, gain=np.sqrt(2), use_wscale=False, lrmul=1):
- fan_in = np.prod(shape[:-1]) # [kernel, kernel, fmaps_in, fmaps_out] or [in, out]
- he_std = gain / np.sqrt(fan_in) # He init
-
- # Equalized learning rate and custom learning rate multiplier.
- if use_wscale:
- init_std = 1.0 / lrmul
- runtime_coef = he_std * lrmul
- else:
- init_std = he_std / lrmul
- runtime_coef = lrmul
-
- # Create variable.
- init = tf.initializers.random_normal(0, init_std)
- return tf.get_variable('weight', shape=shape, initializer=init) * runtime_coef
-
-#----------------------------------------------------------------------------
-# Fully-connected layer.
-
-def dense(x, fmaps, **kwargs):
- if len(x.shape) > 2:
- x = tf.reshape(x, [-1, np.prod([d.value for d in x.shape[1:]])])
- w = get_weight([x.shape[1].value, fmaps], **kwargs)
- w = tf.cast(w, x.dtype)
- return tf.matmul(x, w)
-
-#----------------------------------------------------------------------------
-# Convolutional layer.
-
-def conv2d(x, fmaps, kernel, **kwargs):
- assert kernel >= 1 and kernel % 2 == 1
- w = get_weight([kernel, kernel, x.shape[1].value, fmaps], **kwargs)
- w = tf.cast(w, x.dtype)
- return tf.nn.conv2d(x, w, strides=[1,1,1,1], padding='SAME', data_format='NCHW')
-
-#----------------------------------------------------------------------------
-# Fused convolution + scaling.
-# Faster and uses less memory than performing the operations separately.
-
-def upscale2d_conv2d(x, fmaps, kernel, fused_scale='auto', **kwargs):
- assert kernel >= 1 and kernel % 2 == 1
- assert fused_scale in [True, False, 'auto']
- if fused_scale == 'auto':
- fused_scale = min(x.shape[2:]) * 2 >= 128
-
- # Not fused => call the individual ops directly.
- if not fused_scale:
- return conv2d(upscale2d(x), fmaps, kernel, **kwargs)
-
- # Fused => perform both ops simultaneously using tf.nn.conv2d_transpose().
- w = get_weight([kernel, kernel, x.shape[1].value, fmaps], **kwargs)
- w = tf.transpose(w, [0, 1, 3, 2]) # [kernel, kernel, fmaps_out, fmaps_in]
- w = tf.pad(w, [[1,1], [1,1], [0,0], [0,0]], mode='CONSTANT')
- w = tf.add_n([w[1:, 1:], w[:-1, 1:], w[1:, :-1], w[:-1, :-1]])
- w = tf.cast(w, x.dtype)
- os = [tf.shape(x)[0], fmaps, x.shape[2] * 2, x.shape[3] * 2]
- return tf.nn.conv2d_transpose(x, w, os, strides=[1,1,2,2], padding='SAME', data_format='NCHW')
-
-def conv2d_downscale2d(x, fmaps, kernel, fused_scale='auto', **kwargs):
- assert kernel >= 1 and kernel % 2 == 1
- assert fused_scale in [True, False, 'auto']
- if fused_scale == 'auto':
- fused_scale = min(x.shape[2:]) >= 128
-
- # Not fused => call the individual ops directly.
- if not fused_scale:
- return downscale2d(conv2d(x, fmaps, kernel, **kwargs))
-
- # Fused => perform both ops simultaneously using tf.nn.conv2d().
- w = get_weight([kernel, kernel, x.shape[1].value, fmaps], **kwargs)
- w = tf.pad(w, [[1,1], [1,1], [0,0], [0,0]], mode='CONSTANT')
- w = tf.add_n([w[1:, 1:], w[:-1, 1:], w[1:, :-1], w[:-1, :-1]]) * 0.25
- w = tf.cast(w, x.dtype)
- return tf.nn.conv2d(x, w, strides=[1,1,2,2], padding='SAME', data_format='NCHW')
-
-#----------------------------------------------------------------------------
-# Apply bias to the given activation tensor.
-
-def apply_bias(x, lrmul=1):
- b = tf.get_variable('bias', shape=[x.shape[1]], initializer=tf.initializers.zeros()) * lrmul
- b = tf.cast(b, x.dtype)
- if len(x.shape) == 2:
- return x + b
- return x + tf.reshape(b, [1, -1, 1, 1])
-
-#----------------------------------------------------------------------------
-# Leaky ReLU activation. More efficient than tf.nn.leaky_relu() and supports FP16.
-
-def leaky_relu(x, alpha=0.2):
- with tf.variable_scope('LeakyReLU'):
- alpha = tf.constant(alpha, dtype=x.dtype, name='alpha')
- @tf.custom_gradient
- def func(x):
- y = tf.maximum(x, x * alpha)
- @tf.custom_gradient
- def grad(dy):
- dx = tf.where(y >= 0, dy, dy * alpha)
- return dx, lambda ddx: tf.where(y >= 0, ddx, ddx * alpha)
- return y, grad
- return func(x)
-
-#----------------------------------------------------------------------------
-# Pixelwise feature vector normalization.
-
-def pixel_norm(x, epsilon=1e-8):
- with tf.variable_scope('PixelNorm'):
- epsilon = tf.constant(epsilon, dtype=x.dtype, name='epsilon')
- return x * tf.rsqrt(tf.reduce_mean(tf.square(x), axis=1, keepdims=True) + epsilon)
-
-#----------------------------------------------------------------------------
-# Instance normalization.
-
-def instance_norm(x, epsilon=1e-8):
- assert len(x.shape) == 4 # NCHW
- with tf.variable_scope('InstanceNorm'):
- orig_dtype = x.dtype
- x = tf.cast(x, tf.float32)
- x -= tf.reduce_mean(x, axis=[2,3], keepdims=True)
- epsilon = tf.constant(epsilon, dtype=x.dtype, name='epsilon')
- x *= tf.rsqrt(tf.reduce_mean(tf.square(x), axis=[2,3], keepdims=True) + epsilon)
- x = tf.cast(x, orig_dtype)
- return x
-
-#----------------------------------------------------------------------------
-# Style modulation.
-
-def style_mod(x, dlatent, **kwargs):
- with tf.variable_scope('StyleMod'):
- style = apply_bias(dense(dlatent, fmaps=x.shape[1]*2, gain=1, **kwargs))
- style = tf.reshape(style, [-1, 2, x.shape[1]] + [1] * (len(x.shape) - 2))
- return x * (style[:,0] + 1) + style[:,1]
-
-#----------------------------------------------------------------------------
-# Noise input.
-
-def apply_noise(x, noise_var=None, randomize_noise=True):
- assert len(x.shape) == 4 # NCHW
- with tf.variable_scope('Noise'):
- if noise_var is None or randomize_noise:
- noise = tf.random_normal([tf.shape(x)[0], 1, x.shape[2], x.shape[3]], dtype=x.dtype)
- else:
- noise = tf.cast(noise_var, x.dtype)
- weight = tf.get_variable('weight', shape=[x.shape[1].value], initializer=tf.initializers.zeros())
- return x + noise * tf.reshape(tf.cast(weight, x.dtype), [1, -1, 1, 1])
-
-#----------------------------------------------------------------------------
-# Minibatch standard deviation.
-
-def minibatch_stddev_layer(x, group_size=4, num_new_features=1):
- with tf.variable_scope('MinibatchStddev'):
- group_size = tf.minimum(group_size, tf.shape(x)[0]) # Minibatch must be divisible by (or smaller than) group_size.
- s = x.shape # [NCHW] Input shape.
- y = tf.reshape(x, [group_size, -1, num_new_features, s[1]//num_new_features, s[2], s[3]]) # [GMncHW] Split minibatch into M groups of size G. Split channels into n channel groups c.
- y = tf.cast(y, tf.float32) # [GMncHW] Cast to FP32.
- y -= tf.reduce_mean(y, axis=0, keepdims=True) # [GMncHW] Subtract mean over group.
- y = tf.reduce_mean(tf.square(y), axis=0) # [MncHW] Calc variance over group.
- y = tf.sqrt(y + 1e-8) # [MncHW] Calc stddev over group.
- y = tf.reduce_mean(y, axis=[2,3,4], keepdims=True) # [Mn111] Take average over fmaps and pixels.
- y = tf.reduce_mean(y, axis=[2]) # [Mn11] Split channels into c channel groups
- y = tf.cast(y, x.dtype) # [Mn11] Cast back to original data type.
- y = tf.tile(y, [group_size, 1, s[2], s[3]]) # [NnHW] Replicate over group and pixels.
- return tf.concat([x, y], axis=1) # [NCHW] Append as new fmap.
-
-#----------------------------------------------------------------------------
-# Style-based generator used in the StyleGAN paper.
-# Composed of two sub-networks (G_mapping and G_synthesis) that are defined below.
-
-def G_style(
- latents_in, # First input: Latent vectors (Z) [minibatch, latent_size].
- labels_in, # Second input: Conditioning labels [minibatch, label_size].
- truncation_psi = 0.7, # Style strength multiplier for the truncation trick. None = disable.
- truncation_cutoff = 8, # Number of layers for which to apply the truncation trick. None = disable.
- truncation_psi_val = None, # Value for truncation_psi to use during validation.
- truncation_cutoff_val = None, # Value for truncation_cutoff to use during validation.
- dlatent_avg_beta = 0.995, # Decay for tracking the moving average of W during training. None = disable.
- style_mixing_prob = 0.9, # Probability of mixing styles during training. None = disable.
- is_training = False, # Network is under training? Enables and disables specific features.
- is_validation = False, # Network is under validation? Chooses which value to use for truncation_psi.
- is_template_graph = False, # True = template graph constructed by the Network class, False = actual evaluation.
- components = dnnlib.EasyDict(), # Container for sub-networks. Retained between calls.
- **kwargs): # Arguments for sub-networks (G_mapping and G_synthesis).
-
- # Validate arguments.
- assert not is_training or not is_validation
- assert isinstance(components, dnnlib.EasyDict)
- if is_validation:
- truncation_psi = truncation_psi_val
- truncation_cutoff = truncation_cutoff_val
- if is_training or (truncation_psi is not None and not tflib.is_tf_expression(truncation_psi) and truncation_psi == 1):
- truncation_psi = None
- if is_training or (truncation_cutoff is not None and not tflib.is_tf_expression(truncation_cutoff) and truncation_cutoff <= 0):
- truncation_cutoff = None
- if not is_training or (dlatent_avg_beta is not None and not tflib.is_tf_expression(dlatent_avg_beta) and dlatent_avg_beta == 1):
- dlatent_avg_beta = None
- if not is_training or (style_mixing_prob is not None and not tflib.is_tf_expression(style_mixing_prob) and style_mixing_prob <= 0):
- style_mixing_prob = None
-
- # Setup components.
- if 'synthesis' not in components:
- components.synthesis = tflib.Network('G_synthesis', func_name=G_synthesis, **kwargs)
- num_layers = components.synthesis.input_shape[1]
- dlatent_size = components.synthesis.input_shape[2]
- if 'mapping' not in components:
- components.mapping = tflib.Network('G_mapping', func_name=G_mapping, dlatent_broadcast=num_layers, **kwargs)
-
- # Setup variables.
- lod_in = tf.get_variable('lod', initializer=np.float32(0), trainable=False)
- dlatent_avg = tf.get_variable('dlatent_avg', shape=[dlatent_size], initializer=tf.initializers.zeros(), trainable=False)
-
- # Evaluate mapping network.
- dlatents = components.mapping.get_output_for(latents_in, labels_in, **kwargs)
-
- # Update moving average of W.
- if dlatent_avg_beta is not None:
- with tf.variable_scope('DlatentAvg'):
- batch_avg = tf.reduce_mean(dlatents[:, 0], axis=0)
- update_op = tf.assign(dlatent_avg, tflib.lerp(batch_avg, dlatent_avg, dlatent_avg_beta))
- with tf.control_dependencies([update_op]):
- dlatents = tf.identity(dlatents)
-
- # Perform style mixing regularization.
- if style_mixing_prob is not None:
- with tf.name_scope('StyleMix'):
- latents2 = tf.random_normal(tf.shape(latents_in))
- dlatents2 = components.mapping.get_output_for(latents2, labels_in, **kwargs)
- layer_idx = np.arange(num_layers)[np.newaxis, :, np.newaxis]
- cur_layers = num_layers - tf.cast(lod_in, tf.int32) * 2
- mixing_cutoff = tf.cond(
- tf.random_uniform([], 0.0, 1.0) < style_mixing_prob,
- lambda: tf.random_uniform([], 1, cur_layers, dtype=tf.int32),
- lambda: cur_layers)
- dlatents = tf.where(tf.broadcast_to(layer_idx < mixing_cutoff, tf.shape(dlatents)), dlatents, dlatents2)
-
- # Apply truncation trick.
