vimalnakrani08

vimalnakrani08/stillwarm

Session persistence for llama.cpp — save and restore KV caches to disk. Measured up to 201× faster resume.

0 stars 0 forks 0 watchers Python MIT License

benchmark kv-cache llama-cpp llm-inference python

1 Open Issue Need Help Last updated: Jul 6, 2026

Open Issues Need Help

View All on GitHub

Windows: verify slot save/restore + CI runner about 3 hours ago

help wanted

vimalnakrani08/stillwarm

0

Session persistence for llama.cpp — save and restore KV caches to disk. Measured up to 201× faster resume.

Python

#benchmark#kv-cache#llama-cpp#llm-inference#python