Stable Diffusion Resources

The following resources are available for members.

  1. Training a LoRA model for Stable Diffusion 1.5 – training notebook and training images
  2. Training a LoRA model for Stable Diffusion XL – training notebook and training images
  3. Training a checkpoint model for Stable Diffusion 1.5 – training notebook and training images
  4. Bare-bone, reusable ComfyUI workflow examples

Train LoRA for Stable Diffusion 1.5

To read this content, become a member of this site

Already a member? Log in here.

Train LoRA for Stable Diffusion XL

To read this content, become a member of this site

Already a member? Log in here.

Dreambooth checkpoint training

To read this content, become a member of this site

Already a member? Log in here.

ComfyUI workflows

Text-to-image

Text-to-image workflow (SD 1.5)

Text-to-image workflow (SDXL)

Text-to-image: Compare two models

Text-to-image workflow using LCM LoRA to speed up image generation.

Text-to-image workflow SDXL base + refiner models.

Compare a text-to-image workflow with and without an embedding or keywords.

Condition Average (Prompt interpolation)

Clip Skip

Utilities

Join two prompts or strings.

Image-to-image

Basic Image-to-image workflow

LoRA

Text-to-image with LoRA

Text-to-image with 2 LoRAs

Image-to-image with LoRA

Image-to-image with 2 LoRAs

Compare with and without a LoRA in text-to-image.

Inpainting

Inpaint with a Standard Stable Diffusion model (v1.5 or SDXL)

Inpaint with a Stable Diffusion v1.5 inpainting model

Inpaint with ControlNet Inpainting preprocessor

Text-to-image with automatic inpainting using FaceDetailer in the Impact Pack

Automatic inpainting on an uploaded image using FaceDetailer in the Impact Pack

Upscaling

Text-to-image with an AI upscaler model.

Text-to-image with an AI upscaler model and resize (CR Upscale Image)

Text-to-image with Hi Res Fix (Latent upscale)

Image-to-image with Hi Res Fix (Latent upscale)

Text-to-image with Ultimate SD Upscale (AI upscaler)

Text-to-image with Ultimate SD Upscale (ControlNet Tile)

ControlNet

ControlNet Openpose

ControlNet Openpose Hi Res Fix

ControlNet DWPose

ControlNet Canny

ControlNet Line art/ Line art coarse

ControlNet Line art anime

ControlNet Line art HED

ControlNet Depth Midas

ControlNet Depth Zoe

ControlNet Depth LeReS

ControlNet Depth Anything

2 ControlNets – openpose and depth

2 ControlNets – Hi Res Fix

IP-adapter

See this table for model download and pairing.

IP-adapter original, Plus, or Plus Face (SD 1.5)

IP-adapter, ViT H, or Plus ViT H (SDXL)

IP-adapter Face ID, Face ID Plus, or Plus v2 (SD 1.5)

IP-adapter Face ID or Face ID Plus v2 (SDXL)

IP-adapter Face ID portrait (SD 1.5)

Compare IP-adapter Face and Face ID (SD 1.5)

Stable Diffusion Video

SDV text-to-video

AnimateDiff

AnimateDiff text-to-video

Stable Cascade

Stable Cascade text-to-image

Stable Cascade image-to-image

Stable Cascade image prompt