- Updated server classes (OpenAIServer, SGLangServer, TrlVllmServer, VLLMServer) to accept a ReasoningConfig parameter during initialization.
- Enhanced ReasoningConfig to allow flexible max_tokens without strict validation, accommodating varying provider limits.
- Implemented reasoning configuration injection in APIServer methods for chat and completion handling.
- Updated tests to reflect changes in max_tokens validation logic.
This commit integrates reasoning capabilities into the server handling architecture, improving compatibility with diverse reasoning models.
- Add min_batch_allocation parameter to ensure environments contribute minimum proportion to each batch
- Implement grab_batch_with_minimum_allocations function with proper scaling when allocations exceed 100%
- Add mixed-size group buffering to handle variable-sized data submissions
- Update server to use minimum allocation logic when any env has min_batch_allocation set
- Add comprehensive tests for minimum allocation scenarios
- Update documentation in API README and CONFIG.md
- Update example environments to demonstrate the feature
This feature allows critical environments to guarantee they contribute at least a specified proportion (0.0-1.0) to each training batch, ensuring important data sources are always represented during training.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Made reward field truly optional in messages (no auto-addition)
- Accept custom roles (dog, cat, etc.) beyond standard ones
- Added 24 new tests for edge cases (tuples, unicode, large content)
- Reorganized test structure: moved from testing/ to atroposlib/tests/
- Fixed legacy API tests and removed tests requiring missing data files
All 43 tests pass\! Fixes message handling for SFT use cases.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
