Train a style Lora for Stable Diffusion - SD and SDXL - in 2 min with Civitai

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
[Music] welcome to this new mini guide today we'll learn how to prepare Aura for stable diffusion any version including sdxl in just a few minutes thanks to the civi site if you're not familiar civi is the largest reference Community for stable diffusion here you can find models luras images and much more it's a user-based community training a Laura isn't free but compared to other Solutions it's quite affordable and will cost you 500 bus because is the unit of currency in civii and you can acquire it after signing up either by purchasing with real money or by participating in the community after clicking on create Das trainer Laura select style and then give your Laura a name click next and as always you should have prepared your data set of images and captions it's important to describe each image well in the text file 50 images is a good number for sdxl use 1024x 1024 for SD 1.5 use 512x 512 12 zip everything and drag the zip file onto the website once you've uploaded the zip you can view all your images and texts if you wish you can modify the captions if you hadn't added them this is your chance to do so check the box regarding the rights of the images used and then click on next we're almost there let's wait for the images to upload to the server and then we're on to the final step at this point all that's left to do is choose the version of stable diffusion we want to train for in my case I'll select sdxl it's also possible to change the training parameters but do this only if you're certain about the data you're changing then press submit and you're all set right after we'll see our Laura in a submitted state after pressing submit a message appears indicating that 500 buz will be deducted from our account click on confirm to proceed and there you have it our Laura has been submitted we just need to wait a bit for it to be ready during the training we will see the various steps remaining until completion when the work is finished its status will be ready clicking on the name of the Laura will open the next page on this page you can download the Laura to try it locally or you can publish it on civii publishing is always possible even at a later time after trying it locally don't worry if the example images aren't very beautiful because the Laura will still do its job excellently [Music] in my case I download the Laura and test it on my PC with auto 1111 we do a quick test and there you have it a beautiful image that perfectly represents the train [Music] style remember that after testing it if you like it you can can share your Laura with the community by using the publish button thank you for watching don't forget to like And subscribe
Info
Channel: Digital Art Guide
Views: 1,950
Rating: undefined out of 5
Keywords: Stable Diffusion, Civitai, Lora training, SDXL, image AI, AI models, custom AI, dataset preparation, Buzz currency, model publishing, AI community, Auto1111, machine learning, AI tutorial, technology guide
Id: DUWY9G9AwwE
Channel Id: undefined
Length: 4min 31sec (271 seconds)
Published: Sat Feb 17 2024
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.