amd/Llama-3.1-70B-Instruct-FP8-KV

Quantization vs meta-llama/Llama-3.1-70B-Instruct. amd/Llama-3.1-70B-Instruct-FP8-KV is a quantized version of the other (F8_E4M3/BF16).

On the Hub · Downloads on Model Pulse