GitLab

Change install.md and Update reference to optimize rules directory

Replaced Chinese comments in iqk_mul_mat.inc with English to avoid breaking MSVC compiling

Signed-off-by: Emmanuel Ferdman <emmanuelferdman@gmail.com>

fix: make cpufeature a local import

fix load default max_new_tokens

fix chat template encoding

add check-para

change test

kill serve lead to kill sched and engine

update speed test

Update param

Update Function call

优化提示词，解决部分Deepseek r1的兼容性

fix non stream

Fix cmake config error

Move KV cache creation to balance_serve

Signed-off-by: onepick <jiajuku12@163.com>

Enh: Make Ollama perf data more accurate, consistent with OpenAI's implementation