kodelow/Hy3-NVFP4-W4A16
Text Generation • 172B • Updated • 20 • 8
Quantization · vLLM · NVFP4 · On-prem inference
· Kodelow ·
Independent AI lab
Deployment artifacts for those who run open models on their own hardware: quantized checkpoints, harness-specific post-training.
Everything here runs air-gapped on stock vLLM.
· Espoo, Finland ·