This repository holds workflows crafted for generating AI images through Stable Diffusion using ComfyUI. The workflows are arranged for clarity and ease of use.
otonx_sdxl_base+lora+controlnet+refiner+upscale+facedetail_workflow
This workflow incorporates SDXL models with a refiner. Users have the option to add LoRAs, ControlNet models or T21 Adapters, and an Upscaler. The ControlNet / T21 section is implemented as a Switch logic, allowing users to select between ControlNet models or T21 adapters. After the refiner KSampler, the users have the options to do face fix/detailing.
otonx_sdxl_base+refiner+lcmlora+facedetailer_workflow
My attempt to use LCM Lora with SDXL models.
- ComfyUI: An intuitive interface for crafting AI art pipelines.
- Stable Diffusion: A technique for AI art generation, enhanced by SDXL models.
- SDXL Models: A two-step process with a base model and a refiner model.
- ControlNet Models: Extend diffusion models to allow conditional inputs.
- T21 Adapters: Enable text-to-image conversion efficiently.
- LoRAs and Upscalers: Additional modules for refining and upscaling images.
- Aspect Ratio: The aspect ratio recommended for the Stable Diffusion XL model can be set.
- LoRA Stack: Multiple LoRAs can be used through the LoRA Stack.
- Custom Nodes: A set of custom nodes have been created and are utilized in this project. They can be found at comfyui_otonx_nodes.
- Clone the repository.
- Navigate to the
workflows
directory to access workflow files. - Load the desired workflow file in ComfyUI.
- Modify the workflow nodes and groups as needed.
- Start generating AI art.