Error in seq2seq.py of Method collate

[lypenghao]

/home/wangph/code/pytorch_egs/speech/speech/models/seq2seq.py:235: UserWarning: volatile was removed and now has no effect. Use with torch.no_grad(): instead.
inputs.volatile = True
/home/wangph/code/pytorch_egs/speech/speech/models/seq2seq.py:236: UserWarning: volatile was removed and now has no effect. Use with torch.no_grad(): instead.
labels.volatile = True

Traceback (most recent call last):
File "train.py", line 146, in
run(config)
File "train.py", line 109, in run
dev_loss, dev_cer = eval_dev(model, dev_ldr, preproc)
File "train.py", line 58, in eval_dev
loss = model.loss(batch)
File "/home/wangph/code/pytorch_egs/speech/speech/models/seq2seq.py", line 53, in loss
x, y = self.collate(*batch)
TypeError: collate() missing 2 required positional arguments: 'inputs' and 'labels'

[dmitrii-obukhov]

Probably, this is because batch is zip object, and you can iterate over batch only once. Adding 'batch = list(batch)' at eval_dev() in the for loop helps me.

I'm currently facing the same issue, I added batch = list(batch) and manually replaced collate with collate(batch[0], batch[1]) and the issue becomes index 1 out of range.

[iamxiaoyubei]

+1