😕LoRA vs Dreambooth vs Textual Inversion vs Hypernetworks
Vložit
- čas přidán 14. 01. 2023
- There are 5 methods for teaching specific concepts, objects of styles to your Stable Diffusion: Textual Inversion, Dreambooth, Hypernetworks, LoRA and Aesthetic Gradients. The question is: which one should you use?
In this video we review 3 key research papers, look at the underlying mathematical mechanics behind each method, analyze data from civitai to arrive at an informed and final conclusion.
Discord: / discord
Live Stream in 8 hours: • 😕LoRA vs Dreambooth vs...
======= Links =======
Spreadsheet: docs.google.com/spreadsheets/...
LoRA paper: arxiv.org/abs/2106.09685
Dreambooth Paper: arxiv.org/abs/2208.12242
Textual Inversion Paper: arxiv.org/abs/2208.01618
Dreaming Tulpa: / dreamingtulpa
Driving a machine insane with Dreambooth: • I drove a Machine Insane
Good Tutorials:
Dreambooth tutorial by OlivioSarikas: • DreamBooth for Automat...
Hypernetworks tutorial by Aitrepreneur: • HYPERNETWORK: Train St...
Textual Inversion tutorial by Aitrepreneur: • ULTIMATE FREE TEXTUAL ...
Textual Inversion Paper Walkthough by me: • Textual Inversion with...
LoRA tutorial by me: • 7GB RAM Dreambooth wit...
LoRA tutorial by Nerdy Rodent: • LORA for Stable Diffus...
Aesthetic Embedings tutorial: • How to use Aesthetic G...
======= Music =======
From CZcams Audio Library:
Escapism Yung Logos
Music from freetousemusic.com
‘Late Morning’ by ‘LuKremBo’: • (no copyright music) c...
‘Marshmallow’ by ‘LuKremBo’: • lukrembo - marshmallow...
‘Rose’ by ‘LuKremBo’: • lukrembo - rose (royal...
‘Snow’ by LuKremBo: • lukrembo - snow (royal...
‘Sunset’ by ‘LuKremBo’: • (no copyright music) j...
‘Travel’ by ‘LuKremBo’: • lukrembo - travel (roy...
‘Branch’ by ‘LuKremBo’: • (no copyright music) c...
#stablediffusion #aiart #ai #machinelearning #dreambooth #textual-inversion #hypernetworks #lora #aesthetic-gradients #tutorials #resarch #aesthetic-embeddings - Věda a technologie
The thing about textual inversions is that they create embeddings that are cross combatable with the base models. A textual inversion trained with SD 1.5 will work with all 1.5 based models, and here is the kicker, you can combine them without having to do any model merging. That is HUGE.
yeah, the flexibility of textual inversion is a big factor, also it's really cool conceptually!!
The video really should have mentioned this, it's an incredible advantage for embeddings that was just left out.
Yes, combining two, three or more Dreambooth models is possible, but it takes time and generates yet another 2GB+ model that you need to save somewhere.
Whilst textual inversions can be used flexibly within the prompts in any combination, including weighting them, using as negative prompts, all on the fly with no extra file management
However, textual inversion cannot learn to output things that the base model is not able to do at all. So depending on the base model, it may not be possible to train a textual inversion for a specific concept.
@@expodemita I do not think they are compatible between 1.4/1.5 and 2.0 2.1. 2.0 and 2.1 should be compatible.
@@infocyde2024 2.0 and 2.1 are for sure
Much appreciated, having someone clever distil all of this dense information down and explain it succinctly and with so much enthusiasm is so refreshing!
Thanks for explaining these so well, your visual diagrams are great!
Thanks for your hard work putting this together, very helpful to evolve my understanding of the different approaches. Much appreciated!
Thanks so much for this. Very underrated channel, literally was thinking something like this would be really helpful.
Such an EXCELLENT video. Very very well researched and perfectly presented. Thanks for sharing all your findings and appreciate the time it took.
I love it, love your promises on what we are going to get from your video at the very starting few seconds of the video itself, keep it going man, love your channel and your energy.
This was a very informative video in fact, thank you! And I like your very dramatic delivery of the content! :)
Textual inversion is far better on 2.1 than 1.5, and i think that's why they don't get the same love dreambooth receives. You can also speed up textual inversion training if you spend a few minutes getting the initializing text right so the vectors start in relatively close proximity to their final resting place. The best part imo, is you can combine many embeddings together, something which dreamtbooth doesn't really allow.