- if truncation_psi is not None and truncation_cutoff is not None:
- with tf.variable_scope('Truncation'):
- layer_idx = np.arange(num_layers)[np.newaxis, :, np.newaxis]
- ones = np.ones(layer_idx.shape, dtype=np.float32)
- coefs = tf.where(layer_idx < truncation_cutoff, truncation_psi * ones, ones)
- dlatents = tflib.lerp(dlatent_avg, dlatents, coefs)
-
- # Evaluate synthesis network.
- with tf.control_dependencies([tf.assign(components.synthesis.find_var('lod'), lod_in)]):
- images_out = components.synthesis.get_output_for(dlatents, force_clean_graph=is_template_graph, **kwargs)
- return tf.identity(images_out, name='images_out')
-
-#----------------------------------------------------------------------------
-# Mapping network used in the StyleGAN paper.
-
-def G_mapping(
- latents_in, # First input: Latent vectors (Z) [minibatch, latent_size].
- labels_in, # Second input: Conditioning labels [minibatch, label_size].
- latent_size = 512, # Latent vector (Z) dimensionality.
- label_size = 0, # Label dimensionality, 0 if no labels.
- dlatent_size = 512, # Disentangled latent (W) dimensionality.
- dlatent_broadcast = None, # Output disentangled latent (W) as [minibatch, dlatent_size] or [minibatch, dlatent_broadcast, dlatent_size].
- mapping_layers = 8, # Number of mapping layers.
- mapping_fmaps = 512, # Number of activations in the mapping layers.
- mapping_lrmul = 0.01, # Learning rate multiplier for the mapping layers.
- mapping_nonlinearity = 'lrelu', # Activation function: 'relu', 'lrelu'.
- use_wscale = True, # Enable equalized learning rate?
- normalize_latents = True, # Normalize latent vectors (Z) before feeding them to the mapping layers?
- dtype = 'float32', # Data type to use for activations and outputs.
- **_kwargs): # Ignore unrecognized keyword args.
-
- act, gain = {'relu': (tf.nn.relu, np.sqrt(2)), 'lrelu': (leaky_relu, np.sqrt(2))}[mapping_nonlinearity]
-
- # Inputs.
- latents_in.set_shape([None, latent_size])
- labels_in.set_shape([None, label_size])
- latents_in = tf.cast(latents_in, dtype)
- labels_in = tf.cast(labels_in, dtype)
- x = latents_in
-
- # Embed labels and concatenate them with latents.
- if label_size:
- with tf.variable_scope('LabelConcat'):
- w = tf.get_variable('weight', shape=[label_size, latent_size], initializer=tf.initializers.random_normal())
- y = tf.matmul(labels_in, tf.cast(w, dtype))
- x = tf.concat([x, y], axis=1)
-
- # Normalize latents.
- if normalize_latents:
- x = pixel_norm(x)
-
- # Mapping layers.
- for layer_idx in range(mapping_layers):
- with tf.variable_scope('Dense%d' % layer_idx):
- fmaps = dlatent_size if layer_idx == mapping_layers - 1 else mapping_fmaps
- x = dense(x, fmaps=fmaps, gain=gain, use_wscale=use_wscale, lrmul=mapping_lrmul)
- x = apply_bias(x, lrmul=mapping_lrmul)
- x = act(x)
-
- # Broadcast.
- if dlatent_broadcast is not None:
- with tf.variable_scope('Broadcast'):
- x = tf.tile(x[:, np.newaxis], [1, dlatent_broadcast, 1])
-
- # Output.
- assert x.dtype == tf.as_dtype(dtype)
- return tf.identity(x, name='dlatents_out')
-
-#----------------------------------------------------------------------------
-# Synthesis network used in the StyleGAN paper.
-
-def G_synthesis(
- dlatents_in, # Input: Disentangled latents (W) [minibatch, num_layers, dlatent_size].
- dlatent_size = 512, # Disentangled latent (W) dimensionality.
- num_channels = 3, # Number of output color channels.
- resolution = 1024, # Output resolution.
- fmap_base = 8192, # Overall multiplier for the number of feature maps.
- fmap_decay = 1.0, # log2 feature map reduction when doubling the resolution.
- fmap_max = 512, # Maximum number of feature maps in any layer.
- use_styles = True, # Enable style inputs?
- const_input_layer = True, # First layer is a learned constant?
- use_noise = True, # Enable noise inputs?
- randomize_noise = True, # True = randomize noise inputs every time (non-deterministic), False = read noise inputs from variables.
- nonlinearity = 'lrelu', # Activation function: 'relu', 'lrelu'
- use_wscale = True, # Enable equalized learning rate?
- use_pixel_norm = False, # Enable pixelwise feature vector normalization?
- use_instance_norm = True, # Enable instance normalization?
- dtype = 'float32', # Data type to use for activations and outputs.
- fused_scale = 'auto', # True = fused convolution + scaling, False = separate ops, 'auto' = decide automatically.
- blur_filter = [1,2,1], # Low-pass filter to apply when resampling activations. None = no filtering.
- structure = 'auto', # 'fixed' = no progressive growing, 'linear' = human-readable, 'recursive' = efficient, 'auto' = select automatically.
- is_template_graph = False, # True = template graph constructed by the Network class, False = actual evaluation.
- force_clean_graph = False, # True = construct a clean graph that looks nice in TensorBoard, False = default behavior.
- **_kwargs): # Ignore unrecognized keyword args.
-
- resolution_log2 = int(np.log2(resolution))
- assert resolution == 2**resolution_log2 and resolution >= 4
- def nf(stage): return min(int(fmap_base / (2.0 ** (stage * fmap_decay))), fmap_max)
- def blur(x): return blur2d(x, blur_filter) if blur_filter else x
- if is_template_graph: force_clean_graph = True
- if force_clean_graph: randomize_noise = False
- if structure == 'auto': structure = 'linear' if force_clean_graph else 'recursive'
- act, gain = {'relu': (tf.nn.relu, np.sqrt(2)), 'lrelu': (leaky_relu, np.sqrt(2))}[nonlinearity]
- num_layers = resolution_log2 * 2 - 2
- num_styles = num_layers if use_styles else 1
- images_out = None
-
- # Primary inputs.
- dlatents_in.set_shape([None, num_styles, dlatent_size])
- dlatents_in = tf.cast(dlatents_in, dtype)
- lod_in = tf.cast(tf.get_variable('lod', initializer=np.float32(0), trainable=False), dtype)
-
- # Noise inputs.
- noise_inputs = []
- if use_noise:
- for layer_idx in range(num_layers):
- res = layer_idx // 2 + 2
- shape = [1, use_noise, 2**res, 2**res]
- noise_inputs.append(tf.get_variable('noise%d' % layer_idx, shape=shape, initializer=tf.initializers.random_normal(), trainable=False))
-
- # Things to do at the end of each layer.
- def layer_epilogue(x, layer_idx):
- if use_noise:
- x = apply_noise(x, noise_inputs[layer_idx], randomize_noise=randomize_noise)
- x = apply_bias(x)
- x = act(x)
- if use_pixel_norm:
- x = pixel_norm(x)
- if use_instance_norm:
- x = instance_norm(x)
- if use_styles:
- x = style_mod(x, dlatents_in[:, layer_idx], use_wscale=use_wscale)
- return x
-
- # Early layers.
- with tf.variable_scope('4x4'):
- if const_input_layer:
- with tf.variable_scope('Const'):
- x = tf.get_variable('const', shape=[1, nf(1), 4, 4], initializer=tf.initializers.ones())
- x = layer_epilogue(tf.tile(tf.cast(x, dtype), [tf.shape(dlatents_in)[0], 1, 1, 1]), 0)
- else:
- with tf.variable_scope('Dense'):
- x = dense(dlatents_in[:, 0], fmaps=nf(1)*16, gain=gain/4, use_wscale=use_wscale) # tweak gain to match the official implementation of Progressing GAN
- x = layer_epilogue(tf.reshape(x, [-1, nf(1), 4, 4]), 0)
- with tf.variable_scope('Conv'):
- x = layer_epilogue(conv2d(x, fmaps=nf(1), kernel=3, gain=gain, use_wscale=use_wscale), 1)
-
- # Building blocks for remaining layers.
- def block(res, x): # res = 3..resolution_log2
- with tf.variable_scope('%dx%d' % (2**res, 2**res)):
- with tf.variable_scope('Conv0_up'):
- x = layer_epilogue(blur(upscale2d_conv2d(x, fmaps=nf(res-1), kernel=3, gain=gain, use_wscale=use_wscale, fused_scale=fused_scale)), res*2-4)
- with tf.variable_scope('Conv1'):
- x = layer_epilogue(conv2d(x, fmaps=nf(res-1), kernel=3, gain=gain, use_wscale=use_wscale), res*2-3)
- return x
- def torgb(res, x): # res = 2..resolution_log2
- lod = resolution_log2 - res
- with tf.variable_scope('ToRGB_lod%d' % lod):
- return apply_bias(conv2d(x, fmaps=num_channels, kernel=1, gain=1, use_wscale=use_wscale))
-
- # Fixed structure: simple and efficient, but does not support progressive growing.
- if structure == 'fixed':
- for res in range(3, resolution_log2 + 1):
- x = block(res, x)
- images_out = torgb(resolution_log2, x)
-
- # Linear structure: simple but inefficient.
- if structure == 'linear':
- images_out = torgb(2, x)
- for res in range(3, resolution_log2 + 1):
- lod = resolution_log2 - res
- x = block(res, x)
- img = torgb(res, x)
- images_out = upscale2d(images_out)
- with tf.variable_scope('Grow_lod%d' % lod):
- images_out = tflib.lerp_clip(img, images_out, lod_in - lod)
-
- # Recursive structure: complex but efficient.
- if structure == 'recursive':
- def cset(cur_lambda, new_cond, new_lambda):
- return lambda: tf.cond(new_cond, new_lambda, cur_lambda)
- def grow(x, res, lod):
- y = block(res, x)
- img = lambda: upscale2d(torgb(res, y), 2**lod)
- img = cset(img, (lod_in > lod), lambda: upscale2d(tflib.lerp(torgb(res, y), upscale2d(torgb(res - 1, x)), lod_in - lod), 2**lod))
- if lod > 0: img = cset(img, (lod_in < lod), lambda: grow(y, res + 1, lod - 1))
- return img()
- images_out = grow(x, 3, resolution_log2 - 3)
-
- assert images_out.dtype == tf.as_dtype(dtype)
- return tf.identity(images_out, name='images_out')
-
-#----------------------------------------------------------------------------
-# Discriminator used in the StyleGAN paper.
-
-def D_basic(
- images_in, # First input: Images [minibatch, channel, height, width].
- labels_in, # Second input: Labels [minibatch, label_size].
- num_channels = 1, # Number of input color channels. Overridden based on dataset.
- resolution = 32, # Input resolution. Overridden based on dataset.
- label_size = 0, # Dimensionality of the labels, 0 if no labels. Overridden based on dataset.
- fmap_base = 8192, # Overall multiplier for the number of feature maps.
- fmap_decay = 1.0, # log2 feature map reduction when doubling the resolution.
- fmap_max = 512, # Maximum number of feature maps in any layer.
- nonlinearity = 'lrelu', # Activation function: 'relu', 'lrelu',
- use_wscale = True, # Enable equalized learning rate?
- mbstd_group_size = 4, # Group size for the minibatch standard deviation layer, 0 = disable.
- mbstd_num_features = 1, # Number of features for the minibatch standard deviation layer.
- dtype = 'float32', # Data type to use for activations and outputs.
- fused_scale = 'auto', # True = fused convolution + scaling, False = separate ops, 'auto' = decide automatically.
- blur_filter = [1,2,1], # Low-pass filter to apply when resampling activations. None = no filtering.
- structure = 'auto', # 'fixed' = no progressive growing, 'linear' = human-readable, 'recursive' = efficient, 'auto' = select automatically.
- is_template_graph = False, # True = template graph constructed by the Network class, False = actual evaluation.
- **_kwargs): # Ignore unrecognized keyword args.
-
- resolution_log2 = int(np.log2(resolution))
- assert resolution == 2**resolution_log2 and resolution >= 4
- def nf(stage): return min(int(fmap_base / (2.0 ** (stage * fmap_decay))), fmap_max)
- def blur(x): return blur2d(x, blur_filter) if blur_filter else x
- if structure == 'auto': structure = 'linear' if is_template_graph else 'recursive'
- act, gain = {'relu': (tf.nn.relu, np.sqrt(2)), 'lrelu': (leaky_relu, np.sqrt(2))}[nonlinearity]
-
- images_in.set_shape([None, num_channels, resolution, resolution])
- labels_in.set_shape([None, label_size])
- images_in = tf.cast(images_in, dtype)
- labels_in = tf.cast(labels_in, dtype)
- lod_in = tf.cast(tf.get_variable('lod', initializer=np.float32(0.0), trainable=False), dtype)
- scores_out = None
-
- # Building blocks.
- def fromrgb(x, res): # res = 2..resolution_log2
- with tf.variable_scope('FromRGB_lod%d' % (resolution_log2 - res)):
- return act(apply_bias(conv2d(x, fmaps=nf(res-1), kernel=1, gain=gain, use_wscale=use_wscale)))
- def block(x, res): # res = 2..resolution_log2
- with tf.variable_scope('%dx%d' % (2**res, 2**res)):
- if res >= 3: # 8x8 and up
- with tf.variable_scope('Conv0'):
- x = act(apply_bias(conv2d(x, fmaps=nf(res-1), kernel=3, gain=gain, use_wscale=use_wscale)))
- with tf.variable_scope('Conv1_down'):
- x = act(apply_bias(conv2d_downscale2d(blur(x), fmaps=nf(res-2), kernel=3, gain=gain, use_wscale=use_wscale, fused_scale=fused_scale)))
- else: # 4x4
- if mbstd_group_size > 1:
- x = minibatch_stddev_layer(x, mbstd_group_size, mbstd_num_features)
- with tf.variable_scope('Conv'):
- x = act(apply_bias(conv2d(x, fmaps=nf(res-1), kernel=3, gain=gain, use_wscale=use_wscale)))
- with tf.variable_scope('Dense0'):
- x = act(apply_bias(dense(x, fmaps=nf(res-2), gain=gain, use_wscale=use_wscale)))
- with tf.variable_scope('Dense1'):
- x = apply_bias(dense(x, fmaps=max(label_size, 1), gain=1, use_wscale=use_wscale))
- return x
-
- # Fixed structure: simple and efficient, but does not support progressive growing.
- if structure == 'fixed':
- x = fromrgb(images_in, resolution_log2)
- for res in range(resolution_log2, 2, -1):
- x = block(x, res)
- scores_out = block(x, 2)
-
- # Linear structure: simple but inefficient.
- if structure == 'linear':
- img = images_in
- x = fromrgb(img, resolution_log2)
- for res in range(resolution_log2, 2, -1):
- lod = resolution_log2 - res
- x = block(x, res)
- img = downscale2d(img)
- y = fromrgb(img, res - 1)
- with tf.variable_scope('Grow_lod%d' % lod):
- x = tflib.lerp_clip(x, y, lod_in - lod)
- scores_out = block(x, 2)
-
- # Recursive structure: complex but efficient.
- if structure == 'recursive':
- def cset(cur_lambda, new_cond, new_lambda):
- return lambda: tf.cond(new_cond, new_lambda, cur_lambda)
- def grow(res, lod):
- x = lambda: fromrgb(downscale2d(images_in, 2**lod), res)
- if lod > 0: x = cset(x, (lod_in < lod), lambda: grow(res + 1, lod - 1))
- x = block(x(), res); y = lambda: x
- if res > 2: y = cset(y, (lod_in > lod), lambda: tflib.lerp(x, fromrgb(downscale2d(images_in, 2**(lod+1)), res - 1), lod_in - lod))
- return y()
- scores_out = grow(2, resolution_log2 - 2)
-
- # Label conditioning from "Which Training Methods for GANs do actually Converge?"
- if label_size:
- with tf.variable_scope('LabelSwitch'):
- scores_out = tf.reduce_sum(scores_out * labels_in, axis=1, keepdims=True)
-
- assert scores_out.dtype == tf.as_dtype(dtype)
- scores_out = tf.identity(scores_out, name='scores_out')
- return scores_out
-
-#----------------------------------------------------------------------------
diff --git a/lib/stylegan2/training/networks_stylegan2.py b/lib/stylegan2/training/networks_stylegan2.py
deleted file mode 100755
index 6c96fc19207c943da2fb2e4941816175299e79f8..0000000000000000000000000000000000000000
--- a/lib/stylegan2/training/networks_stylegan2.py
+++ /dev/null
@@ -1,697 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Network architectures used in the StyleGAN2 paper."""
-
-import numpy as np
-import tensorflow as tf
-import dnnlib
-import dnnlib.tflib as tflib
-from dnnlib.tflib.ops.upfirdn_2d import upsample_2d, downsample_2d, upsample_conv_2d, conv_downsample_2d
-from dnnlib.tflib.ops.fused_bias_act import fused_bias_act
-
-# NOTE: Do not import any application-specific modules here!
-# Specify all network parameters as kwargs.
-
-#----------------------------------------------------------------------------
-# Get/create weight tensor for a convolution or fully-connected layer.
-
-def get_weight(shape, gain=1, use_wscale=True, lrmul=1, weight_var='weight'):
- fan_in = np.prod(shape[:-1]) # [kernel, kernel, fmaps_in, fmaps_out] or [in, out]
- he_std = gain / np.sqrt(fan_in) # He init
-
- # Equalized learning rate and custom learning rate multiplier.
- if use_wscale:
- init_std = 1.0 / lrmul
- runtime_coef = he_std * lrmul
- else:
- init_std = he_std / lrmul
- runtime_coef = lrmul
-
- # Create variable.
- init = tf.initializers.random_normal(0, init_std)
- return tf.get_variable(weight_var, shape=shape, initializer=init) * runtime_coef
-
-#----------------------------------------------------------------------------
-# Fully-connected layer.
-
-def dense_layer(x, fmaps, gain=1, use_wscale=True, lrmul=1, weight_var='weight'):
- if len(x.shape) > 2:
- x = tf.reshape(x, [-1, np.prod([d.value for d in x.shape[1:]])])
- w = get_weight([x.shape[1].value, fmaps], gain=gain, use_wscale=use_wscale, lrmul=lrmul, weight_var=weight_var)
- w = tf.cast(w, x.dtype)
- return tf.matmul(x, w)
-
-#----------------------------------------------------------------------------
-# Convolution layer with optional upsampling or downsampling.
-
-def conv2d_layer(x, fmaps, kernel, up=False, down=False, resample_kernel=None, gain=1, use_wscale=True, lrmul=1, weight_var='weight'):
- assert not (up and down)
- assert kernel >= 1 and kernel % 2 == 1
- w = get_weight([kernel, kernel, x.shape[1].value, fmaps], gain=gain, use_wscale=use_wscale, lrmul=lrmul, weight_var=weight_var)
- if up:
- x = upsample_conv_2d(x, tf.cast(w, x.dtype), data_format='NCHW', k=resample_kernel)
- elif down:
- x = conv_downsample_2d(x, tf.cast(w, x.dtype), data_format='NCHW', k=resample_kernel)
- else:
- x = tf.nn.conv2d(x, tf.cast(w, x.dtype), data_format='NCHW', strides=[1,1,1,1], padding='SAME')
- return x
-
-#----------------------------------------------------------------------------
-# Apply bias and activation func.
-
-def apply_bias_act(x, act='linear', alpha=None, gain=None, lrmul=1, bias_var='bias'):
- b = tf.get_variable(bias_var, shape=[x.shape[1]], initializer=tf.initializers.zeros()) * lrmul
- return fused_bias_act(x, b=tf.cast(b, x.dtype), act=act, alpha=alpha, gain=gain)
-
-#----------------------------------------------------------------------------
-# Naive upsampling (nearest neighbor) and downsampling (average pooling).
-
-def naive_upsample_2d(x, factor=2):
- with tf.variable_scope('NaiveUpsample'):
- _N, C, H, W = x.shape.as_list()
- x = tf.reshape(x, [-1, C, H, 1, W, 1])
- x = tf.tile(x, [1, 1, 1, factor, 1, factor])
- return tf.reshape(x, [-1, C, H * factor, W * factor])
-
-def naive_downsample_2d(x, factor=2):
- with tf.variable_scope('NaiveDownsample'):
- _N, C, H, W = x.shape.as_list()
- x = tf.reshape(x, [-1, C, H // factor, factor, W // factor, factor])
- return tf.reduce_mean(x, axis=[3,5])
-
-#----------------------------------------------------------------------------
-# Modulated convolution layer.
-
-def modulated_conv2d_layer(x, y, fmaps, kernel, up=False, down=False, demodulate=True, resample_kernel=None, gain=1, use_wscale=True, lrmul=1, fused_modconv=True, weight_var='weight', mod_weight_var='mod_weight', mod_bias_var='mod_bias'):
- assert not (up and down)
- assert kernel >= 1 and kernel % 2 == 1
-
- # Get weight.
- w = get_weight([kernel, kernel, x.shape[1].value, fmaps], gain=gain, use_wscale=use_wscale, lrmul=lrmul, weight_var=weight_var)
- ww = w[np.newaxis] # [BkkIO] Introduce minibatch dimension.
-
- # Modulate.
- s = dense_layer(y, fmaps=x.shape[1].value, weight_var=mod_weight_var) # [BI] Transform incoming W to style.
- s = apply_bias_act(s, bias_var=mod_bias_var) + 1 # [BI] Add bias (initially 1).
- ww *= tf.cast(s[:, np.newaxis, np.newaxis, :, np.newaxis], w.dtype) # [BkkIO] Scale input feature maps.
-
- # Demodulate.
- if demodulate:
- d = tf.rsqrt(tf.reduce_sum(tf.square(ww), axis=[1,2,3]) + 1e-8) # [BO] Scaling factor.
- ww *= d[:, np.newaxis, np.newaxis, np.newaxis, :] # [BkkIO] Scale output feature maps.
-
- # Reshape/scale input.
- if fused_modconv:
- x = tf.reshape(x, [1, -1, x.shape[2], x.shape[3]]) # Fused => reshape minibatch to convolution groups.
- w = tf.reshape(tf.transpose(ww, [1, 2, 3, 0, 4]), [ww.shape[1], ww.shape[2], ww.shape[3], -1])
- else:
- x *= tf.cast(s[:, :, np.newaxis, np.newaxis], x.dtype) # [BIhw] Not fused => scale input activations.
-
- # Convolution with optional up/downsampling.
- if up:
- x = upsample_conv_2d(x, tf.cast(w, x.dtype), data_format='NCHW', k=resample_kernel)
- elif down:
- x = conv_downsample_2d(x, tf.cast(w, x.dtype), data_format='NCHW', k=resample_kernel)
- else:
- x = tf.nn.conv2d(x, tf.cast(w, x.dtype), data_format='NCHW', strides=[1,1,1,1], padding='SAME')
-
- # Reshape/scale output.
- if fused_modconv:
- x = tf.reshape(x, [-1, fmaps, x.shape[2], x.shape[3]]) # Fused => reshape convolution groups back to minibatch.
- elif demodulate:
- x *= tf.cast(d[:, :, np.newaxis, np.newaxis], x.dtype) # [BOhw] Not fused => scale output activations.
- return x
-
-#----------------------------------------------------------------------------
-# Minibatch standard deviation layer.
-
-def minibatch_stddev_layer(x, group_size=4, num_new_features=1):
- group_size = tf.minimum(group_size, tf.shape(x)[0]) # Minibatch must be divisible by (or smaller than) group_size.
- s = x.shape # [NCHW] Input shape.
- y = tf.reshape(x, [group_size, -1, num_new_features, s[1]//num_new_features, s[2], s[3]]) # [GMncHW] Split minibatch into M groups of size G. Split channels into n channel groups c.
- y = tf.cast(y, tf.float32) # [GMncHW] Cast to FP32.
- y -= tf.reduce_mean(y, axis=0, keepdims=True) # [GMncHW] Subtract mean over group.
- y = tf.reduce_mean(tf.square(y), axis=0) # [MncHW] Calc variance over group.
- y = tf.sqrt(y + 1e-8) # [MncHW] Calc stddev over group.
- y = tf.reduce_mean(y, axis=[2,3,4], keepdims=True) # [Mn111] Take average over fmaps and pixels.
- y = tf.reduce_mean(y, axis=[2]) # [Mn11] Split channels into c channel groups
- y = tf.cast(y, x.dtype) # [Mn11] Cast back to original data type.
- y = tf.tile(y, [group_size, 1, s[2], s[3]]) # [NnHW] Replicate over group and pixels.
- return tf.concat([x, y], axis=1) # [NCHW] Append as new fmap.
-
-#----------------------------------------------------------------------------
-# Main generator network.
-# Composed of two sub-networks (mapping and synthesis) that are defined below.
-# Used in configs B-F (Table 1).
-
-def G_main(
- latents_in, # First input: Latent vectors (Z) [minibatch, latent_size].
- labels_in, # Second input: Conditioning labels [minibatch, label_size].
- truncation_psi = 0.5, # Style strength multiplier for the truncation trick. None = disable.
- truncation_cutoff = None, # Number of layers for which to apply the truncation trick. None = disable.
- truncation_psi_val = None, # Value for truncation_psi to use during validation.
- truncation_cutoff_val = None, # Value for truncation_cutoff to use during validation.
- dlatent_avg_beta = 0.995, # Decay for tracking the moving average of W during training. None = disable.
- style_mixing_prob = 0.9, # Probability of mixing styles during training. None = disable.
- is_training = False, # Network is under training? Enables and disables specific features.
- is_validation = False, # Network is under validation? Chooses which value to use for truncation_psi.
- return_dlatents = False, # Return dlatents in addition to the images?
- is_template_graph = False, # True = template graph constructed by the Network class, False = actual evaluation.
- components = dnnlib.EasyDict(), # Container for sub-networks. Retained between calls.
- mapping_func = 'G_mapping', # Build func name for the mapping network.
- synthesis_func = 'G_synthesis_stylegan2', # Build func name for the synthesis network.
- **kwargs): # Arguments for sub-networks (mapping and synthesis).
-
- # Validate arguments.
- assert not is_training or not is_validation
- assert isinstance(components, dnnlib.EasyDict)
- if is_validation:
- truncation_psi = truncation_psi_val
- truncation_cutoff = truncation_cutoff_val
- if is_training or (truncation_psi is not None and not tflib.is_tf_expression(truncation_psi) and truncation_psi == 1):
- truncation_psi = None
- if is_training:
- truncation_cutoff = None
- if not is_training or (dlatent_avg_beta is not None and not tflib.is_tf_expression(dlatent_avg_beta) and dlatent_avg_beta == 1):
- dlatent_avg_beta = None
- if not is_training or (style_mixing_prob is not None and not tflib.is_tf_expression(style_mixing_prob) and style_mixing_prob <= 0):
- style_mixing_prob = None
-
- # Setup components.
- if 'synthesis' not in components:
- components.synthesis = tflib.Network('G_synthesis', func_name=globals()[synthesis_func], **kwargs)
- num_layers = components.synthesis.input_shape[1]
- dlatent_size = components.synthesis.input_shape[2]
- if 'mapping' not in components:
- components.mapping = tflib.Network('G_mapping', func_name=globals()[mapping_func], dlatent_broadcast=num_layers, **kwargs)
-
- # Setup variables.
- lod_in = tf.get_variable('lod', initializer=np.float32(0), trainable=False)
- dlatent_avg = tf.get_variable('dlatent_avg', shape=[dlatent_size], initializer=tf.initializers.zeros(), trainable=False)
-
- # Evaluate mapping network.
- dlatents = components.mapping.get_output_for(latents_in, labels_in, is_training=is_training, **kwargs)
- dlatents = tf.cast(dlatents, tf.float32)
-
- # Update moving average of W.
- if dlatent_avg_beta is not None:
- with tf.variable_scope('DlatentAvg'):
- batch_avg = tf.reduce_mean(dlatents[:, 0], axis=0)
- update_op = tf.assign(dlatent_avg, tflib.lerp(batch_avg, dlatent_avg, dlatent_avg_beta))
- with tf.control_dependencies([update_op]):
- dlatents = tf.identity(dlatents)
-
- # Perform style mixing regularization.
- if style_mixing_prob is not None:
- with tf.variable_scope('StyleMix'):
- latents2 = tf.random_normal(tf.shape(latents_in))
- dlatents2 = components.mapping.get_output_for(latents2, labels_in, is_training=is_training, **kwargs)
- dlatents2 = tf.cast(dlatents2, tf.float32)
- layer_idx = np.arange(num_layers)[np.newaxis, :, np.newaxis]
- cur_layers = num_layers - tf.cast(lod_in, tf.int32) * 2
- mixing_cutoff = tf.cond(
- tf.random_uniform([], 0.0, 1.0) < style_mixing_prob,
- lambda: tf.random_uniform([], 1, cur_layers, dtype=tf.int32),
- lambda: cur_layers)
- dlatents = tf.where(tf.broadcast_to(layer_idx < mixing_cutoff, tf.shape(dlatents)), dlatents, dlatents2)
-
- # Apply truncation trick.
- if truncation_psi is not None:
- with tf.variable_scope('Truncation'):
- layer_idx = np.arange(num_layers)[np.newaxis, :, np.newaxis]
- layer_psi = np.ones(layer_idx.shape, dtype=np.float32)
- if truncation_cutoff is None:
- layer_psi *= truncation_psi
- else:
- layer_psi = tf.where(layer_idx < truncation_cutoff, layer_psi * truncation_psi, layer_psi)
- dlatents = tflib.lerp(dlatent_avg, dlatents, layer_psi)
-
- # Evaluate synthesis network.
- deps = []
- if 'lod' in components.synthesis.vars:
- deps.append(tf.assign(components.synthesis.vars['lod'], lod_in))
- with tf.control_dependencies(deps):
- images_out = components.synthesis.get_output_for(dlatents, is_training=is_training, force_clean_graph=is_template_graph, **kwargs)
-
- # Return requested outputs.
- images_out = tf.identity(images_out, name='images_out')
- if return_dlatents:
- return images_out, dlatents
- return images_out
-
-#----------------------------------------------------------------------------
-# Mapping network.
-# Transforms the input latent code (z) to the disentangled latent code (w).
-# Used in configs B-F (Table 1).
-
-def G_mapping(
- latents_in, # First input: Latent vectors (Z) [minibatch, latent_size].
- labels_in, # Second input: Conditioning labels [minibatch, label_size].
- latent_size = 512, # Latent vector (Z) dimensionality.
- label_size = 0, # Label dimensionality, 0 if no labels.
- dlatent_size = 512, # Disentangled latent (W) dimensionality.
- dlatent_broadcast = None, # Output disentangled latent (W) as [minibatch, dlatent_size] or [minibatch, dlatent_broadcast, dlatent_size].
- mapping_layers = 8, # Number of mapping layers.
- mapping_fmaps = 512, # Number of activations in the mapping layers.
- mapping_lrmul = 0.01, # Learning rate multiplier for the mapping layers.
- mapping_nonlinearity = 'lrelu', # Activation function: 'relu', 'lrelu', etc.
- normalize_latents = True, # Normalize latent vectors (Z) before feeding them to the mapping layers?
- dtype = 'float32', # Data type to use for activations and outputs.
- **_kwargs): # Ignore unrecognized keyword args.
-
- act = mapping_nonlinearity
-
- # Inputs.
- latents_in.set_shape([None, latent_size])
- labels_in.set_shape([None, label_size])
- latents_in = tf.cast(latents_in, dtype)
- labels_in = tf.cast(labels_in, dtype)
- x = latents_in
-
- # Embed labels and concatenate them with latents.
- if label_size:
- with tf.variable_scope('LabelConcat'):
- w = tf.get_variable('weight', shape=[label_size, latent_size], initializer=tf.initializers.random_normal())
- y = tf.matmul(labels_in, tf.cast(w, dtype))
- x = tf.concat([x, y], axis=1)
-
- # Normalize latents.
- if normalize_latents:
- with tf.variable_scope('Normalize'):
- x *= tf.rsqrt(tf.reduce_mean(tf.square(x), axis=1, keepdims=True) + 1e-8)
-
- # Mapping layers.
- for layer_idx in range(mapping_layers):
- with tf.variable_scope('Dense%d' % layer_idx):
- fmaps = dlatent_size if layer_idx == mapping_layers - 1 else mapping_fmaps
- x = apply_bias_act(dense_layer(x, fmaps=fmaps, lrmul=mapping_lrmul), act=act, lrmul=mapping_lrmul)
-
- # Broadcast.
- if dlatent_broadcast is not None:
- with tf.variable_scope('Broadcast'):
- x = tf.tile(x[:, np.newaxis], [1, dlatent_broadcast, 1])
-
- # Output.
- assert x.dtype == tf.as_dtype(dtype)
- return tf.identity(x, name='dlatents_out')
-
-#----------------------------------------------------------------------------
-# StyleGAN synthesis network with revised architecture (Figure 2d).
-# Implements progressive growing, but no skip connections or residual nets (Figure 7).
-# Used in configs B-D (Table 1).
-
-def G_synthesis_stylegan_revised(
- dlatents_in, # Input: Disentangled latents (W) [minibatch, num_layers, dlatent_size].
- dlatent_size = 512, # Disentangled latent (W) dimensionality.
- num_channels = 3, # Number of output color channels.
- resolution = 1024, # Output resolution.
- fmap_base = 16 << 10, # Overall multiplier for the number of feature maps.
- fmap_decay = 1.0, # log2 feature map reduction when doubling the resolution.
- fmap_min = 1, # Minimum number of feature maps in any layer.
- fmap_max = 512, # Maximum number of feature maps in any layer.
- randomize_noise = True, # True = randomize noise inputs every time (non-deterministic), False = read noise inputs from variables.
- nonlinearity = 'lrelu', # Activation function: 'relu', 'lrelu', etc.
- dtype = 'float32', # Data type to use for activations and outputs.
- resample_kernel = [1,3,3,1], # Low-pass filter to apply when resampling activations. None = no filtering.
- fused_modconv = True, # Implement modulated_conv2d_layer() as a single fused op?
- structure = 'auto', # 'fixed' = no progressive growing, 'linear' = human-readable, 'recursive' = efficient, 'auto' = select automatically.
- is_template_graph = False, # True = template graph constructed by the Network class, False = actual evaluation.
- force_clean_graph = False, # True = construct a clean graph that looks nice in TensorBoard, False = default behavior.
- **_kwargs): # Ignore unrecognized keyword args.
-
- resolution_log2 = int(np.log2(resolution))
- assert resolution == 2**resolution_log2 and resolution >= 4
- def nf(stage): return np.clip(int(fmap_base / (2.0 ** (stage * fmap_decay))), fmap_min, fmap_max)
- if is_template_graph: force_clean_graph = True
- if force_clean_graph: randomize_noise = False
- if structure == 'auto': structure = 'linear' if force_clean_graph else 'recursive'
- act = nonlinearity
- num_layers = resolution_log2 * 2 - 2
- images_out = None
-
- # Primary inputs.
- dlatents_in.set_shape([None, num_layers, dlatent_size])
- dlatents_in = tf.cast(dlatents_in, dtype)
- lod_in = tf.cast(tf.get_variable('lod', initializer=np.float32(0), trainable=False), dtype)
-
- # Noise inputs.
- noise_inputs = []
- for layer_idx in range(num_layers - 1):
- res = (layer_idx + 5) // 2
- shape = [1, 1, 2**res, 2**res]
- noise_inputs.append(tf.get_variable('noise%d' % layer_idx, shape=shape, initializer=tf.initializers.random_normal(), trainable=False))
-
- # Single convolution layer with all the bells and whistles.
- def layer(x, layer_idx, fmaps, kernel, up=False):
- x = modulated_conv2d_layer(x, dlatents_in[:, layer_idx], fmaps=fmaps, kernel=kernel, up=up, resample_kernel=resample_kernel, fused_modconv=fused_modconv)
- if randomize_noise:
- noise = tf.random_normal([tf.shape(x)[0], 1, x.shape[2], x.shape[3]], dtype=x.dtype)
- else:
- noise = tf.cast(noise_inputs[layer_idx], x.dtype)
- noise_strength = tf.get_variable('noise_strength', shape=[], initializer=tf.initializers.zeros())
- x += noise * tf.cast(noise_strength, x.dtype)
- return apply_bias_act(x, act=act)
-
- # Early layers.
- with tf.variable_scope('4x4'):
- with tf.variable_scope('Const'):
- x = tf.get_variable('const', shape=[1, nf(1), 4, 4], initializer=tf.initializers.random_normal())
- x = tf.tile(tf.cast(x, dtype), [tf.shape(dlatents_in)[0], 1, 1, 1])
- with tf.variable_scope('Conv'):
- x = layer(x, layer_idx=0, fmaps=nf(1), kernel=3)
-
- # Building blocks for remaining layers.
- def block(res, x): # res = 3..resolution_log2
- with tf.variable_scope('%dx%d' % (2**res, 2**res)):
- with tf.variable_scope('Conv0_up'):
- x = layer(x, layer_idx=res*2-5, fmaps=nf(res-1), kernel=3, up=True)
- with tf.variable_scope('Conv1'):
- x = layer(x, layer_idx=res*2-4, fmaps=nf(res-1), kernel=3)
- return x
- def torgb(res, x): # res = 2..resolution_log2
- with tf.variable_scope('ToRGB_lod%d' % (resolution_log2 - res)):
- return apply_bias_act(modulated_conv2d_layer(x, dlatents_in[:, res*2-3], fmaps=num_channels, kernel=1, demodulate=False, fused_modconv=fused_modconv))
-
- # Fixed structure: simple and efficient, but does not support progressive growing.
- if structure == 'fixed':
- for res in range(3, resolution_log2 + 1):
- x = block(res, x)
- images_out = torgb(resolution_log2, x)
-
- # Linear structure: simple but inefficient.
- if structure == 'linear':
- images_out = torgb(2, x)
- for res in range(3, resolution_log2 + 1):
- lod = resolution_log2 - res
- x = block(res, x)
- img = torgb(res, x)
- with tf.variable_scope('Upsample_lod%d' % lod):
- images_out = upsample_2d(images_out)
- with tf.variable_scope('Grow_lod%d' % lod):
- images_out = tflib.lerp_clip(img, images_out, lod_in - lod)
-
- # Recursive structure: complex but efficient.
- if structure == 'recursive':
- def cset(cur_lambda, new_cond, new_lambda):
- return lambda: tf.cond(new_cond, new_lambda, cur_lambda)
- def grow(x, res, lod):
- y = block(res, x)
- img = lambda: naive_upsample_2d(torgb(res, y), factor=2**lod)
- img = cset(img, (lod_in > lod), lambda: naive_upsample_2d(tflib.lerp(torgb(res, y), upsample_2d(torgb(res - 1, x)), lod_in - lod), factor=2**lod))
- if lod > 0: img = cset(img, (lod_in < lod), lambda: grow(y, res + 1, lod - 1))
- return img()
- images_out = grow(x, 3, resolution_log2 - 3)
-
- assert images_out.dtype == tf.as_dtype(dtype)
- return tf.identity(images_out, name='images_out')
-
-#----------------------------------------------------------------------------
-# StyleGAN2 synthesis network (Figure 7).
-# Implements skip connections and residual nets (Figure 7), but no progressive growing.
-# Used in configs E-F (Table 1).
-
-def G_synthesis_stylegan2(
- dlatents_in, # Input: Disentangled latents (W) [minibatch, num_layers, dlatent_size].
- dlatent_size = 512, # Disentangled latent (W) dimensionality.
- num_channels = 3, # Number of output color channels.
- resolution = 1024, # Output resolution.
- fmap_base = 16 << 10, # Overall multiplier for the number of feature maps.
- fmap_decay = 1.0, # log2 feature map reduction when doubling the resolution.
- fmap_min = 1, # Minimum number of feature maps in any layer.
- fmap_max = 512, # Maximum number of feature maps in any layer.
- randomize_noise = True, # True = randomize noise inputs every time (non-deterministic), False = read noise inputs from variables.
- architecture = 'skip', # Architecture: 'orig', 'skip', 'resnet'.
- nonlinearity = 'lrelu', # Activation function: 'relu', 'lrelu', etc.
- dtype = 'float32', # Data type to use for activations and outputs.
- resample_kernel = [1,3,3,1], # Low-pass filter to apply when resampling activations. None = no filtering.
- fused_modconv = True, # Implement modulated_conv2d_layer() as a single fused op?
- **_kwargs): # Ignore unrecognized keyword args.
-
- resolution_log2 = int(np.log2(resolution))
- assert resolution == 2**resolution_log2 and resolution >= 4
- def nf(stage): return np.clip(int(fmap_base / (2.0 ** (stage * fmap_decay))), fmap_min, fmap_max)
- assert architecture in ['orig', 'skip', 'resnet']
- act = nonlinearity
- num_layers = resolution_log2 * 2 - 2
- images_out = None
-
- # Primary inputs.
- dlatents_in.set_shape([None, num_layers, dlatent_size])
- dlatents_in = tf.cast(dlatents_in, dtype)
-
- # Noise inputs.
- noise_inputs = []
- for layer_idx in range(num_layers - 1):
- res = (layer_idx + 5) // 2
- shape = [1, 1, 2**res, 2**res]
- noise_inputs.append(tf.get_variable('noise%d' % layer_idx, shape=shape, initializer=tf.initializers.random_normal(), trainable=False))
-
- # Single convolution layer with all the bells and whistles.
- def layer(x, layer_idx, fmaps, kernel, up=False):
- x = modulated_conv2d_layer(x, dlatents_in[:, layer_idx], fmaps=fmaps, kernel=kernel, up=up, resample_kernel=resample_kernel, fused_modconv=fused_modconv)
- if randomize_noise:
- noise = tf.random_normal([tf.shape(x)[0], 1, x.shape[2], x.shape[3]], dtype=x.dtype)
- else:
- noise = tf.cast(noise_inputs[layer_idx], x.dtype)
- noise_strength = tf.get_variable('noise_strength', shape=[], initializer=tf.initializers.zeros())
- x += noise * tf.cast(noise_strength, x.dtype)
- return apply_bias_act(x, act=act)
-
- # Building blocks for main layers.
- def block(x, res): # res = 3..resolution_log2
- t = x
- with tf.variable_scope('Conv0_up'):
- x = layer(x, layer_idx=res*2-5, fmaps=nf(res-1), kernel=3, up=True)
- with tf.variable_scope('Conv1'):
- x = layer(x, layer_idx=res*2-4, fmaps=nf(res-1), kernel=3)
- if architecture == 'resnet':
- with tf.variable_scope('Skip'):
- t = conv2d_layer(t, fmaps=nf(res-1), kernel=1, up=True, resample_kernel=resample_kernel)
- x = (x + t) * (1 / np.sqrt(2))
- return x
- def upsample(y):
- with tf.variable_scope('Upsample'):
- return upsample_2d(y, k=resample_kernel)
- def torgb(x, y, res): # res = 2..resolution_log2
- with tf.variable_scope('ToRGB'):
- t = apply_bias_act(modulated_conv2d_layer(x, dlatents_in[:, res*2-3], fmaps=num_channels, kernel=1, demodulate=False, fused_modconv=fused_modconv))
- return t if y is None else y + t
-
- # Early layers.
- y = None
- with tf.variable_scope('4x4'):
- with tf.variable_scope('Const'):
- x = tf.get_variable('const', shape=[1, nf(1), 4, 4], initializer=tf.initializers.random_normal())
- x = tf.tile(tf.cast(x, dtype), [tf.shape(dlatents_in)[0], 1, 1, 1])
- with tf.variable_scope('Conv'):
- x = layer(x, layer_idx=0, fmaps=nf(1), kernel=3)
- if architecture == 'skip':
- y = torgb(x, y, 2)
-
- # Main layers.
- for res in range(3, resolution_log2 + 1):
- with tf.variable_scope('%dx%d' % (2**res, 2**res)):
- x = block(x, res)
- if architecture == 'skip':
- y = upsample(y)
- if architecture == 'skip' or res == resolution_log2:
- y = torgb(x, y, res)
- images_out = y
-
- assert images_out.dtype == tf.as_dtype(dtype)
- return tf.identity(images_out, name='images_out')
-
-#----------------------------------------------------------------------------
-# Original StyleGAN discriminator.
-# Used in configs B-D (Table 1).
-
-def D_stylegan(
- images_in, # First input: Images [minibatch, channel, height, width].
- labels_in, # Second input: Labels [minibatch, label_size].
- num_channels = 3, # Number of input color channels. Overridden based on dataset.
- resolution = 1024, # Input resolution. Overridden based on dataset.
- label_size = 0, # Dimensionality of the labels, 0 if no labels. Overridden based on dataset.
- fmap_base = 16 << 10, # Overall multiplier for the number of feature maps.
- fmap_decay = 1.0, # log2 feature map reduction when doubling the resolution.
- fmap_min = 1, # Minimum number of feature maps in any layer.
- fmap_max = 512, # Maximum number of feature maps in any layer.
- nonlinearity = 'lrelu', # Activation function: 'relu', 'lrelu', etc.
- mbstd_group_size = 4, # Group size for the minibatch standard deviation layer, 0 = disable.
- mbstd_num_features = 1, # Number of features for the minibatch standard deviation layer.
- dtype = 'float32', # Data type to use for activations and outputs.
- resample_kernel = [1,3,3,1], # Low-pass filter to apply when resampling activations. None = no filtering.
- structure = 'auto', # 'fixed' = no progressive growing, 'linear' = human-readable, 'recursive' = efficient, 'auto' = select automatically.
- is_template_graph = False, # True = template graph constructed by the Network class, False = actual evaluation.
- **_kwargs): # Ignore unrecognized keyword args.
-
- resolution_log2 = int(np.log2(resolution))
- assert resolution == 2**resolution_log2 and resolution >= 4
- def nf(stage): return np.clip(int(fmap_base / (2.0 ** (stage * fmap_decay))), fmap_min, fmap_max)
- if structure == 'auto': structure = 'linear' if is_template_graph else 'recursive'
- act = nonlinearity
-
- images_in.set_shape([None, num_channels, resolution, resolution])
- labels_in.set_shape([None, label_size])
- images_in = tf.cast(images_in, dtype)
- labels_in = tf.cast(labels_in, dtype)
- lod_in = tf.cast(tf.get_variable('lod', initializer=np.float32(0.0), trainable=False), dtype)
-
- # Building blocks for spatial layers.
- def fromrgb(x, res): # res = 2..resolution_log2
- with tf.variable_scope('FromRGB_lod%d' % (resolution_log2 - res)):
- return apply_bias_act(conv2d_layer(x, fmaps=nf(res-1), kernel=1), act=act)
- def block(x, res): # res = 2..resolution_log2
- with tf.variable_scope('%dx%d' % (2**res, 2**res)):
- with tf.variable_scope('Conv0'):
- x = apply_bias_act(conv2d_layer(x, fmaps=nf(res-1), kernel=3), act=act)
- with tf.variable_scope('Conv1_down'):
- x = apply_bias_act(conv2d_layer(x, fmaps=nf(res-2), kernel=3, down=True, resample_kernel=resample_kernel), act=act)
- return x
-
- # Fixed structure: simple and efficient, but does not support progressive growing.
- if structure == 'fixed':
- x = fromrgb(images_in, resolution_log2)
- for res in range(resolution_log2, 2, -1):
- x = block(x, res)
-
- # Linear structure: simple but inefficient.
- if structure == 'linear':
- img = images_in
- x = fromrgb(img, resolution_log2)
- for res in range(resolution_log2, 2, -1):
- lod = resolution_log2 - res
- x = block(x, res)
- with tf.variable_scope('Downsample_lod%d' % lod):
- img = downsample_2d(img)
- y = fromrgb(img, res - 1)
- with tf.variable_scope('Grow_lod%d' % lod):
- x = tflib.lerp_clip(x, y, lod_in - lod)
-
- # Recursive structure: complex but efficient.
- if structure == 'recursive':
- def cset(cur_lambda, new_cond, new_lambda):
- return lambda: tf.cond(new_cond, new_lambda, cur_lambda)
- def grow(res, lod):
- x = lambda: fromrgb(naive_downsample_2d(images_in, factor=2**lod), res)
- if lod > 0: x = cset(x, (lod_in < lod), lambda: grow(res + 1, lod - 1))
- x = block(x(), res); y = lambda: x
- y = cset(y, (lod_in > lod), lambda: tflib.lerp(x, fromrgb(naive_downsample_2d(images_in, factor=2**(lod+1)), res - 1), lod_in - lod))
- return y()
- x = grow(3, resolution_log2 - 3)
-
- # Final layers at 4x4 resolution.
- with tf.variable_scope('4x4'):
- if mbstd_group_size > 1:
- with tf.variable_scope('MinibatchStddev'):
- x = minibatch_stddev_layer(x, mbstd_group_size, mbstd_num_features)
- with tf.variable_scope('Conv'):
- x = apply_bias_act(conv2d_layer(x, fmaps=nf(1), kernel=3), act=act)
- with tf.variable_scope('Dense0'):
- x = apply_bias_act(dense_layer(x, fmaps=nf(0)), act=act)
-
- # Output layer with label conditioning from "Which Training Methods for GANs do actually Converge?"
- with tf.variable_scope('Output'):
- x = apply_bias_act(dense_layer(x, fmaps=max(labels_in.shape[1], 1)))
- if labels_in.shape[1] > 0:
- x = tf.reduce_sum(x * labels_in, axis=1, keepdims=True)
- scores_out = x
-
- # Output.
- assert scores_out.dtype == tf.as_dtype(dtype)
- scores_out = tf.identity(scores_out, name='scores_out')
- return scores_out
-
-#----------------------------------------------------------------------------
-# StyleGAN2 discriminator (Figure 7).
-# Implements skip connections and residual nets (Figure 7), but no progressive growing.
-# Used in configs E-F (Table 1).
-
-def D_stylegan2(
- images_in, # First input: Images [minibatch, channel, height, width].
- labels_in, # Second input: Labels [minibatch, label_size].
- num_channels = 3, # Number of input color channels. Overridden based on dataset.
- resolution = 1024, # Input resolution. Overridden based on dataset.
- label_size = 0, # Dimensionality of the labels, 0 if no labels. Overridden based on dataset.
- fmap_base = 16 << 10, # Overall multiplier for the number of feature maps.
- fmap_decay = 1.0, # log2 feature map reduction when doubling the resolution.
- fmap_min = 1, # Minimum number of feature maps in any layer.
- fmap_max = 512, # Maximum number of feature maps in any layer.
- architecture = 'resnet', # Architecture: 'orig', 'skip', 'resnet'.
- nonlinearity = 'lrelu', # Activation function: 'relu', 'lrelu', etc.
- mbstd_group_size = 4, # Group size for the minibatch standard deviation layer, 0 = disable.
- mbstd_num_features = 1, # Number of features for the minibatch standard deviation layer.
- dtype = 'float32', # Data type to use for activations and outputs.
- resample_kernel = [1,3,3,1], # Low-pass filter to apply when resampling activations. None = no filtering.
- **_kwargs): # Ignore unrecognized keyword args.
-
- resolution_log2 = int(np.log2(resolution))
- assert resolution == 2**resolution_log2 and resolution >= 4
- def nf(stage): return np.clip(int(fmap_base / (2.0 ** (stage * fmap_decay))), fmap_min, fmap_max)
- assert architecture in ['orig', 'skip', 'resnet']
- act = nonlinearity
-
- images_in.set_shape([None, num_channels, resolution, resolution])
- labels_in.set_shape([None, label_size])
- images_in = tf.cast(images_in, dtype)
- labels_in = tf.cast(labels_in, dtype)
-
- # Building blocks for main layers.
- def fromrgb(x, y, res): # res = 2..resolution_log2
- with tf.variable_scope('FromRGB'):
- t = apply_bias_act(conv2d_layer(y, fmaps=nf(res-1), kernel=1), act=act)
- return t if x is None else x + t
- def block(x, res): # res = 2..resolution_log2
- t = x
- with tf.variable_scope('Conv0'):
- x = apply_bias_act(conv2d_layer(x, fmaps=nf(res-1), kernel=3), act=act)
- with tf.variable_scope('Conv1_down'):
- x = apply_bias_act(conv2d_layer(x, fmaps=nf(res-2), kernel=3, down=True, resample_kernel=resample_kernel), act=act)
- if architecture == 'resnet':
- with tf.variable_scope('Skip'):
- t = conv2d_layer(t, fmaps=nf(res-2), kernel=1, down=True, resample_kernel=resample_kernel)
- x = (x + t) * (1 / np.sqrt(2))
- return x
- def downsample(y):
- with tf.variable_scope('Downsample'):
- return downsample_2d(y, k=resample_kernel)
-
- # Main layers.
- x = None
- y = images_in
- for res in range(resolution_log2, 2, -1):
- with tf.variable_scope('%dx%d' % (2**res, 2**res)):
- if architecture == 'skip' or res == resolution_log2:
- x = fromrgb(x, y, res)
- x = block(x, res)
- if architecture == 'skip':
- y = downsample(y)
-
- # Final layers.
- with tf.variable_scope('4x4'):
- if architecture == 'skip':
- x = fromrgb(x, y, 2)
- if mbstd_group_size > 1:
- with tf.variable_scope('MinibatchStddev'):
- x = minibatch_stddev_layer(x, mbstd_group_size, mbstd_num_features)
- with tf.variable_scope('Conv'):
- x = apply_bias_act(conv2d_layer(x, fmaps=nf(1), kernel=3), act=act)
- with tf.variable_scope('Dense0'):
- x = apply_bias_act(dense_layer(x, fmaps=nf(0)), act=act)
-
- # Output layer with label conditioning from "Which Training Methods for GANs do actually Converge?"
- with tf.variable_scope('Output'):
- x = apply_bias_act(dense_layer(x, fmaps=max(labels_in.shape[1], 1)))
- if labels_in.shape[1] > 0:
- x = tf.reduce_sum(x * labels_in, axis=1, keepdims=True)
- scores_out = x
-
- # Output.
- assert scores_out.dtype == tf.as_dtype(dtype)
- scores_out = tf.identity(scores_out, name='scores_out')
- return scores_out
-
-#----------------------------------------------------------------------------
diff --git a/lib/stylegan2/training/training_loop.py b/lib/stylegan2/training/training_loop.py
deleted file mode 100755
index 942d1657c6ec310ba2210879c5399937ac9a355f..0000000000000000000000000000000000000000
--- a/lib/stylegan2/training/training_loop.py
+++ /dev/null
@@ -1,356 +0,0 @@
-# Copyright (c) 2019, NVIDIA Corporation. All rights reserved.
-#
-# This work is made available under the Nvidia Source Code License-NC.
-# To view a copy of this license, visit
-# https://nvlabs.github.io/stylegan2/license.html
-
-"""Main training script."""
-
-import numpy as np
-import tensorflow as tf
-import dnnlib
-import dnnlib.tflib as tflib
-from dnnlib.tflib.autosummary import autosummary
-
-from training import dataset
-from training import misc
-from metrics import metric_base
-
-#----------------------------------------------------------------------------
-# Just-in-time processing of training images before feeding them to the networks.
-
-def process_reals(x, labels, lod, mirror_augment, drange_data, drange_net):
- with tf.name_scope('DynamicRange'):
- x = tf.cast(x, tf.float32)
- x = misc.adjust_dynamic_range(x, drange_data, drange_net)
- if mirror_augment:
- with tf.name_scope('MirrorAugment'):
- x = tf.where(tf.random_uniform([tf.shape(x)[0]]) < 0.5, x, tf.reverse(x, [3]))
- with tf.name_scope('FadeLOD'): # Smooth crossfade between consecutive levels-of-detail.
- s = tf.shape(x)
- y = tf.reshape(x, [-1, s[1], s[2]//2, 2, s[3]//2, 2])
- y = tf.reduce_mean(y, axis=[3, 5], keepdims=True)
- y = tf.tile(y, [1, 1, 1, 2, 1, 2])
- y = tf.reshape(y, [-1, s[1], s[2], s[3]])
- x = tflib.lerp(x, y, lod - tf.floor(lod))
- with tf.name_scope('UpscaleLOD'): # Upscale to match the expected input/output size of the networks.
- s = tf.shape(x)
- factor = tf.cast(2 ** tf.floor(lod), tf.int32)
- x = tf.reshape(x, [-1, s[1], s[2], 1, s[3], 1])
- x = tf.tile(x, [1, 1, 1, factor, 1, factor])
- x = tf.reshape(x, [-1, s[1], s[2] * factor, s[3] * factor])
- return x, labels
-
-#----------------------------------------------------------------------------
-# Evaluate time-varying training parameters.
-
-def training_schedule(
- cur_nimg,
- training_set,
- lod_initial_resolution = None, # Image resolution used at the beginning.
- lod_training_kimg = 600, # Thousands of real images to show before doubling the resolution.
- lod_transition_kimg = 600, # Thousands of real images to show when fading in new layers.
- minibatch_size_base = 32, # Global minibatch size.
- minibatch_size_dict = {}, # Resolution-specific overrides.
- minibatch_gpu_base = 4, # Number of samples processed at a time by one GPU.
- minibatch_gpu_dict = {}, # Resolution-specific overrides.
- G_lrate_base = 0.002, # Learning rate for the generator.
- G_lrate_dict = {}, # Resolution-specific overrides.
- D_lrate_base = 0.002, # Learning rate for the discriminator.
- D_lrate_dict = {}, # Resolution-specific overrides.
- lrate_rampup_kimg = 0, # Duration of learning rate ramp-up.
- tick_kimg_base = 4, # Default interval of progress snapshots.
- tick_kimg_dict = {8:28, 16:24, 32:20, 64:16, 128:12, 256:8, 512:6, 1024:4}): # Resolution-specific overrides.
-
- # Initialize result dict.
- s = dnnlib.EasyDict()
- s.kimg = cur_nimg / 1000.0
-
- # Training phase.
- phase_dur = lod_training_kimg + lod_transition_kimg
- phase_idx = int(np.floor(s.kimg / phase_dur)) if phase_dur > 0 else 0
- phase_kimg = s.kimg - phase_idx * phase_dur
-
- # Level-of-detail and resolution.
- if lod_initial_resolution is None:
- s.lod = 0.0
- else:
- s.lod = training_set.resolution_log2
- s.lod -= np.floor(np.log2(lod_initial_resolution))
- s.lod -= phase_idx
- if lod_transition_kimg > 0:
- s.lod -= max(phase_kimg - lod_training_kimg, 0.0) / lod_transition_kimg
- s.lod = max(s.lod, 0.0)
- s.resolution = 2 ** (training_set.resolution_log2 - int(np.floor(s.lod)))
-
- # Minibatch size.
- s.minibatch_size = minibatch_size_dict.get(s.resolution, minibatch_size_base)
- s.minibatch_gpu = minibatch_gpu_dict.get(s.resolution, minibatch_gpu_base)
-
- # Learning rate.
- s.G_lrate = G_lrate_dict.get(s.resolution, G_lrate_base)
- s.D_lrate = D_lrate_dict.get(s.resolution, D_lrate_base)
- if lrate_rampup_kimg > 0:
- rampup = min(s.kimg / lrate_rampup_kimg, 1.0)
- s.G_lrate *= rampup
- s.D_lrate *= rampup
-
- # Other parameters.
- s.tick_kimg = tick_kimg_dict.get(s.resolution, tick_kimg_base)
- return s
-
-#----------------------------------------------------------------------------
-# Main training script.
-
-def training_loop(
- G_args = {}, # Options for generator network.
- D_args = {}, # Options for discriminator network.
- G_opt_args = {}, # Options for generator optimizer.
- D_opt_args = {}, # Options for discriminator optimizer.
- G_loss_args = {}, # Options for generator loss.
- D_loss_args = {}, # Options for discriminator loss.
- dataset_args = {}, # Options for dataset.load_dataset().
- sched_args = {}, # Options for train.TrainingSchedule.
- grid_args = {}, # Options for train.setup_snapshot_image_grid().
- metric_arg_list = [], # Options for MetricGroup.
- tf_config = {}, # Options for tflib.init_tf().
- data_dir = None, # Directory to load datasets from.
- G_smoothing_kimg = 10.0, # Half-life of the running average of generator weights.
- minibatch_repeats = 4, # Number of minibatches to run before adjusting training parameters.
- lazy_regularization = True, # Perform regularization as a separate training step?
- G_reg_interval = 4, # How often the perform regularization for G? Ignored if lazy_regularization=False.
- D_reg_interval = 16, # How often the perform regularization for D? Ignored if lazy_regularization=False.
- reset_opt_for_new_lod = True, # Reset optimizer internal state (e.g. Adam moments) when new layers are introduced?
- total_kimg = 25000, # Total length of the training, measured in thousands of real images.
- mirror_augment = False, # Enable mirror augment?
- drange_net = [-1,1], # Dynamic range used when feeding image data to the networks.
- image_snapshot_ticks = 50, # How often to save image snapshots? None = only save 'reals.png' and 'fakes-init.png'.
- network_snapshot_ticks = 50, # How often to save network snapshots? None = only save 'networks-final.pkl'.
- save_tf_graph = False, # Include full TensorFlow computation graph in the tfevents file?
- save_weight_histograms = False, # Include weight histograms in the tfevents file?
- resume_pkl = None, # Network pickle to resume training from, None = train from scratch.
- resume_kimg = 0.0, # Assumed training progress at the beginning. Affects reporting and training schedule.
- resume_time = 0.0, # Assumed wallclock time at the beginning. Affects reporting.
- resume_with_new_nets = False): # Construct new networks according to G_args and D_args before resuming training?
-
- # Initialize dnnlib and TensorFlow.
- tflib.init_tf(tf_config)
- num_gpus = dnnlib.submit_config.num_gpus
-
- # Load training set.
- training_set = dataset.load_dataset(data_dir=dnnlib.convert_path(data_dir), verbose=True, **dataset_args)
- grid_size, grid_reals, grid_labels = misc.setup_snapshot_image_grid(training_set, **grid_args)
- misc.save_image_grid(grid_reals, dnnlib.make_run_dir_path('reals.png'), drange=training_set.dynamic_range, grid_size=grid_size)
-
- # Construct or load networks.
