Open madhawav opened 3 years ago
Had the same issue, tweaked the code a bit to:
if len(outputs)>0:
for key in outputs[0].keys():
logs[key] = torch.stack([x[key] for x in outputs]).mean()
else:
logs['val_loss']=torch.tensor(0.)
This is very ad hoc, i think the code needs a 'val' folder as well as a train and test
Hi, When I try to train on a new dataset, it fails with the following error.
Additional Information
My "config_default.yml" Is shown below:
Your help is much appreciated.