Skip to content
Snippets Groups Projects
Commit 4ac50d6a authored by Anjith GEORGE's avatar Anjith GEORGE
Browse files

fix

parent 38a5586f
Branches
Tags
1 merge request!18MCCNN trainer
...@@ -13,7 +13,7 @@ logger = bob.core.log.setup("bob.learn.pytorch") ...@@ -13,7 +13,7 @@ logger = bob.core.log.setup("bob.learn.pytorch")
import time import time
import os import os
"""
#TODO: #TODO:
#0. Add class balancing #0. Add class balancing
#1. Logging to tensorboardX or a simpler logger #1. Logging to tensorboardX or a simpler logger
...@@ -21,8 +21,9 @@ import os ...@@ -21,8 +21,9 @@ import os
#3. Use to(device) instead of .cuda()? #3. Use to(device) instead of .cuda()?
#4. Functionality to select channels from the dataloader: may be move this to the datafolder class #4. Functionality to select channels from the dataloader: may be move this to the datafolder class
#5. Moving more arguments to config? #5. Moving more arguments to config?
#6. Implement the selection of channels as a transform
"""
def comp_bce_loss_weights(target): def comp_bce_loss_weights(target):
""" """
Compute the balancing weights for the BCE loss function. Compute the balancing weights for the BCE loss function.
...@@ -64,7 +65,7 @@ class MCCNNTrainer(object): ...@@ -64,7 +65,7 @@ class MCCNNTrainer(object):
The network to train The network to train
batch_size: int batch_size: int
The size of your minibatch The size of your minibatch
use_gpu: boolean use_gpu: bool
If you would like to use the gpu If you would like to use the gpu
verbosity_level: int verbosity_level: int
The level of verbosity output to stdout The level of verbosity output to stdout
...@@ -80,7 +81,7 @@ class MCCNNTrainer(object): ...@@ -80,7 +81,7 @@ class MCCNNTrainer(object):
The network to train The network to train
batch_size: int batch_size: int
The size of your minibatch The size of your minibatch
use_gpu: boolean use_gpu: bool
If you would like to use the gpu If you would like to use the gpu
adapted_layers: str adapted_layers: str
The blocks in the CNN to adapt; only the ones listed are adapted in the training. The layers are separated by '-' in the The blocks in the CNN to adapt; only the ones listed are adapted in the training. The layers are separated by '-' in the
......
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment