Commits

Branch selector

 main

User selector

bghira

Datepicker

All time

Commit History

Commits on Jan 8, 2025

• fix for #7365, prevent pipelines from overriding provided prompt embeds (#7926)

  Show description for a0acbdc
  

  4 peopleauthored
  a0acbdc

  Copy full SHA for a0acbdc

Commits on Nov 19, 2024

• add skip_layers argument to SD3 transformer model class (#9880)

  Show description for 99c0483
  

  3 peopleauthored
  99c0483

  Copy full SHA for 99c0483

Commits on Jun 18, 2024

• SD3: update default training timestep / loss weighting distribution to logit_normal (#8592)

  Show description for 074a7cc
  

  3 peopleauthored
  074a7cc

  Copy full SHA for 074a7cc

Commits on May 7, 2024

• 7879 - adjust documentation to use naruto dataset, since pokemon is now gated (#7880)

  Show description for 8edaf3b
  

  bghira

  and

  bghira

  authored
  8edaf3b

  Copy full SHA for 8edaf3b

Commits on Apr 2, 2024

• 7529 do not disable autocast for cuda devices (#7530)

  Show description for 8e963d1
  

  3 peopleauthored
  8e963d1

  Copy full SHA for 8e963d1

Commits on Mar 28, 2024

• apple mps: training support for SDXL (ControlNet, LoRA, Dreambooth, T2I) (#7447)

  Show description for d78acde
  

  3 peopleauthored
  d78acde

  Copy full SHA for d78acde

Commits on Mar 26, 2024

• diffusers#7426 fix stable diffusion xl inference on MPS when dtypes shift unexpectedly due to pytorch bugs (#7446)

  Show description for 544710e
  

  3 peopleauthored
  544710e

  Copy full SHA for 544710e

Commits on Oct 5, 2023

• Min-SNR Gamma: correct the fix for SNR weighted loss in v-prediction … (#5238)

  Show description for 02a8d66
  

  3 peopleauthored
  02a8d66

  Copy full SHA for 02a8d66

Commits on Sep 26, 2023

• Min-SNR Gamma: follow-up fix for zero-terminal SNR models on v-prediction or epsilon (#5177)

  Show description for 4a06c74
  

  3 peopleauthored
  4a06c74

  Copy full SHA for 4a06c74

• Timestep bias for fine-tuning SDXL (#5094)

  Show description for 89d8f84
  

  3 peopleauthored
  89d8f84

  Copy full SHA for 89d8f84

Commits on Sep 25, 2023

• SDXL microconditioning documentation should indicate the correct default order of parameters, so that developers know (#5155)

  Show description for 539846a
  

  3 peopleauthored
  539846a

  Copy full SHA for 539846a

Commits on Sep 21, 2023

• Min-SNR gamma support for Dreambooth training (#5107)

  Show description for d558811
  

  3 peopleauthored
  d558811

  Copy full SHA for d558811

Commits on Sep 20, 2023

• SNR gamma fixes for v_prediction training (#5106)

  Show description for 24563ca
  

  bghira

  and

  bghira

  authored
  24563ca

  Copy full SHA for 24563ca

Commits on Sep 19, 2023

• Resolve v_prediction issue for min-SNR gamma weighted loss function (#5096)

  Show description for 74e43a4
  

  3 peopleauthored
  74e43a4

  Copy full SHA for 74e43a4

• Add x-prediction / prediction_type=sample support for SDXL fine-tuning (#5095)

  Show description for 81331f3
  

  3 peopleauthored
  81331f3

  Copy full SHA for 81331f3

Commits on Sep 15, 2023

• Remove logger.info statement from Unet2DCondition code to ensure torch compile reliably succeeds (#4982)

  Show description for a019867
  

  3 peopleauthored
  a019867

  Copy full SHA for a019867

Commits on Sep 5, 2023

• Add --vae_precision option to the SDXL pix2pix script so that we have… (#4881)

  Show description for cfdfcf2
  

  bghira

  and

  bghira

  authored
  cfdfcf2

  Copy full SHA for cfdfcf2

Commits on Jul 12, 2023

• [SDXL] Partial diffusion support for Text2Img and Img2Img Pipelines (#4015)

  Show description for 99b540b
  

  5 peopleauthored
  99b540b

  Copy full SHA for 99b540b