Re-export int8 per-block-32: the per-channel bundle had dead LM-head rows from vocab id ~65024 up (never emitted <|im_end|>) b8a6ac3 verified mlboydaisuke commited on 24 days ago
Re-export with coreai-torch 0.4.1 (fixes OS 27 beta 2+ load failure) 5ad650f verified mlboydaisuke commited on Jul 20
MiniCPM5-1B int8 (Core AI) — iPhone 66.8 tok/s, lossless 6ec16ba verified mlboydaisuke commited on Jun 26