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
adjust 3 bits hyperparameters at auto-round-best (#1081)
ef78c68 Wenhua Cheng committed 11 months ago
Support transformers loading quantized moe model (#1067)
b1b60e4 Wang, Mengni committed 11 months ago
[High Risk]reduce vram usage for optimized RTN mode (#1043)
e1b89d2 Wenhua Cheng committed 12 months ago