Skip to content

Harden FastWordpieceTokenizer, FastBertNormalizer, and PhraseTokenizer with#1521

Open
copybara-service[bot] wants to merge 1 commit into
masterfrom
test_925446199
Open

Harden FastWordpieceTokenizer, FastBertNormalizer, and PhraseTokenizer with#1521
copybara-service[bot] wants to merge 1 commit into
masterfrom
test_925446199

Conversation

@copybara-service
Copy link
Copy Markdown
Contributor

Harden FastWordpieceTokenizer, FastBertNormalizer, and PhraseTokenizer with
bounds verification and initialization checks:

  • Prevent null pointer dereferences by validating required FlatBuffer fields.
  • Prevent OOB reads during detokenization by checking token IDs against vocab size.
  • Secure FastBertNormalizer against adversarial heap slicing by tracking string pool limits.

@copybara-service copybara-service Bot force-pushed the test_925446199 branch 5 times, most recently from b9a0d74 to 6df8131 Compare June 3, 2026 18:07
…er with

  bounds verification and initialization checks:
  * Prevent null pointer dereferences by validating required FlatBuffer fields.
  * Prevent OOB reads during detokenization by checking token IDs against vocab size.
  * Secure FastBertNormalizer against adversarial heap slicing by tracking string pool limits.

PiperOrigin-RevId: 925446199
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant