synchopate/llama-modchip Python
🐋 Run 284B DeepSeek-V4-Flash (<59GiB RAM, 44.4 tok/s pp, 9.1-14.7 tok/s) & 30B MoEs (25-111 tok/s) on a pure 64GB CPU at 256K ctx (0 GPUs). SwiGLU-H2, Stockfish-PVS L3 & 13ms Claude Code restore.
Why It Is Worth Watching
synchopate/llama-modchip is worth watching as a Python project with 878 stars and 62 forks. Its main tags include ai-agents, apple-silicon, avx512, claude-code. The project description says: "🐋 Run 284B DeepSeek-V4-Flash (<59GiB RAM, 44.4 tok/s pp, 9.1-14.7 tok/s) & 30B MoEs (25-111 tok/s) on a pure 64GB CPU at 256K ctx (0 GPUs)…". It was last updated on 2026-11-10. The current dataset does not yet include enough history to judge whether the growth is stable. It is a useful candidate for developers who are comparing tools, tracking open-source trends, or looking for practical examples in this area. Before adopting it, review the README examples, integration cost, license, issue activity, and release cadence.
📈 Trend History (1 points)
|
WeChat updates Follow the Project Worth Watching WeChat account for daily open-source picksDaily project highlights, growth observations, and deeper reads are also published on WeChat for quick reading. |
|