How To Do SDXL LoRA Training On RunPod With Kohya SS GUI Trainer & Use LoRAs With Automatic1111 UI

If you don't have a GPU, or have a strong GPU, or you are using Mac and your computer not supporting Stable Diffusion training, SDXL training, then this is the tutorial you are looking for. In this tutorial, we will use a cheap cloud GPU service provider RunPod to use both Stable Diffusion Web UI Automatic1111 and Stable Diffusion trainer Kohya SS GUI to train SDXL LoRAs. By watching this tutorial, you will be able to training and generate images as easy as your own computer by using RunPod cloud service.

1 click Auto Kohya Installer ⤵️

Tutorial GitHub Readme File ⤵️

SECourses Discord To Get Full Support ⤵️

My LinkedIn ⤵️

My Instagram ⤵️

My Medium ⤵️

My CivitAI ⤵️

Hopefully I will do more research about SDXL training. Will investigate training only unet without text encoder. Moreover, I will investigate and make a workflow about celebrity name based training hopefully. Furthermore, SDXL full DreamBooth training is also on my research and workflow preparation list. Stay subscribed for all.

0:00 Introduction to easy tutorial of using RunPod to do SDXL training
1:55 How to start your RunPod machine for Stable Diffusion XL usage and training
3:18 How to install Kohya on RunPod with a single click installer
4:22 Important things about using SDXL on RunPod
5:38 Step by step installation of Kohya SS GUI trainer on RunPod
7:08 How to terminate automatically started Automatic1111 Instance to free up VRAM
8:00 Where do you need to download and put Stable Diffusion model and VAE files on RunPod
8:58 How to download Stable Diffusion models into RunPod
9:25 How to terminate previously started Automatic1111 Web UI instance to free up VRAM before training,
10:03 How to start Kohya GUI after installation
10:42 How to save and load Kohya training configuration on RunPod
10:54 How to upload and set your training images
12:01 How to install runpodctl
13:15 Why runpodctl is useful
13:51 How to configure Kohya LoRA training on RunPod for Stable Diffusion XL
15:58 What is the importance and effect of Network Rank (Dimension)
19:41 How to update Automatic1111 Web UI on RunPod
20:04 How to understand your Kohya training completed or not
21:57 How to start using your trained or downloaded SDXL LoRA models
22:13 Where the training checkpoint files are saved
22:46 How you should connect to Automatic1111 Web UI interface on RunPod for image generation
23:15 How to set best Stable Diffusion VAE file for best image quality
23:33 How to set full precision VAE on RunPod --no-half-vae
24:18 Where to find good Stable Diffusion prompts for SDXL and SD 1.5 models
25:05 How to activate your trained LoRA via appending to the prompt
25:47 How to improve faces / fix them via using Adetailer automatic inpainting extension
26:44 How to reload last used prompt and all settings with 1 click
26:52 How to enable and use adetailer extension
27:06 In which folder generated images saved and how to download them very fast on RunPod
29:24 How to test face quality and verify your trained LoRA or DreamBooth models / checkpoints
30:10 How to do prompt engineering to improve and fix faces
32:06 How to join our community helping Discord and where to contact to me

[SECourses]

If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on 🥰 ⤵

Technology & Science: News, Tips, Tutorials, Tricks, Best Applications, Guides, Reviews ⤵

Playlist of StableDiffusion Tutorials, Automatic1111 and Google Colab Guides, DreamBooth, Textual Inversion / Embedding, LoRA, AI Upscaling, Pix2Pix, Img2Img ⤵

[bkhyjly]

Hello, can you please provide your own training images, so I can test performance with different set of training images?

[JohanMartinez-ocpu]

I was working on a Lora for stable diffusion xl 1, and I got stuck at the part of updating automatic 111 web UI. I couldn't move forward from there, and I'm very frustrated. Could someone please help me? I followed the manual installation from the video. I'm at the 20:15 mark in the tutorial. Anyone who could please help me update the automatic 1111 on Runpod, please.

[MinuteScalping]

is there a way to train using runpod using custom model, i downloaded with wget to model folder, i get an error, i can only use the preset default model

[patriciaanaka]

I have been using that ADetailer to fix far-off faces and it usually has worked great. Today, though, it has stopped working. The little red rectangle does not appear over the faces and they don't appear to be fixed. Any idea why that would happen? I didn't do anything different and I've tried on two different pods...

[godobegghins]

When the training starts, it gives this warning: one of the inputs have requires_grad=True. Gradients will be None. Will it have effect on the lora results?

[eugenbuzuk]

Could you please advice kohya settings to train very detailed texture, for example sand. Because i've got some pretty blurry result, with very sharp source images.

[akin.kilic.]

Could we have a tutorial for training on Colab please? I have colab pro+ and I really want to utilize all my compute points to train my images for fun. Also the same tutorial could show how to use SDXL on Colab, how to install lora's to it, or try new models etc.

Thank you for all the work sir!

[antoniogutierrez]

in runpod webui 10 there is sdxl installed already. Models alre also loaded just get Vae and clone kohya. I also tested with RTX A6000 48gb which was very smooth and does the same 5200 in about 1 hour also you can generate easily while training with 48gb maybe better since 1 hour less approx. costs the same per/hour

[ArthursArts]

Your tutorials are very insightful. Thanks a lot!
By the way.. my jupyter notebook stays stuck at installing package: tensorrt.
Is this a common issue?

[krazyyworld]

how can i upload large files to runpod...trying to upload my regulization files but all methods seem to fail

[hceely]

So repeats stay the same? if i have 21 images still use 25? for training a person, I want flexibility not just likeness

[patriciaanaka]

I had pretty good success the first time I tried this. Today I tried with another set of images and it is a disaster. the images mostly don't look like the person, there's little specks and streaks all over the image sometimes, and mostly I get results that look nothing like the person. I used 19 images instead of 13 - would that make a huge difference? I thought it would give better results but apparently not... the first image it shows, as the image is developing, looks pretty good - then it updates as the progress gets farther and it applies a bunch of patterns and streaks an stuff to the image, ruining it. do you know what the issue could be?

[krazyyworld]

i installed after detailer but for some reason its not showing up on the ui but its appearing on the list of extensions installed. kindly help me out please

[chiragforu]

do you have a config file for the patreon users? so everything is setup for training images?

[KudaBear-gjbr]

What learning rate for larger datasets, 4-500 images?

[wisdon]

I have an important question: I want to buy a new pc but I cannot afford an high end machine: a self built pc like this one
ryzen 5 3600
motherboard gigabyte b550
rtx 3060 12gb
64gb RAM
700watt powe supply
1terabyte m2 storage memory (7000read 4000write) would be enough to train SD 1.5 models? or even SDXL ones? time isn't so important... and final question: tge the amount of system RAM how much is important? it affects speed or is irrelevant compared to videocard VRAM?

[MinuteScalping]

do you use the command "train unet only" for lora?

[joyeroye]

I lost internet connection during lora sdxl training.. on runpod.. how can i continue training? on runpod I have already explained 12 hours on it.

[chiragforu]

Hello do you have a template for run pod for SDXL LORA?