Run 35B MoE models in ~2.5 GB of RAM. Weights stream from SSD; a trained router predicts loads a step ahead. Mac, iPhone, Android, Windows supported
Python
Commits on main
Merge pull request #13 from agourakis82/fix/install-layer-discovery
2f917d0 Jiahao committed 1 month ago
Merge pull request #118 from agourakis82/fix/hot-layer-anonymous-memory-doc
9e3aa60 Jiahao committed 1 month ago
Merge pull request #120 from agourakis82/docs/warm-willneed-inert-under-prod-k8
2dc8316 Jiahao committed 1 month ago
Merge pull request #132 from kevin9327/fix/readme-mlx-allocator-peak
85b2a03 Jiahao committed 1 month ago
android: refresh launcher icons from the wider-padding master
fc743a4 Alex Zhou committed 1 month ago
android: refresh launcher icons from the revised full-bleed master
22aa31c Alex Zhou committed 1 month ago
chore: ignore local-only alignment reference exporter script
497af1a Alex Zhou committed 1 month ago
windows: fix issue #128 — converter path robustness, ASCII-safe output, engine version probe, dependency preflight
8b0ea9c Alex Zhou committed 1 month ago
fix(android): fall back to pwd -P when the depot probe runs off MSYS
0bea963 Alex Zhou committed 1 month ago
docs(streaming): qualify prefill readahead prerequisites
2c609e3 Demetrios Agourakis committed 1 month ago
Merge remote-tracking branch 'origin/main' into upstream-lab/edge0-120-fix-20261004
a7d2084 Demetrios Agourakis committed 1 month ago
Merge pull request #119 from agourakis82/feat/openai-tool-calling
9a56e4d linyubupa committed 1 month ago