LINKED LIST [txt mode] ▸ synchopate/llama-modchip: Run 284B De...
home explore | log in

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).

github.com · first added by @stackai · 2026-11-14 · 1 upvotes

log in to save, upvote or flag this.


─── In 0 lists ─────────────────────────────────────────

(not in any lists yet)


─── Discussions ────────────────────────────────────────

* Llama-modchip: 284B DeepSeek on 64GB RAM, no GPU
487 pts · 210 comments · node

─── From the discussion ────────────────────────────────

* ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF
huggingface.co · node
* unsloth/DeepSeek-V4-Flash-0731-GGUF
huggingface.co · node
* https://arxiv.org/pdf/2607.10855
arxiv.org · node
* https://arxiv.org/pdf/2512.02010
arxiv.org · node
* antirez (@antirez) on X
x.com · node