Commits

Commit history

1. Commits on Oct 10, 2026

   • Version bump

     jaretburkett

     committed
     f7a1fb9View commit details

     Copy full SHA for f7a1fb9

     Browse repository at this point

   • Add support for Qwen Image 2 Turbo with training adapter

     jaretburkett

     committed
     f370e74View commit details

     Copy full SHA for f370e74

     Browse repository at this point
2. Commits on Oct 9, 2026

   • Version bump

     jaretburkett

     committed
     2fde764View commit details

     Copy full SHA for 2fde764

     Browse repository at this point

   • Add support for Iris-3B

     jaretburkett

     committed
     2fa0bd8View commit details

     Copy full SHA for 2fa0bd8

     Browse repository at this point
3. Commits on Oct 8, 2026

   • Add ability to merge dataset buckets. Useful when using numerous different datasets.

     jaretburkett

     committed
     a949f21View commit details

     Copy full SHA for a949f21

     Browse repository at this point

   • Add threadding to bucket resizing of a batch. Helpful to speed up large batch gathering.

     jaretburkett

     committed
     0cbac2cView commit details

     Copy full SHA for 0cbac2c

     Browse repository at this point
4. Commits on Sep 27, 2026

   • Key table rows by id so deleted rows don't leak pending state to neighbors

     jaretburkett

     committed
     ecee894View commit details

     Copy full SHA for ecee894

     Browse repository at this point
5. Commits on Sep 26, 2026

   • Fix issue loading new comfy checkpoints with ming

     jaretburkett

     committed
     6468a2fView commit details

     Copy full SHA for 6468a2f

     Browse repository at this point

   • Fixed issue on DGX spark that caused unnessary moves of models causing slow down on unified memory

     jaretburkett

     committed
     16ae8d2View commit details

     Copy full SHA for 16ae8d2

     Browse repository at this point
6. Commits on Sep 25, 2026

   • Switch to new training adapters for all minimax models that no longer need contrastive guidance

     jaretburkett

     committed
     60d0c28View commit details

     Copy full SHA for 60d0c28

     Browse repository at this point

   • Add support to train on Fast H3 v2 8 step directly with training adapter

     jaretburkett

     committed
     a776394View commit details

     Copy full SHA for a776394

     Browse repository at this point
7. Commits on Sep 24, 2026

   • Update ui to show DGX spark gpu usage correctly

     jaretburkett

     committed
     5aa3d0cView commit details

     Copy full SHA for 5aa3d0c

     Browse repository at this point

   • Add proper support for DGX spark with easy manager. Thank you NVIDIA!

     jaretburkett

     committed
     60ca869View commit details

     Copy full SHA for 60ca869

     Browse repository at this point

   • Update ming image repo to new one

     jaretburkett

     committed
     cbc6bbdView commit details

     Copy full SHA for cbc6bbd

     Browse repository at this point
8. Commits on Sep 23, 2026

   • Handle vision tokens in ming

     jaretburkett

     committed
     460c29bView commit details

     Copy full SHA for 460c29b

     Browse repository at this point

   • Add support for Ming-Image-0.1-Design with training adapter

     jaretburkett

     committed
     77847d7View commit details

     Copy full SHA for 77847d7

     Browse repository at this point
9. Commits on Sep 22, 2026

   • Version Bump

     jaretburkett

     committed
     0bd3411View commit details

     Copy full SHA for 0bd3411

     Browse repository at this point

   • Add match_target_res for Qwen Image 2.1 so reference images are scaled to the target's pixel area with their own aspect ratio, consistently across text-embedding caching, training, and sampling, an…

     Show description for 07abdbe
     jaretburkett

     committed
     07abdbeView commit details

     Copy full SHA for 07abdbe

     Browse repository at this point

   • Fix support link

     jaretburkett

     committed
     71bc877View commit details

     Copy full SHA for 71bc877

     Browse repository at this point

   • Merge pull request #1055 from vdruts/yue2-safetensors-heads

     Show description for 8f2777a
     jaretburkett

     authored
     8f2777aView commit details

     Copy full SHA for 8f2777a

     Browse repository at this point

   • Fix Qwen Image Edit Plus batch_size > 1 without cached text embeddings by splitting the batch-stacked control tensor into one control image per prompt.

     jaretburkett

     committed
     f073188View commit details

     Copy full SHA for f073188

     Browse repository at this point

   • Fix text encoder unload skipping models whose pipeline does not expose the text encoder (Qwen-Image-2.1, Z-Image, LTX2), so cached-embedding runs actually free the TE.

     jaretburkett

     committed
     a35d833View commit details

     Copy full SHA for a35d833

     Browse repository at this point

   • Fix issue with saving device state when a text encoder is unloaded.

     jaretburkett

     committed
     543a0d7View commit details

     Copy full SHA for 543a0d7

     Browse repository at this point

   • Merge pull request #1036 from ArcGleam/fix/qwen-edit-plus-batch-gt-1

     Show description for ae6ab90
     jaretburkett

     authored
     ae6ab90View commit details

     Copy full SHA for ae6ab90

     Browse repository at this point

   • Merge pull request #1035 from ArcGleam/fix/qwen-edit-plus-control-image-inplace-downsample

     Show description for 9b67238
     jaretburkett

     authored
     9b67238View commit details

     Copy full SHA for 9b67238

     Browse repository at this point

   • Merge pull request #1021 from hitoshic1982/fix/cache-text-encoder-device-preset

     Show description for 3a39d8c
     jaretburkett

     authored
     3a39d8cView commit details

     Copy full SHA for 3a39d8c

     Browse repository at this point

   • Ignore issue review folder

     jaretburkett

     committed
     11b65a3View commit details

     Copy full SHA for 11b65a3

     Browse repository at this point
10. Commits on Sep 21, 2026

    • YuE2: load safetensors semantic heads / NAR adapters (v5, v8, v9) and key the latent cache by head

      vdruts

      committed
      d0ecde7View commit details

      Copy full SHA for d0ecde7

      Browse repository at this point
11. Commits on Sep 20, 2026

    • Save thumbnail with alpha if there is one.

      jaretburkett

      committed
      a8dfcf7View commit details

      Copy full SHA for a8dfcf7

      Browse repository at this point

    • Add support for training with an alpha channel for Qwen Image 2.1

      jaretburkett

      committed
      086b663View commit details

      Copy full SHA for 086b663

      Browse repository at this point

    • Version bump

      jaretburkett

      committed
      31ddc70View commit details

      Copy full SHA for 31ddc70

      Browse repository at this point

    • Change default guidance to 3 for qwen image.

      jaretburkett

      committed
      474948aView commit details

      Copy full SHA for 474948a

      Browse repository at this point

    • Change qwen image 2 to use new lokr format

      jaretburkett

      committed
      f4f5d5dView commit details

      Copy full SHA for f4f5d5d

      Browse repository at this point

    • Add support for Qwen Image 2.1

      jaretburkett

      committed
      c2622edView commit details

      Copy full SHA for c2622ed

      Browse repository at this point
12. Commits on Sep 19, 2026

    • Move the ui model cards to their own plugin folders so plugins can add custom models in the trainer/generator/ui

      jaretburkett

      committed
      8fa15e3View commit details

      Copy full SHA for 8fa15e3

      Browse repository at this point