{}
AICAS 2024 - Final round
Qwen-1.8B full-layer QAT model: Training set-PIQA, groupsize=128
PIQA acc =
