RedHatAI/Meta-Llama-3.1-70B-Instruct-FP8-dynamic
Quantization vs meta-llama/Llama-3.1-70B-Instruct. RedHatAI/Meta-Llama-3.1-70B-Instruct-FP8-dynamic is a quantized version of the other (F8_E4M3/BF16).
3 other repos hold byte-for-byte identical weight files.