OptimizeLLM/Qwen3.8-27B-heretic-MTP-FP8

Quantization vs Qwen/Qwen3.8-27B. OptimizeLLM/Qwen3.8-27B-heretic-MTP-FP8 is a quantized version of the other (F8_E4M3/BF16).

On the Hub · Downloads on Model Pulse