From 9e721cb3d9d38ee998e27cd86a5ba63258a28d2b Mon Sep 17 00:00:00 2001
From: Andre Anjos <andre.dos.anjos@gmail.com>
Date: Sat, 1 Jul 2023 19:08:15 +0200
Subject: [PATCH] [models] Remove runtime checks; Use torchvision normaliser
instead of our own
---
src/ptbench/models/densenet.py | 15 ++----
src/ptbench/models/normalizer.py | 80 +++++++++++++++-----------------
src/ptbench/models/pasa.py | 13 ++----
3 files changed, 46 insertions(+), 62 deletions(-)
diff --git a/src/ptbench/models/densenet.py b/src/ptbench/models/densenet.py
index bc8b3a9d..1aa1b268 100644
--- a/src/ptbench/models/densenet.py
+++ b/src/ptbench/models/densenet.py
@@ -7,8 +7,6 @@ import torch
import torch.nn as nn
import torchvision.models as models
-from .normalizer import TorchVisionNormalizer
-
class Densenet(pl.LightningModule):
"""Densenet module.
@@ -31,7 +29,7 @@ class Densenet(pl.LightningModule):
self.name = "Densenet"
- self.normalizer = TorchVisionNormalizer(nb_channels=nb_channels)
+ self.normalizer = None
# Load pretrained model
weights = None if not pretrained else models.DenseNet121_Weights.DEFAULT
@@ -55,16 +53,13 @@ class Densenet(pl.LightningModule):
imagenet weights, during contruction).
"""
if self.pretrained:
- from .normalizer import TorchVisionNormalizer
+ from .normalizer import make_imagenet_normalizer
- self.normalizer = TorchVisionNormalizer(
- torch.Tensor([0.485, 0.456, 0.406]),
- torch.Tensor([0.229, 0.224, 0.225]),
- )
+ self.normalizer = make_imagenet_normalizer()
else:
- from .normalizer import get_znorm_normalizer
+ from .normalizer import make_z_normalizer
- self.normalizer = get_znorm_normalizer(dataloader)
+ self.normalizer = make_z_normalizer(dataloader)
def training_step(self, batch, batch_idx):
images = batch[1]
diff --git a/src/ptbench/models/normalizer.py b/src/ptbench/models/normalizer.py
index b9ba7eb3..2cc4b956 100644
--- a/src/ptbench/models/normalizer.py
+++ b/src/ptbench/models/normalizer.py
@@ -2,52 +2,23 @@
#
# SPDX-License-Identifier: GPL-3.0-or-later
-"""A network model that prefixes a z-normalization step to any other module."""
+"""A network model that prefixes a subtract/divide step to any other module."""
import torch
import torch.nn
import torch.utils.data
+import torchvision.transforms
-class TorchVisionNormalizer(torch.nn.Module):
- """A simple normalizer that applies the standard torchvision normalization.
+def make_z_normalizer(
+ dataloader: torch.utils.data.DataLoader,
+) -> torchvision.transforms.Normalize:
+ """Computes mean and standard deviation from a dataloader.
- This module does not learn.
+ This function will input a dataloader, and compute the mean and standard
+ deviation by image channel. It will work for both monochromatic, and color
+ inputs with 2, 3 or more color planes.
- Parameters
- ----------
-
- nb_channels : :py:class:`int`, Optional
- Number of images channels fed to the model
- """
-
- def __init__(self, subtract: torch.Tensor, divide: torch.Tensor):
- super().__init__()
- if len(subtract) != len(divide):
- raise ValueError(
- "Lengths of 'subtract' and 'divide' tensors should be the same."
- )
- if len(subtract) not in (1, 3):
- raise ValueError(
- "Length of 'subtract' tensor should be either 1 or 3, depending on the number of color channels."
- )
-
- subtract = torch.as_tensor(subtract)[None, :, None, None]
- divide = torch.as_tensor(divide)[None, :, None, None]
- self.register_buffer("subtract", subtract)
- self.register_buffer("divide", divide)
- self.name = "torchvision-normalizer"
-
- def forward(self, inputs: torch.Tensor):
- """inputs shape [batches, planes, height, width]"""
- return inputs.sub(self.subtract).div(self.divide)
-
-
-def get_znorm_normalizer(
- dataloader: torch.utils.data.DataLoader,
-) -> TorchVisionNormalizer:
- """Returns a normalizer with the mean and std computed from a dataloader's
- unaugmented training set.
Parameters
----------
@@ -55,15 +26,22 @@ def get_znorm_normalizer(
dataloader:
A torch Dataloader from which to compute the mean and std
+
Returns
-------
- An initialized TorchVisionNormalizer
+ An initialized normalizer
"""
- mean = 0.0
- var = 0.0
+ # Peek the number of channels of batches in the data loader
+ batch = next(iter(dataloader))
+ channels = batch[0].shape[1]
+
+ # Initialises accumulators
+ mean = torch.zeros(channels, dtype=batch[0].dtype)
+ var = torch.zeros(channels, dtype=batch[0].dtype)
num_images = 0
+ # Evaluates mean and standard deviation
for batch in dataloader:
data = batch[0]
data = data.view(data.size(0), data.size(1), -1)
@@ -76,5 +54,21 @@ def get_znorm_normalizer(
var /= num_images
std = torch.sqrt(var)
- normalizer = TorchVisionNormalizer(mean, std)
- return normalizer
+ return torchvision.transforms.Normalize(mean, std)
+
+
+def make_imagenet_normalizer() -> torchvision.transforms.Normalize:
+ """Returns the stock ImageNet normalisation weights from torchvision.
+
+ The weights are wrapped in a torch module. This normalizer only works for
+ **RGB (color) images**.
+
+
+ Returns
+ -------
+ An initialized normalizer
+ """
+
+ return torchvision.transforms.Normalize(
+ (0.485, 0.456, 0.406), (0.229, 0.224, 0.225)
+ )
diff --git a/src/ptbench/models/pasa.py b/src/ptbench/models/pasa.py
index cca494f1..61105dbe 100644
--- a/src/ptbench/models/pasa.py
+++ b/src/ptbench/models/pasa.py
@@ -79,12 +79,7 @@ class PASA(pl.LightningModule):
self.dense = nn.Linear(80, 1) # Fully connected layer
def forward(self, x):
- if self.normalizer is None:
- raise TypeError(
- "The normalizer has not been initialized. Make sure to call set_normalizer() after creation of the model."
- )
-
- x = self.normalizer(x)
+ x = self.normalizer(x) # type: ignore
# First convolution block
_x = x
@@ -140,11 +135,11 @@ class PASA(pl.LightningModule):
dataloader:
A torch Dataloader from which to compute the mean and std
"""
- from .normalizer import get_znorm_normalizer
+ from .normalizer import make_z_normalizer
- self.normalizer = get_znorm_normalizer(dataloader)
+ self.normalizer = make_z_normalizer(dataloader)
- def training_step(self, batch, batch_idx):
+ def training_step(self, batch, _):
images = batch[0]
labels = batch[1]["label"]
--
GitLab