Repository
synchopate/llama-modchip
Run 284B DeepSeek-V4-Flash (<59GiB RAM: 41.1 cold pp, 9.1 N=1 tg, up to 25.3 spec tok/s) & 30B MoEs (22.4 N=1 / 111.6 spec tok/s) on 64GB CPU (0 GPUs).
Stars
1.4k
Today
+16
last 24 hours
This week
+193
Forks
134
Clone
git clone https://github.com/synchopate/llama-modchip.gitStar growth
#avx512#cpu-inference#hpc#llama-cpp#mamba2#mixture-of-experts#nemotron#speculative-decoding#zen5#apple-silicon
- Created
- 2026-10-29
- Last commit
- 2026-11-13
- Gained since tracking
- +1,303
Updated Fri, 20 Nov 2026 07:21:13 GMT · description from the project itself