Skip to content
Hot eventLive

Format-Aware Fusion for Fast FP4 Pretraining

1 reports1 sources5 hr ago updated

Get the story

AI overview

2026-10-02 arXiv 发表 Machine Learning Theory 论文,提出 format-aware fusion 方法,通过协同设计量化生产者与规模域、消费者布局,在 native \mxfp{}、global \nvfp{} 和 cooperative-thread-array-local \nvfp{} 上加速 FP4 预训练。

Generated from reports · updated 4 hr ago

Timeline

Follow the coverage from different angles.

Oct 2, 2026
  1. arXiv · Machine Learning Theory
    Format-Aware Fusion for Fast FP4 Pretraining

    本文提出 format-aware fusion 方法,通过协同设计量化生产者与规模域、消费者布局,在 native \mxfp{}、global \nvfp{} 和 cooperative-thread-array-local \nvfp{} 上加速 FP4 预训练。

Heat trend

Current heat 9·Comparable peak 10(Oct 2)·Comparable change over 24 hours –

02.557.510Oct2Oct2Oct2Oct2

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.