trunk/f6585f7f189d6d457fcec520b399a39140a8ba8a: [MPS] Migrate attention from mps graph to metal fully (#201745)
- PyTorch 近 90 天出现 2750 次
- PyTorch 近 90 天第 2707 次发版
- 上一次：同一天稍早 · viable/strict/1792837579: Use vectorized kernels for index_copy (#201814)
发生了什么
Migrates attention from MPS Graph to Metal. Look at the table below for PR2 to look at the speedups for various shapes. Most of the shapes with this PR is faster than previous MPS Graph, though some of the PRs in the stack might be worth merging to squeeze the most performance. But discarding those PRs, this PR already achieves a nice geomean on M5 Pro and M3 ultra machines so I think it's worth to merge this initia…
摘要按规则整理自下方来源原文