LORA + Checkpoint Model Training GUIDE - Get the BEST RESULTS super easy

Great guide, but none of the tools will install. Tried following the many instructions but it literally never works. I always get an error as soon as I run the bat file for Kohya. So I've given up trying to train models. Even with your guide literally nothing works, just endless errors all over the place.

Thank for your tutorial, it really clear. I almost go to the end of Lora training, but the model output being .json file, not safetensor file. I try few time but still the same, did I doing wrong in a step. Please help!

can you train on AMD on windows not linux

So does every individual user have to train a model before they'll actually work?

My gpu isn't great unfortunately :(

There have been a lot of updates for this method, It would be awesome if you could do a new video with all the updates. Anyway, thanks!!

Can the same LORA be used also for videos with stablediffusion?

I'm disapointed, you anounce checkpoint training but you end up training a LoRa!!!

I just downloaded Kohya ss and there is no gui-user.bat in there anymore. When I use gui.bat like the webpage says, it gets it started, but there is nothing to click on to have it open it in a web browser. So, you can't get to the GUI through the method you describe. I figured a work around by using the setup.bat file because I remembered seeing an option to use #5 where is says you can start the gui and open a web browser.

Thank you for guiding me with this excellent video!

For some reason it doesn't seem to work. I can't get it to write the txt files with the key words. It says that it's done, but never actually does anything.

I was not able to train a lora with this tutorial, I have tried to ask for help on your discord server but did not get any up to now..

Does this also work well for objects: if I wanted to do product placement in images, can I successfully train a model on a specific item, such as a car, laptop, etc?

That tag manager is really cool, thanks for that.

Hello Olivio! What Laptop you were using in this video? I have RTX 3050 Ti and it only has 4gb vram, it is interesting that your laptop has 16gb vram.

Hi, just installed Kohya_ss version v21.8.10 and it looks very different fro the one in the video and I cannot find how to generate caption.
Have you got any tutorial on that new version? Thanks

when i press train i get CalledProcessError: Command '['J:\\kohya\\kohya_ss\\venv\\Scripts\\python.exe
ends with returned non-zero exit status 1.

this too: ValueError: torch.cuda.is_available() should be True but is False. xformers' memory efficient attention is only
available for GPU

confused this is new to me

I have been following the steps of a tutorial on how to use LORA and stable difussion, but your file is safetensors and mine is a Json, does anyone know why?

Your guide on LORA and Checkpoint Model Training is a game-changer. The merging trick you shared is pure gold. Subscribed! 🙌

I completed my first training of a LORA based on close to 300 images. When it's finished, I'm left with all my epochs being close to 9mb and the LORA seems to have no effect when tried in ComfyUI. Am I missing something incredibly obvious?

Kohya is a nightmare. Doesn't work at all. I don't recommend it

Can we merge a lora model made with the SDXL base 1.0 with something like realistic vision?

I dont know I followed the video and then my PC crashed when it ran the setup.bat file. I mean my PC flat out crashed, no response. I had to hard turn off my PC and turn it back on. It went through so many restarts while Windows was recovering. Thankfully no damage was done. But I wont be following this guide thats for sure.

with my rtx3050 8gb vram, it take 88hours to train (15 image x 20 train x 10 epoch ) / 1 train batch size = total 3000 steps
i'm using max resolution 768,768

even if i change the resolution to 516x516
it still take a longer time

if i use lora, to train same size steps it take 12 hours

is it mean i need to upgrade my VGA, or are there still a way to fast the train?
I really wanted to make my own check point, i saw on another YT, they train their own model using like google colab,
i never used that before, or any other cloud base, but can i download the checkpoint to use on my computer later after train in cloud?
Edit :
I change the train batch size to 4 and max resolution to 516x516, now my total steps are 750
and i still need time 23 hours to finish the training :")

I had a weird problem that took a day to figure out, it turned out that this json file was empty, I renamed the hugging face directory and it was re-downloaded.

