DMFN
DMFN copied to clipboard
cuDNN error
While executing the training script. I encountered the following error.
Traceback (most recent call last):
File "train.py", line 72, in <module>
model.optimize_parameters()
File "/home/DMFN/models/inpainting_model.py", line 177, in optimize_parameters
l_g_total.backward()
File "/home/anaconda3/envs/dmfn/lib/python3.7/site-packages/torch/tensor.py", line 245, in backward
torch.autograd.backward(self, gradient, retain_graph, create_graph, inputs=inputs)
File "/home/anaconda3/envs/dmfn/lib/python3.7/site-packages/torch/autograd/__init__.py", line 147, in backward
allow_unreachable=True, accumulate_grad=True) # allow_unreachable flag
RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED
You can try to repro this exception using the following code snippet. If that doesn't trigger the error, please include your original repro script when reporting this issue.
import torch
torch.backends.cuda.matmul.allow_tf32 = True
torch.backends.cudnn.benchmark = False
torch.backends.cudnn.deterministic = False
torch.backends.cudnn.allow_tf32 = True
data = torch.randn([2, 256, 64, 64], dtype=torch.float, device='cuda', requires_grad=True)
net = torch.nn.Conv2d(256, 256, kernel_size=[3, 3], padding=[1, 1], stride=[1, 1], dilation=[1, 1], groups=1)
net = net.cuda().float()
out = net(data)
out.backward(torch.randn_like(out))
torch.cuda.synchronize()
ConvolutionParams
data_type = CUDNN_DATA_FLOAT
padding = [1, 1, 0]
stride = [1, 1, 0]
dilation = [1, 1, 0]
groups = 1
deterministic = false
allow_tf32 = true
input: TensorDescriptor 0x7fb8e80d6120
type = CUDNN_DATA_FLOAT
nbDims = 4
dimA = 2, 256, 64, 64,
strideA = 1048576, 4096, 64, 1,
output: TensorDescriptor 0x7fb8e80c8380
type = CUDNN_DATA_FLOAT
nbDims = 4
dimA = 2, 256, 64, 64,
strideA = 1048576, 4096, 64, 1,
weight: FilterDescriptor 0x7fb8e80d2500
type = CUDNN_DATA_FLOAT
tensor_format = CUDNN_TENSOR_NCHW
nbDims = 4
dimA = 256, 256, 3, 3,
Pointer addresses:
input: 0x7fb936000000
output: 0x7fb93a000000
weight: 0x7fb924e90200
Additional pointer addresses:
grad_output: 0x7fb93a000000
grad_weight: 0x7fb924e90200
Backward filter algorithm: 5
While execution of the suggested code snippet I did get any error or warnings.
After typing python -m torch.utils.collect_env
I got the following.
Collecting environment information...
