Fix Mistral tokenizer missing spaces in decode_stream (Issue #1822)#2211
Open
mrshibly wants to merge 3 commits into
Open
Fix Mistral tokenizer missing spaces in decode_stream (Issue #1822)#2211mrshibly wants to merge 3 commits into
mrshibly wants to merge 3 commits into
Conversation
mrshibly
requested review from
KaelanDt,
andyland,
k223kim,
lantiga,
lianakoleva and
t-vi
as code owners
for more information, see https://pre-commit.ci
mrshibly
commented
Mar 5, 2026
mrshibly
left a comment
mrshibly
left a comment
Author
There was a problem hiding this comment.
Summary Fixes #1822 where litgpt chat generated text without spaces for Mistral and other HuggingFace tokenizers.
Changes Unified Tokenizer.decode_stream to use a running buffer approach for all backends. This ensures that BPE-based tokenizers correctly preserve leading spaces and context-dependent characters during streaming, which were previously stripped when decoded individually.
Testing Verified manually. The fix aligns the HuggingFace backend logic with the robust implementation already used for SentencePiece.
lianakoleva
approved these changes
Mar 21, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary Fixes #1822 where litgpt chat generated text without spaces for Mistral and other HuggingFace tokenizers.
Changes Unified Tokenizer.decode_stream to use a running buffer approach for all backends. This ensures that BPE-based tokenizers correctly preserve leading spaces and context-dependent characters during streaming, which were previously stripped when decoded individually.
Testing Verified manually. The fix aligns the HuggingFace backend logic with the robust implementation already used for SentencePiece.