Loading vocab file C:\Users\Andy/.cache\huggingface\hub\models--openai--clip-vit-large-patch14\snapshots\8d052a0f05efbaefbc9e8786ba291cfdf93e5bff\vocab.json

@oliviosarikas do you know how feasible it is to install Stable Diffusion on a Mac? for example Mac Mini M2 Pro? Will it work by running it on a virtual machine to windows or will it be better via bootcamp? or is there a Mac direct version?

Thanks

Of everything I learned, you blew my mind with the f2 trick

Hai thanks for the videos I'm new on the AI generated field, so what is the difference with Checkpoint and LoRa?

hello The discord invite didnt work please repost it

fckn bold you dont know nothing, you only make reviews for like , get a JOB AND A REAL LIFE

Good evening Olivio,
So I just followed your guide to create my first LORA but I find myself with 22 hours of training!
I must have made a mistake somewhere ;-) However, I have an RTX 4070 TI and a Core I9 13900k.
120 photos of the same model in high quality (FHD) half are faces and the other is the full length model.
I chose 10 steps, I think my concern comes from there. I'm still going to let it finish just out of curiosity.
Thank you for the guide which will have nevertheless been very useful to me.

By far the most thorough and easiest to understand guide. I followed someone else's this past week and they skipped.

The Cudda setup and explanation
The Folder Structuring setup explanation
The explanation of the different tabs, up until now I thought the Dreambooth was just a different way to train Lora's....anyway their guide was really confusing and missed a lot of stuff you cover.

lovely vid, keep on these good works!

you are awesome. internet is getting better bcs of u 🎉

❤

I like Olivia's videos. The only difficulty is that I watch all his videos but I can't train stable diffution using LoRA. Every time I try to train I get errors indicating "CUDA out of memory". I've been trying for a month but I couldn't. I want to sign up as a member but I don't know if it's worth it since I couldn't solve even this basic question here or on the Discor channel.

Has anyone here already solved "CUDA out of memory" problem?

I have 10GB RTX3080 and I still get CUDA out of memory. I do have 3050 and 3080 running in parallel but I am not sure how to configure Kohya to use both.

this is heavily outdated, I'm trying to install Kohya and there's so many things that ain't available or as described in the video. For example, the cudnn installation, its very different now. Also I don't have the upgrade.ps1

Great stuff, thanks

Your definition of Epoch means nothing. What do you mean "Generation of the model"? 👎

Please make one for online but with the same idea, like the idea of full resolution

@OlivioSarikas why dont you spend 30 seconds on extracting lora from your checkpiont merge? :)

great info

Thanks. This video helped a lot. However i do have aproblem with Kohya. All the buttons are not working so cannot really do anything. The is a warnig about LD Library in Terminbal. Am i missing some files or what is going on?

when captioning, whats the diff between BLIP and WD14? Why do you use wd14?

Fantastic tutorial Olivio! Love all your AI videos. Can you use LORA to train other objects and keep them consistent? Using your technique, I've been trying to create a LORA of a specific baseball glove and a specific helmet, and apply it to an image of a baseball player in stadium. Most of the outputs I get are quite monstrous like 10 finger alien gloves. I've given it a ton of data like close up/medium shots of the glove from multiple angles, medium and wide shot of someone wearing the glove. Still nothing works. Any thoughts on how to train a LORA to keep something like a glove consistent? I'm using using SD v1.5 pruned, photon v1, realistic vision v40 as the checkpoints in A1111.

yes plese make a video training checkpoint models using online servers

I dont whats wrong with your or whats wrong with my PC. It doenst want to work. I really hoped after coupled of days to get it working, nope. I think LORA doesnt want to work. PermissionError: [Errno 13] Permission denied: 'D:\\Kohya\\kohya_ss\\setup\\..\\cudnn_windows\\cudnn_windows' . everytime the same thing. I searched everywhere how to install it properly, but there nobody except olivio. Sry Olivio, I am not against you, I am just angry and that there is no other video who can explain it to me easier or from step to step from the beginning