Commits

Branch selector

 main

User selector

All users

All time

Commit History

Commits on Apr 21, 2026

• Merge pull request #386 from AI-Hypercomputer:ninatu/wan-vace-batch-size-fix

  Show description for ae22683
  Google-ML-Automation

  committed
  ae22683

  Copy full SHA for ae22683

• Merge pull request #385 from AI-Hypercomputer:prisha/ltx2_opt

  Show description for b224fd2
  Google-ML-Automation

  committed
  b224fd2

  Copy full SHA for b224fd2

• Merge pull request #382 from AI-Hypercomputer:ltx2_bugfix

  Show description for 0b6410b
  Google-ML-Automation

  committed
  0b6410b

  Copy full SHA for 0b6410b

Commits on Apr 20, 2026

• Update README and ignore kernels in CI (#387)

  eltsai

  authored
  bc3bca0

  Copy full SHA for bc3bca0

• PR #359: Ring attention integration and other optimizations

  Show description for c60e65d
  Google-ML-Automation

  committed
  c60e65d

  Copy full SHA for c60e65d

• LTX2 Performance enhancements

  prishajain1

  committed
  77973e3

  Copy full SHA for 77973e3

• Fix batch size calculation in Wan VACE pipeline for precomputed embeddings

  Show description for b4fd384
  ninatu

  committed
  b4fd384

  Copy full SHA for b4fd384

Commits on Apr 19, 2026

• fix(ltx2): resolve flash attention block size mismatch and missing config overrides

  Show description for 490dd70
  mbohlool

  committed
  490dd70

  Copy full SHA for 490dd70

Commits on Apr 17, 2026

• Merge pull request #381 from AI-Hypercomputer:wan2p2vae

  Show description for ad6391a
  Google-ML-Automation

  committed
  ad6391a

  Copy full SHA for ad6391a

• Merge pull request #379 from AI-Hypercomputer:ninatu/fix-sharding-mismatch

  Show description for 2965670
  Google-ML-Automation

  committed
  2965670

  Copy full SHA for 2965670

• Merge pull request #376 from AI-Hypercomputer:ulysses-attention-benchmark

  Show description for 702cadd
  Google-ML-Automation

  committed
  702cadd

  Copy full SHA for 702cadd

• Fix: Overhaul WAN checkpointers for robust multi-host restoration

  Show description for 69f7701
  

  ninatu

  and

  martinarroyo

  committed
  69f7701

  Copy full SHA for 69f7701

Commits on Apr 16, 2026

• add support for wan vae 2.2 & fix hacky wan vae 2.1

  jzr99

  committed
  1da5d2f

  Copy full SHA for 1da5d2f

• fix wan command for generation. (#380)

  Show description for 18f6f0f
  

  entrpn

  and

  jfacevedo-google

  authored
  18f6f0f

  Copy full SHA for 18f6f0f

• Add Ulysses attention support

  csgoogle

  committed
  a4e0ae7

  Copy full SHA for a4e0ae7

• Merge pull request #377 from AI-Hypercomputer:mldiag

  Show description for 7f698e3
  Google-ML-Automation

  committed
  7f698e3

  Copy full SHA for 7f698e3

• Merge pull request #371 from AI-Hypercomputer:fixbiassharding

  Show description for a3747b1
  Google-ML-Automation

  committed
  a3747b1

  Copy full SHA for a3747b1

Commits on Apr 15, 2026

• Fix transformer sharding and cross-attention flash block sizes

  csgoogle

  committed
  2ddf8ab

  Copy full SHA for 2ddf8ab

Commits on Apr 14, 2026

• Add Google Cloud ML Diagnostics and profiling support

  Show description for 1d35d2f
  mbohlool

  committed
  1d35d2f

  Copy full SHA for 1d35d2f

• Merge pull request #353 from AI-Hypercomputer:ninatu/grad_clip

  Show description for 384d211
  Google-ML-Automation

  committed
  384d211

  Copy full SHA for 384d211

• Merge pull request #351 from AI-Hypercomputer:ninatu/logging

  Show description for 2d237f3
  Google-ML-Automation

  committed
  2d237f3

  Copy full SHA for 2d237f3

• Add gradient clipping options to optimizer

  Show description for 35076d3
  

  ninatu

  and

  martinarroyo

  committed
  35076d3

  Copy full SHA for 35076d3

• Refactor: Output maxlogging to both logger and stdout

  Show description for bf2a842
  

  ninatu

  and

  martinarroyo

  committed
  bf2a842

  Copy full SHA for bf2a842

Commits on Apr 13, 2026

• Merge pull request #364 from AI-Hypercomputer:sagarchapara/wananimate

  Show description for 46cae70
  Google-ML-Automation

  committed
  46cae70

  Copy full SHA for 46cae70

• Remove accidentally committed file with invalid space-only filename (#375)

  Show description for d799d57
  Perseus14

  authored
  d799d57

  Copy full SHA for d799d57

• Merge pull request #372 from AI-Hypercomputer:ltx2_lora

  Show description for 07f56d2
  Google-ML-Automation

  committed
  07f56d2

  Copy full SHA for 07f56d2

• adding missing dep for nightly regression tests (#374)

  prishajain1

  authored
  12b267c

  Copy full SHA for 12b267c

• Merge pull request #352 from AI-Hypercomputer:ninatu/wan_training

  Show description for 66859e9
  Google-ML-Automation

  committed
  66859e9

  Copy full SHA for 66859e9

• add wan animate transformer

  csgoogle

  committed
  6f5e027

  Copy full SHA for 6f5e027

• Improve Wan model training and add Wan-VACE training support.

  Show description for 58448d2
  

  ninatu

  and

  martinarroyo

  committed
  58448d2

  Copy full SHA for 58448d2

• Add LTX2 LoRA inference support

  prishajain1

  committed
  9810681

  Copy full SHA for 9810681

Commits on Apr 10, 2026

• Merge pull request #373 from AI-Hypercomputer:prisha/ltx2_fixes

  Show description for ceca471
  Google-ML-Automation

  committed
  ceca471

  Copy full SHA for ceca471

• use State instead of GraphState

  Show description for f04046c
  

  Cristian Garcia

  authored and

  Google-ML-Automation

  committed
  f04046c

  Copy full SHA for f04046c

Commits on Apr 9, 2026

• Merge pull request #366 from AI-Hypercomputer:mehdy-temp-upscaler

  Show description for 993a5a6
  Google-ML-Automation

  committed
  993a5a6

  Copy full SHA for 993a5a6

Commits on Apr 8, 2026

• LTX-2 Minor Fixes

  prishajain1

  committed
  c5bb862

  Copy full SHA for c5bb862