LoRA training

From Interaction Station Wiki
Jump to navigation Jump to search

Training a LoRA with Kohya

These instructions should work on the computers in WH.02.110

Preparing for training

Step 1: Collect the images you want to use for training

Step 2: Create the folder structure
Create a folder on the desktop and name it (your name, or name of your project).
Download this zip file File:Folder structure and configuration file.zip , and unzip it in the folder you just created. The zip will create the basic file structure, and a configuration file.

Folder structure 1.png

Inside the img folder, there needs to be another folder. The name of this folder has to be <REPEATS><underscore><TRIGGERWORD><space><CLASS>

Folder structure 2.png

Copy your images into this folder

Images in folder.png

Step 3: Create captions

We are now ready to create the captions for each image.
On the desktop find the kohya icon, and double click it.

Kohya icon.png

A terminal window opens, and after a couple of seconds you should see a url (127.0.0.1:7860) that you can open in a browser to get a GUI. In the GUI select the Utilities Tab, and inside that the Captioning Tab and there select BLIP Captioning.
Make sure the "Image folder to caption", points to the folder containing your images. Also put your trigger word in "Prefix to add to BLIP caption". Now click the "Caption Images" button

Captioning.png

You won't see anything happening in the GUI but there should be some activity in the terminal

Captioning progress terminal.png

Created captions.png

Training your LoRA

Step 4: Adjust the configuration If you didn't already launch the kohya software

Make sure you select the LoRA tab

Load the configuration file that you downloaded earlier. We are going to leave it mostly at the default settings, but there are a few things we need to adjust.

Basic configuration.png

Change the model name to the name you want it to have, and make sure the paths to the image folder and model folder are correct.

In the parameters section we can configure to generate sample images during training. Adjust the prompts so they use your triggerword and suit your purpose.

Sample prompts.png

Step 5: Start training Hit the training button, and wait for the first samples to appear. (A sample folder will be created in the model folder)

Training terminal.png

Samples.png

Using your LoRA in Stable Diffusion WebUI

Step 6: Copy your LoRA to Stable Diffusion

Select the LoRA model in the model folder that looks most promising to you according to the samples, and copy it to the LoRA folder in Stable Diffusion WebUI. That folder is located in /home/interactionstation/MachineLearning/stable-diffusion-webui/model/Lora

Select LoRA.png

Sd lora folder.png

Step 7: Test your LoRA In the WebUI check the LoRA tab, and refresh it to see your model. Clicking on it will add your LoRA to the prompt.

Sd lora tab.png

Step 8: Play with your LoRA

You can use your LoRA in combination with different compatible checkpoint models.

Change checkpoint model.png

You can also combine multiple LoRA models, and change the weight of the LoRA. Also see what happens when you change the cfg scale.

Multiple LoRA's.png