isogen's picture
Upload folder using huggingface_hub
9aea40e verified
metadata
base_model: Qwen/Qwen3-0.6B-Base

EXL3 quantization of Qwen3-0.6B-Base, 8 bits per weight, including output layers.

HumanEval (argmax)

Model Q4 Q6 Q8 FP16
Qwen3-0.6B-exl3-8bpw-h8 0.0% 38.4% 40.9% 40.2%
Qwen3-0.6B-Base-exl3-8bpw-h8 0.0% 36.0% 37.2% 36.6%