How can you get the initializing text right before the training?
@@leonardom862 By running image to text I suppose ?
Ironically I haven't been able to run dreambooth yet,I switched to linux for AI... something broken with PyTorch2.0 and Cuda11.7 only thing affected is dreambooth training. Turn on gradient checkpoint and it cant train, turn it off and I cant make it to the first epoch without running out of 24GB of vram? I hope this gets fixed soon.
Totally agree. Suuper crucial to be able to call multiple embeddings in a prompt!
Superb video. I don't think I've ever seen a tutorial/explaination for anything that is this good.
Awesome comparison mate, great addition with the statistics, thanks a lot
you deserve more subscribers, only channel I found that actually delivers what you need to know
I greatly appreciate this video sir! It is really helpful for me to have context of how things actually work behind the scenes to make mental connections and improve how I interact with the external program.
Liked and Subscribed, Thank you for all the hard work!
This video is exactly what I needed, and you went about it in the best way possible. Thanks for this
You explained that amazingly, very easy to understand - also things move fast because it seems like LoRA is now the most popular.
Thanks a ton for this breakdown, I've been struggling with this same question for a few weeks now. I had already come to a similar conclusion myself, but this was very validating.
Dreambooth is preferred, but the models sizes make it so cumbersome and challenging to test different versions. With textual inversion, the file sizes are insignificant, and you can stack them on top of each other, making them very flexible.
I haven't actually evaluated embeddedings (textual inversion) yet for quality because the animation notebook I use doesn't support them, but the developer just made it compatible, so I'm looking forward to testing it out more.
Wow. Thanks for this video, esp. the first part which gave just enough detail to understand the trade-offs and underlying approaches.
i'm a total beginner to AI, and i suck at math, but you somehow managed to clear a shit ton of confusion. I was hooked on Dreambooth tutorials and trust me, you don't want that. I literally thought i was not going to be able to get started simply because of the massive resources it required.
Trust me, you are really good at explaning things :)
Really appreaciate the help
I’ve been trying to install dream booth for 3 days now. No success. Ready to walk in front of a bus
Great explanation! I've learned more about how AI art works from this video alone than all my previous watched videos combined. Everyone tends to say how to configure things without explaining how it works.
Really appreciate this, so many videos showing how to do this stuff, but not how it works, and specially not how it works dumbed down to a level I can understand. Very cool, thankyou
One interesting piece of data is Lora has quite a high faves per download rating while only being out for a short period of time
yeah, I saw that too.... good sign!
What a great video! Thanks for your academic sharing and empirical results❤
Outstanding job explaining these concepts! Well done!
That was a really helpful video that definitely saved me a bunch of time trying to understand these differences by myself :P
saving people time makes me super happy, thanks!
You are really good at explaining this stuff. Thanks!
Thanks for making such great and informative video. Keep up the good work
That is the type of quality content that I'm digging for.
I want to understand Stable Diffusion and everything related. But my attention span/knowledge about programming is not enough that I can just read papers about it. So I need videos, with visuals, and easy explainations. And your video was Perfect. Liked + Subscribed :)
This is great. Thank you for sharing your knowledge, and about Excalidraw.
Thank You a lot. This has been a really good explanation that I felt missing.
This video is dope. Super clear and informative. Thank you!!!
I love your videos, always on point !
really great video! finally understand the differences. just the conclusion is already out of date, since we're moving so incredibly fast. lora, is the most popular format on civitai now. understandable, since training is the quickest, even though ti's end-result is much smaller.
Loved this breakdown. You need more followers!
Hey Koiboi. Great video. When you made this video, as you said yourself, LoRA was still very new and the stats are probably not accurate. Now that a good amount of time has passed, I would love to watch an updated analysis video on the effectiveness of LoRA compared to Dreambooth and Textual Inversion.
Either way, this is the most informative video I've watched so far comparing these fine-tuning models. Liked and subbed 👍.
this is the best video about the topic ive ever seen, thanks so much
Best video I ever seen.
Best vibes!
Thanks so much
What I was looking for in that video was the comparison of usability in different scenarios. Which model is good for faces which one for style transfer etc. I'm missing that, other than that quite comprehensive comparison. Good job!
Great info! And coincides with what I learned on Computerphile's channel. Slowly but surely my mind is able to wrap around with what we're dealing with.
thanks for making those complex concepts easy to understand!
