CoolFace
Modelpublic

spectator2026/MiniMax-M3-AWQ-int4

sourceHugging Faceotherupdated 3mo agoView on Hugging Face
5likes1.2kdownloads
9 commits on main
154bf4a3mo ago

card: link the fp8-KV serving note to patches/

spectator2026
52473713mo ago

patches: add e5m2 fp8 KV-cache serving patches (A100, ~1.5x KV)

spectator2026
8cdbc0c3mo ago

serving notes: e5m2 fp8 KV-cache works on A100 (~1.5x KV, modest KLD cost)

spectator2026
5b421883mo ago

Upload README.md with huggingface_hub

spectator2026
8d9429d3mo ago

Delete folder serving with huggingface_hub

spectator2026
b1ceb5f3mo ago

Upload folder using huggingface_hub

spectator2026
b0a7f853mo ago

Upload folder using huggingface_hub

spectator2026
7d1ba113mo ago

Add files using upload-large-folder tool

spectator2026
ef207ff3mo ago

initial commit

spectator2026