error when running code about machine-learning-collection HOT 3 CLOSED

Do you have CUDA enabled GPU? The default argument for running the Transformer from scratch code is cuda. If you want to run it on CPU you need to change this call. Specifically change the call of

model = Transformer(src_vocab_size, trg_vocab_size, src_pad_idx, trg_pad_idx).to(
        device
)

to

model = Transformer(src_vocab_size, trg_vocab_size, src_pad_idx, trg_pad_idx, device='cpu').to(
        device
)

Edit:
I think since we're setting the device a few lines below we should actually do device=device, rather than device='cpu' and this is definitely an improvement to the code. Will fix this

from machine-learning-collection.

I'm closing this, let me know if you still wonder about this!

from machine-learning-collection.

thank you very much, I couldn´t test it yet but thank you for the suggestion to fix the issue, will try it as soon as I can, best wishes

from machine-learning-collection.