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.

Sign up for GitHub

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

Jump to bottom

mask token id == 3 during pretraining? #231

Open

whaleloops opened this issue Nov 8, 2022 · 0 comments

whaleloops commented Nov 8, 2022

I noticed that kMaskWordTokenId (mask2 as defined in the paper) is 3 as defined below.
https://github.com/google-research/pegasus/blob/main/pegasus/ops/pretrain_parsing_ops.cc#L69

However, the id of token 'a' is also 3 in sentencepiece vocab from "gs://t5-data/vocabs/cc_all.32000/sentencepiece.model"

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment