v1.0.0

Features

• [Feature] Update diffusers IP Adapter #99
• [Feature] Support Style LoRA #102
• [Feature] Supoport Instruction-tuning SDXL #103
• [Feature] Support xformers #105
• [Feature] Support PixArt Alpha #106
• [Feature] Support OFT #107
• [Feature] Support PixArt DreamBooth Style #108
• [Feature] Support DeepSpeed/CorossalAI #109
• [Feature] Support ControlNetXS #110
• [Feature] Support Inpaint #111
• [Feature] Support RandomChoice for random mask #112
• [Feature] Support whole mask #113
• [Feature] Support Finetune pretrained IP-Adapter #115
• [Feature] Support Kandinsky v2.2 #116
• [Feature] Support Kandinskyv3 #117
• [Feature] Support pre compute bucket ids #118
• [Feature] Fix like xtuner style repo #119
• [Feature] Pure python config #121
• [Feature] Use default args for pretrained weights name #122

Fix

• [Fix] LoRA default alpha #104
• [Bugfix] Inpaint masked image #114

Docs

• [Docs] ControlNet data prepare #100
• [Docs] Aspect Ratio Bucketing #101
• [Enhance] Update docs #120

v0.3.0

Features

• [Feature] Support DeepFloyd IF #76
• [Feature] Support DeBias Estimation Loss #79
• [Feature] Support SSD-1B #83
• [Feature] Faster training #84 [Feature] Faster train with torch.compile #85 [Enhance] Support compile mode #86 [Enhance] Support fast norm controlnet #88 [Feature] Support compile controlnet and fused GN #95
• [Feature] Support Noise Methods #87
• [Feature] Support InstructPix2Pix #89
• [Feature] Support TimeSteps Bias #90
• [Feature] Wuerstchen #91
• [Feature] Support Peft #92
• [Feature] Support v_prediction #94
• [Feature] Support LCM #96

Enhance

• [Enhance] Add model zoo #75
• [Feature] Apply more ruff rules #77
• [Enhance] Support IP Adapter Batch Size>1 #78
• [Enhance] Use pyproject toml #80
• [Enhance] Update docs #81
• [Enhance] Add blog post #82
• [Fix] LoRA docs #93

v0.2.1

Fix

• [Enhance] Fix tests lint #74
• [Enhance] Fix config lint #72

v0.2.0

What's Changed

• [Project] Face Expression #46
• [Feature] Add csv param to datasets #47
• [Enhance] Update face expression project sdxl lora #48
• [Feature] Support ControlNet Small #49
• [Feature] Support multi aspectratio resize #50
• [Feature] Support zunko lora projects #51
• [Feature] Support google/dreambooth datasets #52
• [Enhance] Support devcontainer #57
• [Enhance] Update libs #58
• [Feature] IP Adapter #59
• [Feature] Pre compute Text Embeddings #60
• [Feature] Support IP Adapter Plus #61
• [Enhance] Fix more epochs for IP Adapter Plus #62
• [Feature] Support negative prompt for inference #63
• [Feature] Support T2I Adapter #64
• [Feature] Support Erasing Concepts from Diffusion Models #65
• [Feature] Support IP Adapter Pipeline for Inference #66
• [Feature] Support Distill SDXL training #67
• [Feature] Support show vae tool #68
• [Feature] Support Ruff #69
• [Feature] support mypy #70

Fix

• [Fix] Fix README.md #41
• [Fix] minor fix #45
• add sponsor #54

v0.1.4

What's Changed

• [Feature] Publish docs #35
• fix typos #36 @omahs
• [Feature] api docs #37
• [Bugfix] Delete LoRA EMA #38
• [Feat] Colaboratory Example #39

v0.1.3

What's Changed

• [Feature] Support ControlNet #31
• [Feature] Human face DreamBooth with Distill SD. #32

v0.1.2

What's Changed

• [Enhance] Update DreamBooth #21
• [Feature] Support SDXL DreamBooth #22
• [Feature] Refactor Stable Diffusion and LoRA #23
• [Feature] Refactor SDXL and SDXL-LoRA #25
• [Enhance] Update README #26 Thank you for the issue #24 from @Ezra-Yu
• [Feature] Support Distill SD DreamBooth #27 [Enhance] Add training time for distill sd #29

v0.1.1

What's Changed

• [Docs] Update get started #15
• [Docs] Add config docs #16
• [BugFix] DreamBooth Dataset #17
• [Enhance] Add Example Notebook #19 Thank you for the issue #14 from @andysingal
• [Bump] v0.1.1 #20

v0.1.0

First release

What's Changed

• [Feature] Add Stable Diffusion v15 training #1
• [Feature] SNR weighting gamma L2 loss #2
• [Feature] Add inference demo #3
• [Feature] Add Unet EMA training #4
• [Feature] Add LoRA #5
• [Feature] Support train text_encoder #6
• [Feature] Support SDXL #7
• [Feature] Support DreamBooth #8
• [Bugfix] crop top left calc #9
• Add .dockerignore #10
• [Feature] Support local data #11
• [Feature] Support DDP #12
• [Docs] prepare dataset #13