python/sglang/srt/models/deepseek.py · 2695ab05375456b4a2ac625cca12b95c3b90e5ae · change / sglang · GitLab

Find file Blame History Permalink

Fix loading KV quantization scale; Enable modelopt kv cache (#4686) · 2695ab05
Yun Dai authored Apr 08, 2025
```
Co-authored-by: qingquansong <ustcsqq@gmail.com>
```
2695ab05

deepseek.py 16.5 KB