[Combined PR] Security Hardening, Standardization, and Robustness Improvements #507
Conversation
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
This commit addresses: 1. Security Fix (gemma google-deepmind#469): Replaces unsafe eval() with AST-based _SafeEvaluator in Calculator tool. 2. Architecture: Renames num_embed to vocab_size across the codebase for consistency. 3. Compatibility: Fixes legacy math statement SyntaxErrors for Python 3.8/3.9. 4. Cleanup: Removes stale nucleus sampling TODO.
… tests - Uses errors='replace' in _decode_bytes to prevent UnicodeDecodeError. - Adds gemma/gm/data/_tasks_test.py for permanent verification.
…nology - Fixes multiple typos in classification example and transformer comments (google-deepmind#423). - Refactors last token slicing to use jnp.take_along_axis in core Transformer. - Renames attention_types to layers_types repository-wide for architectural consistency.
Ashutosh0x
changed the title
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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 of Improvements
This PR consolidates several critical engineering improvements for the Gemma repository:
1. Security Hardening for Calculator (#469)
eval()with a strict AST-based evaluator (_SafeEvaluator).pi,e), and 10+ core functions.2. Architectural Standardization
num_embed->vocab_size(25+ files) for consistency with industry standards.attention_types->layers_typesto support non-attention layer types (e.g., identity).matchstatements toif/elifblocks for Python 3.8/3.9 compatibility.3. JAX Performance & Readability
jnp.take_along_axisin_transformer.pyandgemma3n/_transformer.py, following maintainerTODOrecommendations.4. Data Pipeline Robustness (#504)
_decode_bytesin_tasks.pywitherrors='replace'to prevent crashes on invalid UTF-8 sequences.gemma/gm/data/_tasks_test.py.5. Quality Assurance
TODOcomments after verifying feature completion.Verified through exhaustive unit tests, architectural audits, and compilation checks.