Thank you so much for this video! Amazing work
LORA works only with an extension, and many people don't know how to use it yet, hence lower ratings. Great video btw! Visual comparision would have been great as well! As far as I can remember, there was one in LORA blogpost, showing how textual inversion may be less flexible than dreambooth or lora, and the latter two were showing comparatively similar results.
Auto added compatibility now! But it was only added recently. (I still use the extension, I find the drop-down much easier to use than how auto implemented it, plus it gives you the ability to tweak the weight of both U-Net and the Text Encoder -- super cool!)
i did everything i should do and i never get lora to run. it was no issue with the textual inversion, tho.
controlnet has almost made lora obsolete for anything other than oddities
Thanks koiboi! I absolutely love these in depth videos.
Any plans to give ControlNet this sort of treatment?
This video is AMAZING! Thank you SO MUCH.
Really great video, thanks for sharing all your research!
glad it helped!
That was an insanely good explanation. Thank you!
Thanks for the input, good research!!
THANK YOU THANK YOU THANK YOU!! Great video. Great insight.
thanks for the short explanation. Loved it!
Absolutely incredible video, thank you!
Very nice summary- thank you 🙏
All different now! LoRA is by far the best all round method now and hugely gaining popularity... Great video by the way, excellent explanations!
Great video and explanation! I really want TI to be the future but I agree, the quality of dreambooth training is usually better.
thank's for the data point!
Thank you for model explanation. Really loved your content so far.
At the end of civitai comparison, I’m curious if we split data to use cases, object embedding vs style embedding would have different performance/preference.
that's a super hard question to answer :(
brilliant video, thank you for all the efforts!
I just started with the whole txt2img models and therefore I really thank you for this video!
It was great to watch and I got a lot of information. I really appreciate it.
Also, would you mind sharing the excalidraw board? Either as a exported png or as a file, it would be a great resource for my own documentation, if it's not too much to ask.
Incredable explanation! Thanks a lot.
Thank you for an informative and engaging video!
This is so well taught man thank you so much
I've been messing with Loras and they seem to work really well. You can also do a good amount of mix and matching with loras whereas a full model checkpoint only allows you to use that one model at a time. if I had a fruits lora and a vegetables lora, then I could just turn them both on to get fruits and vegies in my random prompt that doesn't ask for fruits or vegies. If I later just want fruit then I could just remove the vegies lora.
I think loras are going to be big going forward, most people just don't know about them yet.
I get the feeling that Textual Inversion is the go-to for when you have a new idea you want to teach the model (like a specific character or subject), and Lora is great for when you have a concept you don't want to stop and explain to the model, or may have difficulty doing so. They're very similar things, but not quite the same.
For example; loras are great for mimicking a specific art style, because instead of having to describe "I want a painted animation style like this specific style, but with eyes drawn just so", you can train a lora and then just say "" at the end of your prompt, and since it isn't actually part of the prompt, this clears up tokens for describing the actual thing you want depicted in that style.
This is exactly how i feel about lora. It’s disappointing that people don’t seem to gasp the same values of how beneficial loras can be.
@@treyslider6954 And also, as described in the Automatic1111 docs, Textual Inversion can't teach COMPLETELY new concepts.
The example they gave is that if you trained a model that only knew how to make apples on images of bananas, it wouldn't learn what a banana is, it would just make long yellow apples (in the best-case scenario). Because it's not actually changing model weights, it's better for teaching a style than a new subject, because unless the subject is very similar to something it's seen, it can't learn it.
LoRAs can teach a model something it's never seen before, because they are directly inserting weights into the model, meaning it's actually modifying the model and not the input going into it.
Basically, Textual Inversion for simple styles, LoRA for anything complicated.
You're an absolute legend. Great video
Very clear explanation, thank you!
What an insanely helpful video! I'm still holding out hope the quality of hypernetworks improves (I've had fantastic results with it, but updates often break it and nobody really knows what they're doing so guides are not great)
It shares some of the same advantages as TI (smaller file size, can be transferred between models easily) and I really hate having giant checkpoints just to add single concepts.
I was excited to learn about LoRA, but it looks like it can't be used without first adding it to a checkpoint, so its lost some appeal for me. Can you train multiple concepts to a checkpoint with LoRA one at a time and have them all retain coherency?
Much appreciated! Having this nodes of wisdom to operate with AI models is a huge contribution to society! Props to you.
