64 4 months ago

tools
ollama run uaysk0327/nemotron-3-nano:30b-q4_k_xl

Models

View all →

Readme

Able to use 92160 Context if you enable q8_0 kv cache quantize in 24GB VRAM, ollama’s official model is using q4_k_m and this model is using q4_k_xl so it can be fit in 24GB VRAM