Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

How to generate hypothese with fconv for DrNMT? (ende) #5563

Open
123THU opened this issue Nov 8, 2024 · 0 comments
Open

How to generate hypothese with fconv for DrNMT? (ende) #5563

123THU opened this issue Nov 8, 2024 · 0 comments

Comments

@123THU
Copy link

123THU commented Nov 8, 2024

❓ Questions and Help

I followed the instructions from DrNMT to train a base MT model(fconv) first, and then hoped to use the MT model to prepare three files: source, target and hypothese. However, when I started to use the generated hypotheses file, it showed:

AssertionError: /mnt/workspace/lastversion/venv/fairseq/examples/discriminative_reranking_nmt/data/train_hypothese.txt should contain 11883537 hypotheses but only has 3961176. (--beam=3)

So I want to ask what are the correct files(source and target, I choose train.en, train.de)), and how should I generate the correct hypothese file.

Thank you so much for relpying!!!!

This is the code I use to generate hypotheses with fconv

fairseq-generate data-bin/wmt17_en_de \ --path checkpoints/fconv_wmt_en_de/checkpoint_best.pt \ --beam 3 --remove-bpe --gen-subset train > output_train.txt

environment

  • Ubuntu
  • pip fiarseq
@123THU 123THU changed the title How to generate hypothese with fcov for DrNMT? How to generate hypothese with fconv for DrNMT? (ende) Nov 8, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant