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
docs(paper): add Edge0 technical report (The Other Half of the Memory Wall)
linyubupaca40e7d6f72e
diff --git a/paper/main.pdf b/paper/main.pdfnew file mode 100644index 0000000..205ed3bBinary files /dev/null and b/paper/main.pdf differ