v0.13.1
新功能
New Features
- The model inference service stress testing now supports generating prompts of specified length using random values. Refer to the user guide for more details.
- Compatible with evaluation during training using the ms-swift training framework. Reference
- Fixed bugs related to framework stability.
What's Changed
- compat swift by @Yunnglin in #397
- Update perf random dataset by @Yunnglin in #399
- fix dump config by @Yunnglin in #401
- Filter illegal characters & Use asyncio.new_event_loop() instead of deprecated get_event_loop() by @Mengxi12345 in #411
- reduce save items and update doc by @Yunnglin in #406
- Fix loop eval by @Yunnglin in #415
- add perf no test connection by @Yunnglin in #417
- update doc by @Yunnglin in #422
New Contributors
- @Mengxi12345 made their first contribution in #411
Full Changelog: v0.13.0...v0.13.1
Contributors
Yunnglin and Mengxi12345