Skip to content

Bugfix: Correct context length for text-gen dataset #581

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

Merged
jambayk merged 1 commit into from
Sep 19, 2023
Merged

Bugfix: Correct context length for text-gen dataset #581

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
2 changes: 1 addition & 1 deletion olive/data/component/text_generation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -168,7 +168,7 @@ def text_gen_corpus_pre_process(_dataset, tokenizer, all_kwargs):
# no randomization, just use contiguous blocks of tokens
if args.corpus_strategy == TextGenCorpusStrategy.JOIN_SLIDING_WINDOW:
# we use the stride as both the step between sequences and the context size
step, context = args.stride, args.stride
step, context = args.stride, seqlen - args.stride
else:
# JOIN corpus_strategy
# text is split into non-overlapping sequences and there is no context
Expand Down