Text Generation
Transformers
PyTorch
Chinese
English
llama
text-generation-inference

The problem of pad_token

#10
by kang1 - opened

I noticed that in added_token, you set pad_token_id as 64000, but in config.json, you set pad_token_id as 0, is there exist some mistakes?

Thanks for pointing out this problem. The real pad token id is 0. I have deleted the unused add_token file.

fireballoon changed discussion status to closed

Sign up or log in to comment