Repository navigation
Conversation
- types: add FinishReason.MALFORMED_RESPONSE, which the API returns (googleapis#2937). - _api_client: use `await asyncio.sleep` instead of `time.sleep` in the httpx branch of the async upload retry, so the backoff no longer blocks the event loop (googleapis#2938). - _transformers: keep annotations next to a `$ref` (e.g. a field-level `description`) when inlining the referenced definition in response_schema, instead of overwriting them with the shared model's (googleapis#3007). - _common: serialize Pydantic's deferred model build through a lock, since `defer_build=True` is not thread safe and concurrent first use could see a half-built model (googleapis#2991). - _api_client: in `__del__`, close each aiohttp session on the event loop that created it and keep strong references to the scheduled close tasks so they are not destroyed while pending (googleapis#2995). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01R97bRW4GzCnhFEm2VhymSq
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01R97bRW4GzCnhFEm2VhymSq
fix: address several open SDK bugs (googleapis#2937, googleapis#2938, googleapis#3007, googleapis#2991, googleapis#2995)
chore(main): release 2.11.0
Author
|
Cerrado para limpieza de ramas. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
This pull request ensures an explicit default
top_p=0.95is maintained across generation configuration classes ingoogle/genai/types.py(GenerateContentConfig,GenerationConfig,Model, andLiveConnectConfig).Motivation & Context
Provides a reliable nucleus sampling anchor to eliminate reasoning drift, token repetition loops, and hallucinations during thinking modes under default
temperature: 1.0.This directly addresses the SDK control surface requirements and reasoning calibration tracked in Issue #2954.
Key Changes
default=0.95insideField(...)fortop_pingoogle/genai/types.py.Verification
mypytype checking tests with zero syntax or duplicate argument errors.@adri22235).