Skip to content

[WS2][CUDA][MiniMax-H3] sp_norm_adaln: sequence-parallel norm/AdaLN/gated residual, byte-equal to WS1 (RFC #420 stack 9/10) - #494

Open
fusheng-ji wants to merge 384 commits into
RL-Align:test-h3from
fusheng-ji:feat/h3-sp-norm-adaln
Open

fusheng-ji wants to merge 384 commits into
RL-Align:test-h3from
fusheng-ji:feat/h3-sp-norm-adaln

Commits

This pull request is big! We're only showing the most recent 250 commits
  1. Commits on Oct 10, 2026

  2. Commits on Oct 13, 2026

  3. Commits on Oct 19, 2026

  4. Commits on Oct 21, 2026

  5. Commits on Oct 23, 2026

  6. Commits on Oct 25, 2026

  7. Commits on Oct 26, 2026

  8. Commits on Oct 27, 2026