lora
lora copied to clipboard
cloneofsimo
The selected model is not LoRA or not trained by `sd-scripts`? // error in automatic1111 UI with LoRa
The selected model is not LoRA or not trained by `sd-scripts`?
create LoRA for Text Encoder: 72 modules.
create LoRA for U-Net: 192 modules.
LoRa was trained via this collab
Referencing this GitHub
When I load this LoRa in Automatic1111 UI, this error happens.
How it can be fixed?
There's this conversion script but haven't really tried it yet https://gist.github.com/laksjdjf/dc43fc6c5cb4d9000fbbf18515f1e4d8
Compatibility is a current issue. Currently, you can merge it to ckpt and use it, although that will remove the nice feature of LoRA. I'll work on it in the future.
@oscarnevarezleal @cian0 @cloneofsimo @CombainerA19 @anotherjesse Hi, dears. I have add some code to support the Automatic111 WebUI. What you should do is just runing 'train_lora_dreambooth.py', then you will find that there is another lora_weight_webui.safetensors in the output dir. That is the right weight for WebUI. reference my project in https://github.com/tengshaofeng/lora_tbq
Oh nice, but how about the textual inversion that gets trained? And is it compatible with dev branch?
train_lora_dreambooth.py
im sorry but the instructions are very poor, how exactly do you use this to convert a trained lora of sd 2.1 for use in webui?
train_lora_dreambooth.py
im sorry but the instructions are very poor, how exactly do you use this to convert a trained lora of sd 2.1 for use in webui?
It is support for sd1.x, but not sure for sd2.1. You can modify that base my code.
@oscarnevarezleal @cian0 @cloneofsimo @CombainerA19 @anotherjesse Hi, dears. I have add some code to support the Automatic111 WebUI. What you should do is just runing 'train_lora_dreambooth.py', then you will find that there is another lora_weight_webui.safetensors in the output dir. That is the right weight for WebUI. reference my project in https://github.com/tengshaofeng/lora_tbq
The instruction is very poor
