search

vladmandic / automatic

SD.Next: Advanced Implementation of Stable Diffusion and other Diffusion-based generative image models
https://github.com/vladmandic/automatic
GNU Affero General Public License v3.0
5.3k stars 377 forks source link

[Feature]: add support for third party stable cascade finetunes #3311

Open A500000 opened 6 days ago

A500000 commented 6 days ago

Feature description

This would add support for using my own stable cascade safetensors finetunes instead of the hf blobs currently being used

Version Platform Description

No response

vladmandic commented 5 days ago

you should be able to load any stable cascade finetunes as-is? copying @Disty0 for more info as he did a SC finetune himself.

Disty0 commented 5 days ago

Load the Stable Cascade base (the decoder) from Huggingface as the main model first, then load a Unet (prior) as the Unet.

Put the Unet safetensors in Unet folder and put the text encoder (if you use one) in there too. Text encoder name should be the Unet Name + _text_encoder

Example Unet: sc_unet.safetensors Example TE: sc_unet_text_encoder.safetensors