The Model is based on Alexnet. I was training it with python 3.7 pytorch 0.4.1.
But in the forward process it showed the following errors:
Traceback (most recent call last):
File “interpretable_CNN-Zhang-learned_template.py”, line 683, in
model=train_model(model,dataloaders,loss_func,optimizer,scheduler,device,100)
File “interpretable_CNN-Zhang-learned_template.py”, line 494, in train_model
outputs=model(inputs)
File “/home/TUE/20183494/miniconda3/envs/env_new/lib/python3.7/site-packages/torch/nn/modules/module.py”, line 477, in call
result = self.forward(*input, **kwargs)
File “interpretable_CNN-Zhang-learned_template.py”, line 402, in forward
x=self.dropout(x)
File “/home/TUE/20183494/miniconda3/envs/env_new/lib/python3.7/site-packages/torch/nn/modules/module.py”, line 477, in call
result = self.forward(*input, **kwargs)
File “/home/TUE/20183494/miniconda3/envs/env_new/lib/python3.7/site-packages/torch/nn/modules/dropout.py”, line 53, in forward
return F.dropout(input, self.p, self.training, self.inplace)
File “/home/TUE/20183494/miniconda3/envs/env_new/lib/python3.7/site-packages/torch/nn/functional.py”, line 595, in dropout
return functions.dropout.Dropout.apply(input, p, training, inplace)
File "/home/TUE/20183494/miniconda3/envs/env_new/lib/python3.7/site-packages/torch/nn/functions/dropout.py", line 40, in forward
ctx.noise.bernoulli(1 - ctx.p).div(1 - ctx.p)
RuntimeError: Creating MTGP constants failed. at /opt/conda/conda-bld/pytorch_1535493744281/work/aten/src/THC/THCTensorRandom.cu:34
Is anybody encounter this problem? I don’t understand what caused this problem, is there anybody can help me ?
Many thanks