PyTorch version: 1.8.0+cu111
Is debug build: False
CUDA used to build PyTorch: 11.1
ROCM used to build PyTorch: N/A
OS: Ubuntu 16.04.7 LTS (x86_64)
GCC version: (Ubuntu 9.4.0-1ubuntu1~16.04) 9.4.0
Clang version: Could not collect
CMake version: version 3.21.3
Python version: 3.7 (64-bit runtime)
Is CUDA available: True
CUDA runtime version: Could not collect
GPU models and configuration: GPU 0: GeForce GTX 1070
Nvidia driver version: 455.45.01
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.7.0.5
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn.so.5.1.10
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn.so.8.0.4
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8.0.4
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_adv_train.so.8.0.4
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8.0.4
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8.0.4
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8.0.4
/usr/local/cuda-11.1/targets/x86_64-linux/lib/libcudnn_ops_train.so.8.0.4
HIP runtime version: N/A
MIOpen runtime version: N/A
Versions of relevant libraries:
[pip3] efficientnet-pytorch==0.6.3
[pip3] facenet-pytorch==2.5.2
[pip3] numpy==1.19.5
[pip3] numpydoc==1.1.0
[pip3] pytorch-fid==0.1.1
[pip3] pytorch-ignite==0.4.7
[pip3] pytorch2keras==0.2.4
[pip3] pytorch3d==0.5.0
[pip3] segmentation-models-pytorch==0.1.3
[pip3] torch==1.8.0+cu111
[pip3] torch-geometric==1.7.2
[pip3] torch-model-archiver==0.4.0
[pip3] torch-scatter==2.0.7
[pip3] torch-sparse==0.6.10
[pip3] torch-workflow-archiver==0.1.0
[pip3] torchaudio==0.8.0
[pip3] torchserve==0.4.0
[pip3] torchvision==0.9.0+cu111
[conda] blas 1.0 mkl
[conda] efficientnet-pytorch 0.6.3 pypi_0 pypi
[conda] facenet-pytorch 2.5.2 pypi_0 pypi
[conda] mkl 2021.3.0 h06a4308_520
[conda] mkl-service 2.4.0 py37h7f8727e_0
[conda] mkl_fft 1.3.0 py37h42c9631_2
[conda] mkl_random 1.2.2 py37h51133e4_0
[conda] numpy 1.19.5 pypi_0 pypi
[conda] numpydoc 1.1.0 pyhd3eb1b0_1
[conda] pytorch-fid 0.1.1 pypi_0 pypi
[conda] pytorch-ignite 0.4.7 pypi_0 pypi
[conda] pytorch2keras 0.2.4 pypi_0 pypi
[conda] pytorch3d 0.5.0 pypi_0 pypi
[conda] segmentation-models-pytorch 0.1.3 pypi_0 pypi
[conda] torch 1.8.0+cu111 pypi_0 pypi
[conda] torch-geometric 1.7.2 pypi_0 pypi
[conda] torch-model-archiver 0.4.0 pypi_0 pypi
[conda] torch-scatter 2.0.7 pypi_0 pypi
[conda] torch-sparse 0.6.10 pypi_0 pypi
[conda] torch-workflow-archiver 0.1.0 pypi_0 pypi
[conda] torchaudio 0.8.0 pypi_0 pypi
[conda] torchserve 0.4.0 pypi_0 pypi
[conda] torchsul 0.1.26 pypi_0 pypi
[conda] torchvision 0.9.0+cu111 pypi_0 pypi
Could you please guide me on this?
@LeftAttention https://github.com/Zheng222/DMFN/blob/b6f2258ef0ac57a3cca16ee0aad7ea990834af03/train.py#L17
You can try to change to torch.backends.cudnn.benchmark = False
I tried that same issue. For the first batch it is running fine but for the second batch it is throwing this error during back propagation. Initially I thought this may be due to different input format but that is not the cause of this issue. In CPU it works. I am not able to figure out the cause of this issue.
@LeftAttention You can refer to my environment.
PyTorch version: 1.9.0+cu102
Clang version: Could not collect
CMake version: version 3.16.6
Libc version: glibc-2.17
Python version: 3.6 (64-bit runtime)
Python platform: Linux-4.15.0-142-generic-x86_64-with-Ubuntu-16.04-xenial
Is CUDA available: True
CUDA runtime version: Could not collect
GPU models and configuration:
GPU 0: GeForce RTX 2080 Ti
GPU 1: GeForce RTX 2080 Ti
Nvidia driver version: 440.33.01
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.7.6.5
/usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn.so.8.2.1
/usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8.2.1
/usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_adv_train.so.8.2.1
/usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8.2.1
/usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8.2.1
/usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8.2.1
/usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_ops_train.so.8.2.1
HIP runtime version: N/A
MIOpen runtime version: N/A
I tried that same issue. For the first batch it is running fine but for the second batch it is throwing this error during back propagation. Initially I thought this may be due to different input format but that is not the cause of this issue. In CPU it works. I am not able to figure out the cause of this issue.
Thanks. I will check.