tinytorchtrainer

Successor of pytorch-pretrained-cnns minus pytorch lightning

Scripts:

train.py: Training with various parameters
eval.py: Runs evaluation of a given checkpoint with validation dataset
eval_adv_rob.py: Runs evaluation of a given checkpoint with validation dataset under adversarial attacks of AutoAttack
plot_convolutions.py: Plots the 3x3 conv filters of a given checkpoint
plot_feature_maps.py: Plots the feature maps of selected layers of a given checkpoints for a given input sample
init_data.py: Prepares the data independent of training (usefull before sweeps)
ckpt_info.py: Dumps info stored in a checkpoint
to_onnx.py: Exports checkpoints to ONNX

Credits

Name		Name	Last commit message	Last commit date
Latest commit History 199 Commits
.github/workflows		.github/workflows
data		data
data_samples		data_samples
models		models
tests		tests
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
ckpt_info.py		ckpt_info.py
eval.py		eval.py
eval_adv_rob.py		eval_adv_rob.py
eval_cifar10_1.py		eval_cifar10_1.py
eval_cifar10_c.py		eval_cifar10_c.py
hubconf.py		hubconf.py
init_data.py		init_data.py
intermediate_fmaps.py		intermediate_fmaps.py
plot_convolutions.py		plot_convolutions.py
plot_feature_maps.py		plot_feature_maps.py
requirements.txt		requirements.txt
to_onnx.py		to_onnx.py
train.py		train.py
upload_checkpoints.py		upload_checkpoints.py
utils.py		utils.py