NVFP4-first inference engine for consumer Blackwell. OMMA.SF.16864, KV64 cache (128K context), multi-kernel dispatch. DenForge engine backend. RTX 5070 Ti (GB203-300-A1, SM120).
-
Updated
May 20, 2026 - C++
#
omma
Here is 1 public repository matching this topic...
Improve this page
Add a description, image, and links to the omma topic page so that developers can more easily learn about it.
Add this topic to your repo
To associate your repository with the omma topic, visit your repo's landing page and select "manage topics."