1
Max-Window Scale Estimation for Near-Lossless HiF8 W8A8 Quantization-Aware Training
提出Max-Window尺度估计方法，实现HiF8 W8A8近无损量化感知训练的新突破。
arXiv:2605.26189v1 Announce Type: cross Abstract: Quantization-aware training (QAT) with low-bit floating-point formats enables efficient LLM deployme…