Merge branch 'main' of https://github.com/Scorpinaus/ComfyUI-DiffusersLoader
This commit is contained in:
@@ -33,9 +33,11 @@ Follow these steps to install and set up the project:
|
||||
|
||||
## Features
|
||||
The CombinedDiffusersLoader supports loading of diffusers checkpoints for:
|
||||
- SD1.5
|
||||
- SD 1.5
|
||||
- SD 2.1
|
||||
- SDXL
|
||||
- SD3
|
||||
- AuraFlow
|
||||
|
||||
### Workflow Example
|
||||
The combined loader work in the same manner as existing checkpoints loader as seen in this workflow:
|
||||
@@ -49,7 +51,8 @@ Take note for SD3:
|
||||
- Selecting all will create a combined T5 encoder in the same folder named combined_text_encoder.safetensors for the first use.
|
||||
|
||||
## Limitations & Future Improvements
|
||||
Add support for AuraFlow and other compatible diffusers format checkpoints
|
||||
Add support for other compatible diffusers format checkpoints
|
||||
For AuraFlow and SD3, the loading and combining for the combined_model.safetensors will have a delay for low-vram or low-system ram machines
|
||||
|
||||
|
||||
## Contributing
|
||||
|
||||
Reference in New Issue
Block a user