- with tf.device('/gpu:0'):
- if resume_pkl is None or resume_with_new_nets:
- print('Constructing networks...')
- G = tflib.Network('G', num_channels=training_set.shape[0], resolution=training_set.shape[1], label_size=training_set.label_size, **G_args)
- D = tflib.Network('D', num_channels=training_set.shape[0], resolution=training_set.shape[1], label_size=training_set.label_size, **D_args)
- Gs = G.clone('Gs')
- if resume_pkl is not None:
- print('Loading networks from "%s"...' % resume_pkl)
- rG, rD, rGs = misc.load_pkl(resume_pkl)
- if resume_with_new_nets: G.copy_vars_from(rG); D.copy_vars_from(rD); Gs.copy_vars_from(rGs)
- else: G = rG; D = rD; Gs = rGs
-
- # Print layers and generate initial image snapshot.
- G.print_layers(); D.print_layers()
- sched = training_schedule(cur_nimg=total_kimg*1000, training_set=training_set, **sched_args)
- grid_latents = np.random.randn(np.prod(grid_size), *G.input_shape[1:])
- grid_fakes = Gs.run(grid_latents, grid_labels, is_validation=True, minibatch_size=sched.minibatch_gpu)
- misc.save_image_grid(grid_fakes, dnnlib.make_run_dir_path('fakes_init.png'), drange=drange_net, grid_size=grid_size)
-
- # Setup training inputs.
- print('Building TensorFlow graph...')
- with tf.name_scope('Inputs'), tf.device('/cpu:0'):
- lod_in = tf.placeholder(tf.float32, name='lod_in', shape=[])
- lrate_in = tf.placeholder(tf.float32, name='lrate_in', shape=[])
- minibatch_size_in = tf.placeholder(tf.int32, name='minibatch_size_in', shape=[])
- minibatch_gpu_in = tf.placeholder(tf.int32, name='minibatch_gpu_in', shape=[])
- minibatch_multiplier = minibatch_size_in // (minibatch_gpu_in * num_gpus)
- Gs_beta = 0.5 ** tf.div(tf.cast(minibatch_size_in, tf.float32), G_smoothing_kimg * 1000.0) if G_smoothing_kimg > 0.0 else 0.0
-
- # Setup optimizers.
- G_opt_args = dict(G_opt_args)
- D_opt_args = dict(D_opt_args)
- for args, reg_interval in [(G_opt_args, G_reg_interval), (D_opt_args, D_reg_interval)]:
- args['minibatch_multiplier'] = minibatch_multiplier
- args['learning_rate'] = lrate_in
- if lazy_regularization:
- mb_ratio = reg_interval / (reg_interval + 1)
- args['learning_rate'] *= mb_ratio
- if 'beta1' in args: args['beta1'] **= mb_ratio
- if 'beta2' in args: args['beta2'] **= mb_ratio
- G_opt = tflib.Optimizer(name='TrainG', **G_opt_args)
- D_opt = tflib.Optimizer(name='TrainD', **D_opt_args)
- G_reg_opt = tflib.Optimizer(name='RegG', share=G_opt, **G_opt_args)
- D_reg_opt = tflib.Optimizer(name='RegD', share=D_opt, **D_opt_args)
-
- # Build training graph for each GPU.
- data_fetch_ops = []
- for gpu in range(num_gpus):
- with tf.name_scope('GPU%d' % gpu), tf.device('/gpu:%d' % gpu):
-
- # Create GPU-specific shadow copies of G and D.
- G_gpu = G if gpu == 0 else G.clone(G.name + '_shadow')
- D_gpu = D if gpu == 0 else D.clone(D.name + '_shadow')
-
- # Fetch training data via temporary variables.
- with tf.name_scope('DataFetch'):
- sched = training_schedule(cur_nimg=int(resume_kimg*1000), training_set=training_set, **sched_args)
- reals_var = tf.Variable(name='reals', trainable=False, initial_value=tf.zeros([sched.minibatch_gpu] + training_set.shape))
- labels_var = tf.Variable(name='labels', trainable=False, initial_value=tf.zeros([sched.minibatch_gpu, training_set.label_size]))
- reals_write, labels_write = training_set.get_minibatch_tf()
- reals_write, labels_write = process_reals(reals_write, labels_write, lod_in, mirror_augment, training_set.dynamic_range, drange_net)
- reals_write = tf.concat([reals_write, reals_var[minibatch_gpu_in:]], axis=0)
- labels_write = tf.concat([labels_write, labels_var[minibatch_gpu_in:]], axis=0)
- data_fetch_ops += [tf.assign(reals_var, reals_write)]
- data_fetch_ops += [tf.assign(labels_var, labels_write)]
- reals_read = reals_var[:minibatch_gpu_in]
- labels_read = labels_var[:minibatch_gpu_in]
-
- # Evaluate loss functions.
- lod_assign_ops = []
- if 'lod' in G_gpu.vars: lod_assign_ops += [tf.assign(G_gpu.vars['lod'], lod_in)]
- if 'lod' in D_gpu.vars: lod_assign_ops += [tf.assign(D_gpu.vars['lod'], lod_in)]
- with tf.control_dependencies(lod_assign_ops):
- with tf.name_scope('G_loss'):
- G_loss, G_reg = dnnlib.util.call_func_by_name(G=G_gpu, D=D_gpu, opt=G_opt, training_set=training_set, minibatch_size=minibatch_gpu_in, **G_loss_args)
- with tf.name_scope('D_loss'):
- D_loss, D_reg = dnnlib.util.call_func_by_name(G=G_gpu, D=D_gpu, opt=D_opt, training_set=training_set, minibatch_size=minibatch_gpu_in, reals=reals_read, labels=labels_read, **D_loss_args)
-
- # Register gradients.
- if not lazy_regularization:
- if G_reg is not None: G_loss += G_reg
- if D_reg is not None: D_loss += D_reg
- else:
- if G_reg is not None: G_reg_opt.register_gradients(tf.reduce_mean(G_reg * G_reg_interval), G_gpu.trainables)
- if D_reg is not None: D_reg_opt.register_gradients(tf.reduce_mean(D_reg * D_reg_interval), D_gpu.trainables)
- G_opt.register_gradients(tf.reduce_mean(G_loss), G_gpu.trainables)
- D_opt.register_gradients(tf.reduce_mean(D_loss), D_gpu.trainables)
-
- # Setup training ops.
- data_fetch_op = tf.group(*data_fetch_ops)
- G_train_op = G_opt.apply_updates()
- D_train_op = D_opt.apply_updates()
- G_reg_op = G_reg_opt.apply_updates(allow_no_op=True)
- D_reg_op = D_reg_opt.apply_updates(allow_no_op=True)
- Gs_update_op = Gs.setup_as_moving_average_of(G, beta=Gs_beta)
-
- # Finalize graph.
- with tf.device('/gpu:0'):
- try:
- peak_gpu_mem_op = tf.contrib.memory_stats.MaxBytesInUse()
- except tf.errors.NotFoundError:
- peak_gpu_mem_op = tf.constant(0)
- tflib.init_uninitialized_vars()
-
- print('Initializing logs...')
- summary_log = tf.summary.FileWriter(dnnlib.make_run_dir_path())
- if save_tf_graph:
- summary_log.add_graph(tf.get_default_graph())
- if save_weight_histograms:
- G.setup_weight_histograms(); D.setup_weight_histograms()
- metrics = metric_base.MetricGroup(metric_arg_list)
-
- print('Training for %d kimg...\n' % total_kimg)
- dnnlib.RunContext.get().update('', cur_epoch=resume_kimg, max_epoch=total_kimg)
- maintenance_time = dnnlib.RunContext.get().get_last_update_interval()
- cur_nimg = int(resume_kimg * 1000)
- cur_tick = -1
- tick_start_nimg = cur_nimg
- prev_lod = -1.0
- running_mb_counter = 0
- while cur_nimg < total_kimg * 1000:
- if dnnlib.RunContext.get().should_stop(): break
-
- # Choose training parameters and configure training ops.
- sched = training_schedule(cur_nimg=cur_nimg, training_set=training_set, **sched_args)
- assert sched.minibatch_size % (sched.minibatch_gpu * num_gpus) == 0
- training_set.configure(sched.minibatch_gpu, sched.lod)
- if reset_opt_for_new_lod:
- if np.floor(sched.lod) != np.floor(prev_lod) or np.ceil(sched.lod) != np.ceil(prev_lod):
- G_opt.reset_optimizer_state(); D_opt.reset_optimizer_state()
- prev_lod = sched.lod
-
- # Run training ops.
- feed_dict = {lod_in: sched.lod, lrate_in: sched.G_lrate, minibatch_size_in: sched.minibatch_size, minibatch_gpu_in: sched.minibatch_gpu}
- for _repeat in range(minibatch_repeats):
- rounds = range(0, sched.minibatch_size, sched.minibatch_gpu * num_gpus)
- run_G_reg = (lazy_regularization and running_mb_counter % G_reg_interval == 0)
- run_D_reg = (lazy_regularization and running_mb_counter % D_reg_interval == 0)
- cur_nimg += sched.minibatch_size
- running_mb_counter += 1
-
- # Fast path without gradient accumulation.
- if len(rounds) == 1:
- tflib.run([G_train_op, data_fetch_op], feed_dict)
- if run_G_reg:
- tflib.run(G_reg_op, feed_dict)
- tflib.run([D_train_op, Gs_update_op], feed_dict)
- if run_D_reg:
- tflib.run(D_reg_op, feed_dict)
-
- # Slow path with gradient accumulation.
- else:
- for _round in rounds:
- tflib.run(G_train_op, feed_dict)
- if run_G_reg:
- for _round in rounds:
- tflib.run(G_reg_op, feed_dict)
- tflib.run(Gs_update_op, feed_dict)
- for _round in rounds:
- tflib.run(data_fetch_op, feed_dict)
- tflib.run(D_train_op, feed_dict)
- if run_D_reg:
- for _round in rounds:
- tflib.run(D_reg_op, feed_dict)
-
- # Perform maintenance tasks once per tick.
- done = (cur_nimg >= total_kimg * 1000)
- if cur_tick < 0 or cur_nimg >= tick_start_nimg + sched.tick_kimg * 1000 or done:
- cur_tick += 1
- tick_kimg = (cur_nimg - tick_start_nimg) / 1000.0
- tick_start_nimg = cur_nimg
- tick_time = dnnlib.RunContext.get().get_time_since_last_update()
- total_time = dnnlib.RunContext.get().get_time_since_start() + resume_time
-
- # Report progress.
- print('tick %-5d kimg %-8.1f lod %-5.2f minibatch %-4d time %-12s sec/tick %-7.1f sec/kimg %-7.2f maintenance %-6.1f gpumem %.1f' % (
- autosummary('Progress/tick', cur_tick),
- autosummary('Progress/kimg', cur_nimg / 1000.0),
- autosummary('Progress/lod', sched.lod),
- autosummary('Progress/minibatch', sched.minibatch_size),
- dnnlib.util.format_time(autosummary('Timing/total_sec', total_time)),
- autosummary('Timing/sec_per_tick', tick_time),
- autosummary('Timing/sec_per_kimg', tick_time / tick_kimg),
- autosummary('Timing/maintenance_sec', maintenance_time),
- autosummary('Resources/peak_gpu_mem_gb', peak_gpu_mem_op.eval() / 2**30)))
- autosummary('Timing/total_hours', total_time / (60.0 * 60.0))
- autosummary('Timing/total_days', total_time / (24.0 * 60.0 * 60.0))
-
- # Save snapshots.
- if image_snapshot_ticks is not None and (cur_tick % image_snapshot_ticks == 0 or done):
- grid_fakes = Gs.run(grid_latents, grid_labels, is_validation=True, minibatch_size=sched.minibatch_gpu)
- misc.save_image_grid(grid_fakes, dnnlib.make_run_dir_path('fakes%06d.png' % (cur_nimg // 1000)), drange=drange_net, grid_size=grid_size)
- if network_snapshot_ticks is not None and (cur_tick % network_snapshot_ticks == 0 or done):
- pkl = dnnlib.make_run_dir_path('network-snapshot-%06d.pkl' % (cur_nimg // 1000))
- misc.save_pkl((G, D, Gs), pkl)
- metrics.run(pkl, run_dir=dnnlib.make_run_dir_path(), data_dir=dnnlib.convert_path(data_dir), num_gpus=num_gpus, tf_config=tf_config)
-
- # Update summaries and RunContext.
- metrics.update_autosummaries()
- tflib.autosummary.save_summaries(summary_log, cur_nimg)
- dnnlib.RunContext.get().update('%.2f' % sched.lod, cur_epoch=cur_nimg // 1000, max_epoch=total_kimg)
- maintenance_time = dnnlib.RunContext.get().get_last_update_interval() - tick_time
-
- # Save final snapshot.
- misc.save_pkl((G, D, Gs), dnnlib.make_run_dir_path('network-final.pkl'))
-
- # All done.
- summary_log.close()
- training_set.close()
-
-#----------------------------------------------------------------------------
diff --git a/lib/utils/__init__.py b/lib/utils/__init__.py
deleted file mode 100644
index bad63140a6f1927a9253f58b06b1c08c96c0683a..0000000000000000000000000000000000000000
--- a/lib/utils/__init__.py
+++ /dev/null
@@ -1,92 +0,0 @@
-from pkgutil import extend_path
-__path__ = extend_path(__path__, __name__)
-import numpy as np
-import os
-from ..stylegan2.dnnlib import tflib
-
-def fix_randomness(seed=None):
- config = {'rnd.np_random_seed': seed,
- 'rnd.tf_random_seed': 'auto'}
- tflib.init_tf(config)
-
-def adjust_dynamic_range(image, target_range, dtype):
- minval = np.min(image)
- maxval = np.max(image)
- return ((target_range[0]*(maxval - image) + target_range[1]*(image-minval))/(maxval - minval)).astype(dtype)
-
-def normalize(v):
- """
- Return a normlaized version of the input vector
- """
- return v/np.linalg.norm(v)
-
-def lerp(p0, p1, n, start=0.0, end=1.0):
- """
- Linear interpolation between two points
- Inputs:
- p0, p1: Rank-1 numpy vectors with same dimension D
- n: int, total number of points to return
- start, end : control the range of the interpolation, i.e. the range of interpolation parameter t.
- Interpolated points are computed as (1-t)*p0 + t*p1 where t can takes linearly spaced values
- in the range [start, end] (included).
-
- Returns:
- p : Numpy vector of shape (n, D) containing all interpolated points
- """
- t = np.linspace(start, end, n)[:, np.newaxis]
- p = (1-t) * p0[np.newaxis, :] + t * p1[np.newaxis, :]
- return p
-
-def logerp(p0, p1, n, start=-10.0, end=0.0):
- """
- Logarithmic interpolation between two points
- Inputs:
- p0, p1: Rank-1 numpy vectors with same dimension D
- n: int, total number of points to return
- start, end : control the range of the interpolation, i.e. the range of interpolation parameter t.
- Interpolated points are computed as (1-t)*p0 + t*p1 where t can takes logarithmically spaced values
- in the range [10^start, 10^end] (included).
-
- Returns:
- p : Numpy vector of shape (n, D) containing all interpolated points
- """
- t = np.logspace(start, end, n)[:, np.newaxis]
- p = (1-t) * p0[np.newaxis, :] + t * p1[np.newaxis, :]
- return p
-
-def rand_cos_dist(v, cos_dist):
- """
- From an input vector `v` and a target cosine distance `cos_dist`,
- returns a random unit vector w with same dimension as `v` such that
- cosine_distance(`v`, `w`) = `cos_dist`.
- """
- costheta = 1-cos_dist
- # Form the unit vector parallel to v:
- u = v / np.linalg.norm(v)
- latent_dim = v.shape[0]
-
- # Pick a random vector:
- r = np.random.randn(latent_dim)
-
- # Form a vector perpendicular to v:
- uperp = r - r.dot(u)*u
-
- # Make it a unit vector:
- uperp = uperp / np.linalg.norm(uperp)
-
- # w is the linear combination of u and uperp with coefficients costheta
- # and sin(theta) = sqrt(1 - costheta**2), respectively:
- w = costheta*u + np.sqrt(1 - costheta**2)*uperp
-
- return w
-
-def get_task():
-
- if 'SGE_TASK_ID' in os.environ and os.environ['SGE_TASK_ID'] != 'undefined':
- current_task = int(os.environ['SGE_TASK_ID']) - 1
- num_tasks = int(os.environ['SGE_TASK_LAST'])
- else:
- current_task = 0
- num_tasks = 1
-
- return current_task, num_tasks
diff --git a/lib/utils/plotting.py b/lib/utils/plotting.py
deleted file mode 100644
index efffdcaa2436ddfc926c70cbb62567c1eff92467..0000000000000000000000000000000000000000
--- a/lib/utils/plotting.py
+++ /dev/null
@@ -1,27 +0,0 @@
-import matplotlib.pyplot as plt
-import bob.io.image
-import numpy as np
-
-def facegrid(face_images, nrows, ncols, figsize=None, labels=None):
- """
- Produces an image grid of size (nrows, ncols) showing each image
- contained in the input `face_images` list.
- An optional `labels` list can also be provided, in which case the `labels` will be
- used as title for each subplot in the grid.
- """
- if figsize is None:
- figsize = (2*ncols, 2*nrows)
- if labels is None:
- labels = [None]*len(face_images)
- fig, ax = plt.subplots(nrows=nrows, ncols=ncols, squeeze=False, figsize=figsize, tight_layout=True)
- for i, (face, label) in enumerate(zip(face_images, labels)):
- currax = ax[i//ncols, i%ncols]
- if face is not None:
- currax.imshow(bob.io.image.to_matplotlib(face))
- currax.axis('off')
- if label is not None:
- currax.set_title(label)
-
- return fig, ax
-
-
diff --git a/project_db.py b/project_db.py
deleted file mode 100644
index a4e3e825ba611d78da8f2e6ef0151fe41c112304..0000000000000000000000000000000000000000
--- a/project_db.py
+++ /dev/null
@@ -1,157 +0,0 @@
-import os
-import click
-
-import bob.io.base
-import bob.io.image
-from bob.extension import rc
-
-from lib.utils import get_task, fix_randomness
-from lib.stylegan2.preprocessor import FFHQCropper
-from lib.stylegan2.generator import StyleGAN2Generator
-from lib.stylegan2.projector import Projector
-
-from bob.extension.scripts.click_helper import (
- ResourceOption,
- ConfigCommand,
-)
-
-from lib.utils import fix_randomness
-
-CROP_DIR = "img_aligned"
-PROJ_DIR = "projected"
-LAT_DIR = "w_latents"
-
-
-def is_processed(f, output_dir):
- return os.path.exists(f.make_path(os.path.join(output_dir, LAT_DIR), ".h5"))
-
-
-@click.command(
- cls=ConfigCommand,
- help="Project a bob.bio.base.database.BioDatabase into the StyleGAN2 latent space.",
-)
-@click.option(
- "--database",
- "-d",
- entry_point_group="bob.bio.database",
- required=True,
- cls=ResourceOption,
- help="BioDatabase of face images that should be projected",
-)
-@click.option(
- "--output-dir",
- "-o",
- default=rc['bob.synface.multipie_projections'],
- cls=ResourceOption,
- help="Root directory of the output files",
-)
-@click.option(
- "--num_steps",
- "-n",
- type=int,
- default=1000,
- cls=ResourceOption,
- help="Number of steps during projection. Lower values will decrease total projection runtime but could also lead to lesser quality latents. Keeping the default is the safest option.",
-)
-@click.option(
- "--group",
- "-g",
- default="world",
- cls=ResourceOption,
- help="Either `world`, `dev` or `eval`. By default, project the `world` set",
-)
-@click.option(
- "--checkpoint",
- "-c",
- is_flag=True,
- cls=ResourceOption,
- help="Activate flag to checkpoint cropped faces and projected faces",
-)
-@click.option(
- "--force",
- "-f",
- is_flag=True,
- cls=ResourceOption,
- help="`Activate flag to overwrite computed latent if they already exist",
-)
-@click.option(
- "--seed",
- "-s",
- type=int,
- cls=ResourceOption,
- help="Seed to control stochasticity during projection.",
-)
-def project(
- database,
- output_dir,
- group="world",
- num_steps=1000,
- checkpoint=False,
- force=False,
- seed=None,
- **kwargs
-):
- failure_file_path = os.path.join(output_dir, "failure.dat")
- fix_randomness(seed=seed)
-
- cropper = FFHQCropper()
- generator = StyleGAN2Generator()
- projector = Projector(num_steps=num_steps)
- projector.set_network(generator.network)
-
- task_id, num_tasks = get_task()
-
- if group == "world":
- files = database.training_files()
- else:
- files = database.test_files(group)
-
- if not force:
- files = [f for f in files if not is_processed(f, output_dir)]
-
- if os.path.exists(failure_file_path):
- with open(failure_file_path, "r") as failure_file:
- fail_cases = [item.rstrip() for item in failure_file]
-
- files = [f for f in files if f.path not in fail_cases]
-
- subfiles = files[task_id :: num_tasks]
- print("{} files remaining. Handling {} of them".format(len(files), len(subfiles)))
-
- for i, f in enumerate(subfiles):
- print("{} : {}".format(i, f))
- image = f.load(database.original_directory, database.original_extension)
-
- try:
- cropped = cropper(image)
- except:
- with open(failure_file_path, "a") as failure_file:
- failure_file.write(f.path + "\n")
- print("Failure to crop {} !".format(f))
- break
-
- if checkpoint:
- bob.io.base.save(
- cropped,
- f.make_path(os.path.join(output_dir, CROP_DIR), ".png"),
- create_directories=True,
- )
-
- projected = projector(cropped)
-
- if checkpoint:
- bob.io.base.save(
- projected.image,
- f.make_path(os.path.join(output_dir, PROJ_DIR), ".png"),
- create_directories=True,
- )
-
- bob.io.base.save(
- projected.w_latent,
- f.make_path(os.path.join(output_dir, LAT_DIR), ".h5"),
- create_directories=True,
- )
-
-
-if __name__ == "__main__":
- project()
diff --git a/requirements.txt b/requirements.txt
deleted file mode 100644
index 089f1abe678a25e3bdd9f06da16152d3b03da86f..0000000000000000000000000000000000000000
--- a/requirements.txt
+++ /dev/null
@@ -1,3 +0,0 @@
-setuptools
-numpy
-bob.extension
\ No newline at end of file
diff --git a/setup.py b/setup.py
deleted file mode 100644
index bb8afb3c045007dea9c6b096548a8b4f81096729..0000000000000000000000000000000000000000
--- a/setup.py
+++ /dev/null
@@ -1,65 +0,0 @@
-#!/usr/bin/env python
-# -*- coding: utf-8 -*-
-
-from setuptools import setup, dist
-dist.Distribution(dict(setup_requires=['bob.extension']))
-
-from bob.extension.utils import find_packages
-from bob.extension.utils import load_requirements
-
-install_requires = load_requirements()
-
-
-setup(
-
- name='bob.paper.synthetic_dataset',
- version=open("version.txt").read().rstrip(),
- description='New package',
-
- url='https://gitlab.idiap.ch/bob/bob.paper.ijcb2021.synthetic_dataset',
- license='GPLv3',
-
- # there may be multiple authors (separate entries by comma)
- author='Laurent Colbois',
- author_email='laurent.colbois@idiap.ch',
-
- # there may be a maintainer apart from the author - you decide
- #maintainer='?',
- #maintainer_email='email@example.com',
-
- # you may add more keywords separating those by commas (a, b, c, ...)
- keywords = "bob",
-
- long_description=open('README.rst').read(),
-
- # leave this here, it is pretty standard
- packages=find_packages(),
- include_package_data=True,
- zip_safe = False,
-
- install_requires=install_requires,
-
- entry_points={
- # add entry points (scripts, bob resources here, if any)
- 'console_scripts': [
- ],
-
-
- },
-
- # check classifiers, add and remove as you see fit
- # full list here: https://pypi.org/classifiers/
- # don't remove the Bob framework unless it's not a bob package
- classifiers = [
- 'Framework :: Bob',
- 'Development Status :: 4 - Beta',
- 'Intended Audience :: Science/Research',
- 'License :: OSI Approved :: GNU General Public License v3 (GPLv3)',
- 'Natural Language :: English',
- 'Programming Language :: Python',
- 'Programming Language :: Python :: 3',
- 'Topic :: Scientific/Engineering :: Artificial Intelligence',
- 'Topic :: Software Development :: Libraries :: Python Modules',
- ],
-
-)
\ No newline at end of file
diff --git a/version.txt b/version.txt
deleted file mode 100644
index 79d48fc4015cc8015db8c42188f074088e94ee37..0000000000000000000000000000000000000000
--- a/version.txt
+++ /dev/null
@@ -1 +0,0 @@
-0.0.1b0
\ No newline at end of file