Experimenting With OneTrainer - OneTrainer vs Kohya - Realism vs Stylization - Reg images vs No-Reg images
Added 2024-01-08 14:03:48 +0000 UTCExperimenting With OneTrainer - OneTrainer vs Kohya - Realism vs Stylization - Reg images vs No-Reg images
I feel like reg images using in OneTrainer improved realism and also lessened the overfitting as expected
OneTrainer realism looks better than Kohya meanwhile Kohya generalization looks better
Download attached images to see full resolution
OneTrainer Stable Diffusion XL (SDXL) Fine Tuning Best Presets
Massive 4K Resolution Woman & Man Class Ground Truth Stable Diffusion Regularization Images Dataset
The Very Best Workflow For SDXL DreamBooth / Full Fine Tuning - Results Of 100+ Full Trainings
Comments
Sadly not because it requires a Desktop GUI. I plan to research for such template on RunPod.
Furkan Gözükara
2024-01-23 14:12:23 +0000 UTCBig improvements to realism. Is there a runpod installer script for OneTrainer?
Faraad Haashim
2024-01-23 07:21:39 +0000 UTCyes they recently merged into code and I plan to do hopefully soon
Furkan Gözükara
2024-01-22 10:45:57 +0000 UTCyou can train on Pixart using onetrain, i just started the training not sure about the results
Hassan Alhassan
2024-01-22 08:56:51 +0000 UTCnot yet sadly.
Furkan Gözükara
2024-01-20 23:57:37 +0000 UTCDid you get a chance to test with the face masking feature?
mike oxmaul
2024-01-20 23:35:43 +0000 UTCi am starting machines now to start testing
Furkan Gözükara
2024-01-15 22:12:36 +0000 UTCno problem man, thanks for the hard work!
Idraen
2024-01-15 21:15:42 +0000 UTCstill working on that. so sorry that i still couldn't release
Furkan Gözükara
2024-01-13 22:20:55 +0000 UTCAny chance to get a SD 1.5 config?
Idraen
2024-01-13 21:16:08 +0000 UTCit is true for sd 1.5. for SDXL i dont know maybe due to incorrect implementation i got horrific results nothing good with EMA. did a lot of testing
Furkan Gözükara
2024-01-11 14:12:39 +0000 UTCWhat about training EMA? I recall you praising Kohya's 1.5 version w EMA training for being one that gave the best results. The configs here (which have worked very well for me) have the EMA training. I think that is default. Is that function disabled in these configs due to not (yet) being tested, or have you determined the best option is to leave it off (at least how it works at the moment in Open trainer)?
John Dopamine
2024-01-11 12:53:42 +0000 UTCi am yet to test masking
Furkan Gözükara
2024-01-11 11:05:18 +0000 UTCI would like to see comparison between masked vs usual training as well !!
Umut T.
2024-01-11 10:52:03 +0000 UTCye kohya is great too
Furkan Gözükara
2024-01-11 10:48:28 +0000 UTCI love my Kohya, no ty. If I wanna generalization, I just train on sdxl dpo... I believe style loras are better without generalization.
Umut T.
2024-01-11 10:47:18 +0000 UTCThanks for the instructions on reg images!! I need to do comparisons w/ masking on vs. off. I'm not sure if it just masks the person and doesn't look at the rest at all or if it just weights things differently. If it only looks inside the mask then yea, maybe not as useful as I was thinking. I got good enough results w/ it on though so I'll try it on w/ reg images sometime.
John Dopamine
2024-01-08 17:36:05 +0000 UTC1 : added our reg images as new concept but made the repeat 0.003. so 0.003 * 5200 = 15 images used each epoch 2 : i only have a txt file for training dataset that has ohwx man and for reg images just man. so a single txt file that contains the rare token and class token 3 : i didnt try masking yet. i don't think it will improve quality. do you have good results with it? i believe better dataset better than masking
Furkan Gözükara
2024-01-08 16:09:39 +0000 UTCDo you explain anywhere how you incorporated reg images? Did you just make a concept for "Regularisation images", put 3k images in there w/o captions (or maybe a single caption txt file w/ the class name? ie "Person"?) Also have you used captions or a single caption txt for your trials so far? I got really nice results w/ your configs already, but I only put a simple txt file w/ the subject (trigger) name in it - I never captioned w/ kohya for single likenesses. Lastly, per the developer in that thread you created on their github I tried mask traininng last night and had good results. I used the "remgb" option and it quickly made masks for just the person - pretty cool. Do you know in that case if it just looks at the person, or does it look at both but can differentiate better? I don't believe Kohya has a masking feature so I wanted to try it. I'm so glad you picked this to look into. I see others asking for different things, but for me this is the best topic you've taken on after Kohya's training. Plus the OT developer is so active that I'm sure he'll be very quick to add support for future models (OT supports PicsArt training, Wurscher training, etc - dev is all in w/ the different models).
John Dopamine
2024-01-08 14:51:07 +0000 UTC
