qwen2.5 long context

Open ysjprojects opened this issue 10 months ago • 1 comments

Qwen2.5 7B Instruct and Qwen2.5 14B Instruct extended to 1 million context length

https://qwenlm.github.io/blog/qwen2.5-1m/

Feb 04 '25 05:02 ysjprojects

Looks great, thank you @ysjprojects . we limit the default kv-cache size, though?

Feb 04 '25 10:02 t-vi

PR should be ready to merge, the failing test case is unrelated to the model.

May 16 '25 07:05 ysjprojects

Thank you @ysjprojects

Jun 15 '25 15:06 KaelanDt