You explained this so well. My smooth brain couldn't understand these different methods for the longest time \uwu/
Awesome explanation, thank you!
Nice video, even though I watched it 6 months later, lots of things happened, your video is still extremely helpful (except the LoRA part 😊)
Thank you so much for this video! 🙏
This video is incredibly helpful. I'm probably going to use either LoRA or Dreambooth, as Textual Inversion can't teach brand new subjects as well as you can by directly inserting or modifying weights in the model.
Very nice introduction! Thank you!
the fact that you dont have10x more subscribers or views boggles me. i guess not enough sex and drama. i hate to be cynical but holy S*t this is a important subject. and you break things down so normies like me can understand. Thank you sincerely.
thanks so much dude!! I made a few off-topic passion-project videos that my audience didn't really understand, so I think youtube doesn't trust my content... something like that.
Quality audience < quantity audience!
@@lewingtonn youtube is tyrannical these days. I guess with the amount of videos being posted everyday they have to do some thing. but one would think searching would solve the issue of relevance and quality, but "the algorithm" obviously chooses who it vets, and who it vets is obviously etc etc etc. We could have this conversation for hours maybe even days lmao. but no really thanks for your content man. seriously.
@@ArtfulRascal8 sounds like you probably know more about this than me lmao, but thanks honestly!
could you please do a part 2 where you
- explain aesthetic gradients for educational purposes, and maybe provide data on user feedback like you did at the end for the others.
- explain lycoris, which from what i understand is lora + 4 random good ideas, but id love to see someone on your level break it down a bit better.
- give us updated data on the other forms now that more feedback is available (you mentioned not having a big enough sample size to judge the newest tech).
that would be insanely helpful. thanks!
Great tutorial mate, thank you!
Amazing video with the explanation on the heuristics. Have you updated your excel with the usage now after 4 months and would you change the opinion based on your quantitative data from civitai
Great breakdown of the details, and pros and cons of each technology! Which model would you use if you plan to train a few characters, then re-use them a lot together to write a visual story frame by frame (as in colored manga)? On a low to mid end laptop GPU...
Why am I only now seeing this? Great video and thanks for the feature ❤
That's so much work! Thank you man
Great video. Big thanks
Great video, really helped me understand this
what a great list of checkpoints you have, a man of culture 🤣
clearly explained, much appreciated!
Very great job! Thank you!🥰
That was absolutely awesome. Thanks for that, I wish you'd do a deep dive tutorial of the most recent update of dreambooth in automatic 1111, some settings seems to have major impact in the training while not being very much covered. (Gradient anyone? XD)
Yeah, every time I grasp the settings a bit better, new settings get added.
Sir, thanks for your effort and detailed explanation 🫡🫡🫡learned so much🙇♂🙇♂🙇♂
Amazing work thank you. New sub.
Really good guide. One constructive critical point: when calculating for a trend of data: do it by time, not in total. Dreambooth was the first one, so you biased the numbers in favor, simply because Dreambooth existed for longer. For the favorites, you could do Favorites/Downloads.
Nicely done. I know from experience that training Dreambooth requires at least 10 GB of VRAM, so its not accessible to everyone. I am definitely going to check out the other methods.
Great Video! Thank you very much!
Thank you, sir, for this explanation!
wow that was great, thank you so much!
I really like to combine textual inversions with LoRAs to further control the result. Not always perfect but can get really good results.
ok you had me @00:27 , would be cool to see a video on civitai
thanks a lot for your effort, great job
Great video and it's explained very well. I feel smarter now.
Very interesting video, thanks.
Finetuning is using Dreambooth with multiple text/image embedding right ?
This is awesome! thank you!
I've used all these methods besides dreambooth. And from my experience on training a specific person LORA has gave me the best results and it's also the quickest of the methods i've tried as well which is a bonus. You can also use them on any model and mix them together ect. The only problem i've had is using it to produce two different people at the same time. As you can't go over a total weight of 1.0 but more realistically like 0.8 and the more you use together the lower the weight you have to use for each. But that can be solved using inpainting or probably other methods as well.
Awesome work, man!
Well-suited review for any of those fancy ways to generate best 2d girl for yourself
thank you for the great explanation!
I just started training Lora in the dreambooth extension in Automatic1111 and the files it's outputting for me for SD2.1 are about 5 MB pt files, nowhere near what you said Lora sizes should be.