Model Recipes
Meta

meta-llama/Llama-3.3-70B-Instruct

Llama 3.3 70B dense model with NVIDIA FP8/FP4 quantized variants for Hopper and Blackwell GPUs

View on HuggingFaceView on ModelScope
dense70B131,072 ctxvLLM 0.12.0+text
Guide