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.

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

set unique UNK token #435

Draft
wants to merge 1 commit into
base: main
Choose a base branch
from
Draft

set unique UNK token #435

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions tuning/sft_trainer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -292,6 +292,14 @@ def train(
tokenizer.pad_token = configs.DEFAULT_PAD_TOKEN
else:
tokenizer.eos_token = configs.DEFAULT_EOS_TOKEN
if tokenizer.unk_token == tokenizer.eos_token:
logger.warning(
"UNK token set to default, to make it different from eos token"
)
if tokenizer.eos_token != configs.DEFAULT_UNK_TOKEN:
tokenizer.unk_token = configs.DEFAULT_UNK_TOKEN
else:
tokenizer.eos_token = configs.DEFAULT_EOS_TOKEN

# TODO: lower priority but understand if resizing impacts inference quality and why its needed.
# It makes sense if we manipulate tokenizer that we also save it and provide it to inference.
Expand Down
Loading