You can train LORA models using dreambooth extension in the A1111 sd webui, the real question is how to obtain the 144MB lora model/embedding you can use with ease with pretty much any SD model (via additional networks button) instead of the small 5MB .pt which can't be used this way. Any suggestions on that?

@mofme69 I was asking about this and what settings)

@vell I figured out how to make "Extra models" compatible LoRA out of Dreambooth Extension trained LoRA - you simply need to extract LoRA from a checkpoint generated during training using this toolset: https://github.com/bmaltais/kohya_ss and save the extracted LoRA to /models/Lora directory. Works as charm.

@mofme69 Thanks, I'll try

@mofme69 in the gitub, Installation instruction for kohya, one can read :

Open a regular user Powershell terminal and type the following inside: [...]

Could you please answer those few questions, as the installation is not as obvious as it seems to be for some of us :

- Once in the powershell, am I supposed to copy paste a line from the Installation instructions, wait for it to finish, then copy paste and run the next line ? Can I (should I) copy paste the whole thing and let it run ?

- And more important : should I first be in a specific folder of my hard drive (withing the PowerShell) before running those lines ?

Those Lora are a f@#cking revolution by the way. No need to merge checkpoints anymore. Just need to prompt multiple Loras to get a "merged" result during the generation. And it saves loads of space too ;)

Waiting forward for your help, I'd love to participate to this heavy Lora posting of models extracted.

Cheers.

@hansolocambo
> - Once in the powershell, am I supposed to copy paste a line from the Installation instructions, wait for it to finish, then copy paste and run the next line ?

Yes, exactly like that. Execute line-by-line.

> - And more important : should I first be in a specific folder of my hard drive (withing the PowerShell) before running those lines ?

You should be inside the folder of the cloned repository. As per the instructions:
git clone https://github.com/bmaltais/kohya_ss.git <<< this clones the repository to your hard drive

cd kohya_ss <<< this enters the cloned repository