2605.26628v1 May 26, 2026 cs.AI

Tail-Aware HiFloat4: W4A4 Post-Training Quantization for Wan2.2

Xin Di
Xin Di
Citations: 374
h-index: 11
Long Peng
Long Peng
Citations: 823
h-index: 16
Zhengjun Zha
Zhengjun Zha
Citations: 8
h-index: 1
Zhanfeng Feng
Zhanfeng Feng
Citations: 99
h-index: 4
Shuai Guo
Shuai Guo
Citations: 115
h-index: 6
Yang Cao
Yang Cao
Citations: 24
h-index: 1

This report describes Tail-Aware HiFloat4, our submission to the low-bit text-to-video generation quantization challenge. Our method adapts the public ViDiT-Q post-training quantization pipeline to Wan2.2 under the HiFloat4 numerical format. We quantize the main linear layers in both Wan2.2 transformer modules with W4A4 HiFloat4 fake quantization, keep numerically sensitive boundary modules in high precision, and introduce an activation-tail-aware percentile calibration module for channel-mask construction. Together with compact PTQ-state restoration, this design reduces the influence of rare calibration outliers while keeping the runtime HiFloat4 arithmetic and sampling pipeline unchanged.

0 Citations
0 Influential
8 Altmetric
40.0 Score
Original PDF

No Analysis Report Yet

This paper hasn't been analyzed by Gemini yet.

Log in to request an AI analysis.

댓글

댓글을 작성하려면 로그인하세요.

아직 댓글이 없습니다. 첫 번째 댓글을 남겨보세요!