opencompass
opencompass copied to clipboard
[Update] Add configs for Gemma, OREAL, and Qwen2.5 models
Description:
- Add VLLM configs for Qwen2.5 models (7B/14B/32B/72B) supporting 128k long context.
- Add VLLM configs for Gemma models (3-4B/12B/27B) with 128k context support.
- Add TurboMind config for OREAL-32B.