PTQ-Level Efficiency, QAT-Level Accuracy | 训练后量化的效率,量化感知训练的精度
Commits on main
support two more calib datasets and fix embedding layer bug (#653)
7d72403 Wenhua Cheng committed 1 year ago
Integrate RTN quantization into GGUF packing to enhance robustness (#644)
06e0f52 Heng Guo committed 1 year ago
support packing immediately for gguf to reduce ram usage (#638)
3b282dc Wenhua Cheng committed 1 year ago
fix gguf bug of moe models and lmhead/embedding bits setting regression (#628)
20eed5b Heng Guo committed 1 year ago
fix lots of gguf critic bugs and support imatrix in rtn mode (#595)
f9921cd Wenhua Cheng committed 1 year ago