LORA: Install Guide and Super-High Quality Training - with Community Models!!!

The Easy Starter Guide to Installing LORA on Automatic 1111 for Stable Diffusion. Follow my super easy Lora setup guide and learn how to train your Lora files for super-high quality portraits. Use Realistic Vision V1.3 as the base model for extremely detailed and realistic results. Get better portraits with Lora, the super fast training tool that creates smaller files than Dreambooth. With this Stable Diffusion 1.5 model and Lora settings you get perfect results. All in my comprehensive Lora tutorial. Start enhancing your portraits today with the ultimate starter guide for Lora on Automatic 1111.

LORA Install Guide, Stable Diffusion, for Automatic 1111. Setup for Lora - Super Easy Guide. How to train your Lora Files. Get Super-High Quality Portraits by Training Lora with Realistic Vision V1.3 as the Base Model. Extremely Detailed and Realistic Results. Get better portraits with Lora. This Lora Tutorials shows you the model and Lora Settings to get the perfect results. Lora is a super fast training for Stable Diffusion. Lora creates smaller files than Dreambooth. Get better Results with Lora. The Ultimate Starter Guide for Lora on Automatic 1111

#### Links from the Video ####
And yes, I have also watched the Aitrepreneur Video on this. ;)

Support my Channel:

[hybridroid]

Are we not gon talk bout the intro? bars ! 😂

Whoa! 🤩 Thanks for the walk through!

This looks neat and can be a powerful tool to add to my D&D player and NPC builds!
I will give this a shot as soon as I finish my PC build.

[ahsenia]

Literally everything has changed. A new tutorial would be nice. Thank you!

[Mobay]

I would suggest everyone not to change the Powershell execution policy. It is there to prevent malicious code from running on you computer. There are no steps in this repo, that actually needs it. You can simply run the 11 commands one at a time in the prompt, without changing the execution policy. If you do change the policy, change it back to the system default to continue to have a secure system.

[michail_]

Thank you very much, Oliver! With you I figured out what to do and how to do it. In principle, it's not that difficult, but in order not to sit down to read a ton of material, it's easier to watch your video. And with time and experimentation, I'll be able to figure out what functions in "Lora" are needed and how best to adjust the model.

[DM-dyvn]

No need to name the image folder "100 underscore test", if there is only one folder.
Just "100" is enough telling LoRA to run 100 epochs (every picture 100 times).
The other two folders will be created automatically. No need to create them beforehand.
Just provide some path to somewhere of your choice.

You will need to name the image folders with underscore IF you want to train several concepts. Then you name the folders like 100_Concept1, 100_Concept2 etc.

[buenosaires]

You have explained the structure well, thanks Olivio! Expecting the updated version of Lora.

[dokwmrq]

Hey Olivio!Thanks for your guide! I am a foreigners and not good at English, but your video is very clearly and only watch once to understand! Thank you!

[ewersoncpimenta]

I'm in love with this content and they way you took your time to explain it.

Very well done, Olivio. Thank you.

[AG_before]

I have liked this because of the hard work you have put into it and the wonderful tutorial. 👍

But my brain exploded about 5 minutes into it. I need to go _way_ back!

[gratteur]

All your videos are really great. Thank you !

[Sheevlord]

Thanks for making the video!

Now I understand how people manage to release new LoRAs every day - the training takes less time than I expected. Gotta try it out myself soon.

[DM-dyvn]

Thank you, Olivio, for the explanations!
This manual setup of the folder structure scared me off, and I gave up.
I'll give a second try now =)

[jeffreyuthale]

Olivio !looking good ! Thanks for the tutorial!

[Beestlucas]

Thanks for this! You are a great teacher.

[DM-dyvn]

Hello Olivio, a higher batch size is not a free ride. It will speed up the process by averaging the image information in a batch. Thus, effectively providing less unique shots. Might work great when training a style on a big amount of images. I started with 34 images and increased the batch size like 1, 2, 4, 8. The model was getting visibly weaker and less precise at batch size 4 and higher. Well, 34 / 4 = 8 unique shots while 20 are needed

[GustavLonergan]

I could hear you explaning thing to me for days, thank you for the informative video!

[nrvwyqc]

Hi Olivio, thanks for your hard work in presenting this tutorial and give confidence in trying this software. Wish you success!

[gzus]

I will try the adventure from scratch in here, thanks you for you're hudge works, you won a student! xD ^^

[NeonSparks]

any chance of an updated version of this? it seems so much has changed since you made it :/