I lost my job so i have some extra time on my hands
I lost my job so i have some extra time on my hands
Stime to learn now i guess
Fi hope you can find asap
Sgoing to try and just do freelance
Swhat is the it/s supposed to be?
Fdepends on gpu
Fwhat is your gpu
Srtx 4090
Si did the 15gb fast preset
F1-1.2 range
Sokay that's what i'm getting was hoping i could get it faster haha
Fyouy need to sacrifice quality with xformers
For batch size + xformers
Soh i can use xformers? i think i need to rewatch and look for that part
Fyes but that can reduce quality
Fxformers + turn off gradient checkpoint
Fwill speed up
Sgradient is under learning rate scaler?
Shmm xformers only took it up to 1.4
Fye you cant get much faster
Fdid you disable gradient checkpoint?
Si did
Fok then that is what you can get
Ftry batch size 2
Fbut it may give out of vram error likely
HI'm currently trying to learn how to train a model like realvisXL (https://civitai.com/models/139562/realvisxl-v40).
I've been training LoRA models, but have no experience with checkpoint training.
To train a target-focused (realism) and also flexible model like realvisXL, do I just need a lot more images followed by detailed captions? Or is there something more to it?
If you have any suggestions or links to share, I'd be really grateful.
I've been training LoRA models, but have no experience with checkpoint training.
To train a target-focused (realism) and also flexible model like realvisXL, do I just need a lot more images followed by detailed captions? Or is there something more to it?
If you have any suggestions or links to share, I'd be really grateful.
Fyou need a big dataset
Fproper captioning
Fand training
HHow big to be exact?
Do I need 100+ image? or 1000+?
And for the captions, is it ok to use captioning tools such as blip or wd1.4? It seems impossible to manually caption thousands of images. Thanks.
Do I need 100+ image? or 1000+?
And for the captions, is it ok to use captioning tools such as blip or wd1.4? It seems impossible to manually caption thousands of images. Thanks.
Fyou want to do a model like real vis XL 4?
Fyou need 10s of thousands high quality image to fine to SDXL base into such model
Fyou can use CogVLM for captioning
Fespecially CogVLM v2
HI'm wondering if the Dreambooth method would be effective for training with 10,000 high-quality images. Based on the publicly available SDXL base model, there are various fine-tuning methods such as Dreambooth, LoRA, HyperNetwork, and Textual Inversion. From what I know, Dreambooth seems to be the most appropriate. I would appreciate it if you could share any specifically recommended methods for the final training process you emphasized, or if there are any noteworthy precedents for training with such a large dataset.
(P.S. I've watched all of your fine-tuning tutorial videos on YouTube multiple times. I'm a big fan! Thank you!)
(P.S. I've watched all of your fine-tuning tutorial videos on YouTube multiple times. I'm a big fan! Thank you!)
Fdreambooth without class images = fine tuning
Fit is best method
Hby without class images you mean without regularization images right?
Fclass = reg same meaning
HOkay thanks a lot! Lastly, I'm curious if it would be appropriate to use a dynamic optimizer in this case as well. I've consistently used the Prodigy optimizer and cosine LR scheduler with LoRA, and I'm wondering if they would be effective when training with a large number of images.
Of course, frequent experimentation should be prioritized, but it would be greatly helpful if there's any guidance on where to set the starting point. Thank you!
Of course, frequent experimentation should be prioritized, but it would be greatly helpful if there's any guidance on where to set the starting point. Thank you!
Fi tested them and i found adafactor works better
Fbut it is a choice
Fbefore doing a big training
Fmake smaller ones and see if works
HThanks bunch. I'll try smaller version first and share the results if it turns out to be ok.
SDr Furkan iS POSIBLE TO USE STABLE DIFUSION 3 for training in one trainer?
Fi think yes
Fbut i havent had chance yet
Fi plan hopefully soon
STHank you you are very kind