GitHub - gavamedia/deltafin: Run Kimi K3, a 2.8T-parameter Mixture-of-Experts LLM, on a single Apple Silicon Mac. Streams MXFP4 experts on demand over HTTP into a local disk cache — fused NEON kernels, Metal/MPS compute, exact reproducible decoding, and an OpenAI-compatible API server for local chat and coding agents.

Running Kimi K3 on a M1 Mac

Run Kimi K3, a 2.8T-parameter Mixture-of-Experts LLM, on a single Apple Silicon Mac. Streams MXFP4 experts on demand over HTTP into a local disk cache — fused NEON kernels, Metal/MPS compute, exact...

100 points by tito 85 days ago github.com 69 comments

Summary

Learn how to run the Kimi K3 language model on an M1 Mac, utilizing its powerful hardware to efficiently manage a large number of parameters and stream experts on demand. This setup leverages advanced computing techniques like fused NEON kernels and Metal/MPS for optimized performance.

Comments (69 of ) ⌛

Loading more comments... ⌛

Scroll down to load more comments

No comments available.