Closed AlexanderTrg closed 1 year ago
all labels in train ds from IMBD are 1
from torchtext.datasets import IMDB train_iter = IMDB(split='train') labels = [label for label, text in train_iter] print(sum(labels)/len(labels)) 1.0
pytorch 2.0.0+cu118 torchtext 0.15.1+cpu
Closing because this is a duplicate of https://github.com/pytorch/text/issues/2041. Let's track the issue there
🐛 Bug
all labels in train ds from IMBD are 1
from torchtext.datasets import IMDB train_iter = IMDB(split='train') labels = [label for label, text in train_iter] print(sum(labels)/len(labels)) 1.0
pytorch 2.0.0+cu118 torchtext 0.15.1+cpu