[ML] Fix Array out of bounds exception in the XLM Roberta tokenizer #106655
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Outdated suggestions cannot be applied.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.
The following
ArrayIndexOutOfBoundsException
has been observed tokenising certain Unicode characters in the XLM Roberta tokenizer. The overflow comes from a fixed sized reusable buffer that holds the normalised form of a single unicode character encoded as UTF-8. The normalised form can be much longer than the allocated buffer size (8 chars), for example 'ﷺ' normalises to 33 characters. That is surprising so I compared the output to other SentencePiece implementations and the results are in agreement.The fix here is to spend a few bytes increasing the buffer size.
This PR also removes some unused fields.