Scitator
commented
3 years ago could you please do the following as a hotfix:
import os
from torch import nn, optim
from torch.utils.data import DataLoader
from catalyst import dl, utils
from catalyst.data.transforms import ToTensor
from catalyst.contrib.datasets import MNIST
from catalyst.data.loader import BatchPrefetchLoaderWrapper
model = nn.Sequential(nn.Flatten(), nn.Linear(28 * 28, 10))
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=0.02)
loaders = {
"train": DataLoader(
MNIST(os.getcwd(), train=True, download=True, transform=ToTensor()), batch_size=32
),
"valid": DataLoader(
MNIST(os.getcwd(), train=False, download=True, transform=ToTensor()), batch_size=32
),
}
for key in loaders.keys():
loaders[key] = BatchPrefetchLoaderWrapper(
loaders[key], num_prefetches=32)
class CustomSupervisedRunner(dl.SupervisedRunner):
def get_loaders(self, stage):
return self._loaders
runner = CustomSupervisedRunner(
input_key="features", output_key="logits", target_key="targets", loss_key="loss"
)
# model training
runner.train(
model=model,
criterion=criterion,
optimizer=optimizer,
loaders=loaders,
num_epochs=1,
callbacks=[
dl.AccuracyCallback(input_key="logits", target_key="targets", topk_args=(1, 3, 5)),
dl.PrecisionRecallF1SupportCallback(
input_key="logits", target_key="targets", num_classes=10
),
dl.AUCCallback(input_key="logits", target_key="targets"),
# catalyst[ml] required ``pip install catalyst[ml]``
# dl.ConfusionMatrixCallback(input_key="logits", target_key="targets", num_classes=10),
],
logdir="./logs",
valid_loader="valid",
valid_metric="loss",
minimize_valid_metric=True,
verbose=True,
load_best_on_end=True,
)? We will drop a correct solution next release ;)
🐛 Bug Report
loaders = get_loaders_from_params(initial_seed=initial_seed, **loaders)seems does not work withBatchPrefetchLoaderWrapperHow To Reproduce
Steps to reproduce the behavior: run the following code sample
Code sample
Screenshots
Expected behavior
The code should be training.
Environment
Catalyst version: 21.03.2 PyTorch version: 1.8.1 Is debug build: No CUDA used to build PyTorch: 11.1 TensorFlow version: N/A TensorBoard version: N/A
OS: linux GCC version: (GCC) 4.8.5 20150623 (Red Hat 4.8.5-39) CMake version: Could not collect
Python version: 3.8 Is CUDA available: Yes CUDA runtime version: Could not collect GPU models and configuration: GPU 0: GeForce RTX 2080 Ti Nvidia driver version: 455.32.00 cuDNN version: Could not collect
Versions of relevant libraries: [pip3] catalyst==21.3.2 [pip3] numpy==1.20.2 [pip3] tensorboardX==2.1 [pip3] torch==1.8.1 [pip3] torchaudio==0.8.0a0+e4e171a [pip3] torchfile==0.1.0 [pip3] torchnet==0.0.4 [pip3] torchvision==0.9.1 [conda] blas 1.0 mkl
[conda] catalyst 21.3.2 pypi_0 pypi [conda] cudatoolkit 11.1.1 h6406543_8 conda-forge [conda] ffmpeg 4.3 hf484d3e_0 pytorch [conda] mkl 2020.4 h726a3e6_304 conda-forge [conda] mkl-service 2.3.0 py38h1e0a361_2 conda-forge [conda] mkl_fft 1.3.0 py38h5c078b8_1 conda-forge [conda] mkl_random 1.2.0 py38hc5bc63f_1 conda-forge [conda] numpy 1.20.2 pypi_0 pypi [conda] pytorch 1.8.1 py3.8_cuda11.1_cudnn8.0.5_0 pytorch [conda] tensorboardx 2.1 pypi_0 pypi [conda] torchaudio 0.8.1 py38 pytorch [conda] torchfile 0.1.0 pypi_0 pypi [conda] torchnet 0.0.4 pypi_0 pypi [conda] torchvision 0.9.1 py38_cu111 pytorch
Additional context
Checklist