You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Traceback (most recent call last):
File "/home/shared/tabzilla/TabSurvey/tabzilla_experiment.py", line 137, in __call__
result = cross_validation(model, self.dataset, self.time_limit)
File "/home/shared/tabzilla/TabSurvey/tabzilla_utils.py", line 247, in cross_validation
train_predictions, train_probs = curr_model.predict_wrapper(X_train)
File "/home/shared/tabzilla/TabSurvey/models/basemodel.py", line 108, in predict_wrapper
self.predictions, self.prediction_probabilities = self.predict(X)
File "/home/shared/tabzilla/TabSurvey/models/basemodel_torch.py", line 163, in predict
self.predict_proba(X)
File "/home/shared/tabzilla/TabSurvey/models/basemodel_torch.py", line 170, in predict_proba
probas = self.predict_helper(X)
File "/home/shared/tabzilla/TabSurvey/models/hopular_model.py", line 141, in predict_helper
preds = self.model(batch_X[0])
File "/opt/conda/envs/torch/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(*input, **kwargs)
File "/home/shared/tabzilla/TabSurvey/models/hopular/blocks.py", line 791, in forward
embeddings = self.embeddings(input).unsqueeze(dim=0)
File "/opt/conda/envs/torch/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(*input, **kwargs)
File "/home/shared/tabzilla/TabSurvey/models/hopular/blocks.py", line 119, in forward
input_embedded = torch.cat(tuple(feature_embedding(
File "/home/shared/tabzilla/TabSurvey/models/hopular/blocks.py", line 119, in <genexpr>
input_embedded = torch.cat(tuple(feature_embedding(
File "/opt/conda/envs/torch/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(*input, **kwargs)
File "/opt/conda/envs/torch/lib/python3.10/site-packages/torch/nn/modules/container.py", line 141, in forward
input = module(input)
File "/opt/conda/envs/torch/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(*input, **kwargs)
File "/opt/conda/envs/torch/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 103, in forward
return F.linear(input, self.weight, self.bias)
RuntimeError: mat1 and mat2 shapes cannot be multiplied (256x1 and 2x32)
The text was updated successfully, but these errors were encountered:
encountered with dataset
openml__collins__3567
:The text was updated successfully, but these errors were encountered: