InhwanBae
commented
1 year ago Hi @Jasmine302, Your error seems to be the same as the one I noticed when I recently reran the source code. As mentioned in issue#1, it might help to use PyTorch version 11.1 or below, or to try another seed.
The method of the torch.distributions.log_prob() operation for probability distribution in the latest version of PyTorch has changed, which seems to cause problems with loss backpropagation. If any other issues arise, let me know!
Jasmine302
3123155028
I'm getting this error on the hotel dataset when reproducing your code Traceback (most recent call last): File "train.py", line 201, in
main()
File "train.py", line 181, in main
train(epoch)
File "train.py", line 106, in train
V_init, V_pred, V_refi, valid_mask = model(S_obs, S_trgt)
File "/home/jnu/xxq/Social-STGCNN-master/venv/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl
return forward_call(*input, **kwargs)
File "/home/jnu/xxq/GraphTERN-main/graphtern/model.py", line 129, in forward
mix = Categorical(torch.nn.functional.softmax(temp[:, :, :, 4], dim=-1))
File "/home/jnu/xxq/Social-STGCNN-master/venv/lib/python3.8/site-packages/torch/distributions/categorical.py", line 66, in init
super(Categorical, self).init(batch_shape, validate_args=validate_args)
File "/home/jnu/xxq/Social-STGCNN-master/venv/lib/python3.8/site-packages/torch/distributions/distribution.py", line 56, in init
raise ValueError(
ValueError: Expected parameter probs (Tensor of shape (1, 25, 8)) of distribution Categorical(probs: torch.Size([1, 25, 8])) to satisfy the constraint Simplex(), but found invalid values:
tensor([[[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan],
[nan, nan, nan, nan, nan, nan, nan, nan]]], device='cuda:0',
grad_fn=)
Have you encountered?