Version names includes the base model it was trained from, use appropriate version for each model. Rouwei version wont work well with noob, the opposite is also true. Old noob versions would be too weak on a fresh noob checkpoints also, eps and vpred should also be used with their dedicated versions. Don't forget to check version descriptions for more specific info and list of supported style tags.
Anima v4 info: Update for recent anima preview 3 version. List of styles is same.
Anima v3.5 info: Same version as anima-p2_v3, but trained of full range of blocks. Should™ just work better overall. Out of dataset test still fine. List of styles the same, grid
Anima v3 info: V3 is version trained from and for anima preview 0.2. Still won't affect data outside of dataset so much. Style list is same with v2, grid with styles.
Anima v2 info: A bit late, since 0.2 preview is already out, but whatever. It was trained from 0.1 version of preview model. Catastrophic forgetting was addressed and improved, at least it's not so disastrous now. Mixing between styles was also a little bit improved from empirical tests. List of all styles and grid. It would not work as well with 0.2 preview.
Anima v1 info: Anima version of lora comes with limited dataset. List of styles that comes with it. More like a test version. Model is in early stage and too incoherent at something above 1mp. If you decide to upscale, you would need to use tiled methods, unless you're okay with artifacts. Style tags should be invoked with @ before them. Mixing somewhat works, but not how it was with default xl weighting, rather like compel mode.
MIO - mix in one versions with mix of ai styles and some real artists in dataset.
BIO - best in one versions, containing only artificial images in dataset. 90% filled with naiv3 or styles from local models with very distinctive look only.
Description
Version trained from Noob-vpred-1.0 checkpoint. Dataset was massively updated from previous 1.337 version, most of it still naiv3 or distinctive local xl models styles. List of all styles in this version and hello Asuka grid with styles
FAQ
Comments (24)
List of all styles in this version and hello Asuka grid with styles << these links are broken fyi
Works on my end, it's just catbox, check connectivity to it
@bakariso tried vpn but failed. if you don't mind can you upload both of link elsewhere ?
@lowduty Yeah here https://pomf2.lain.la/f/h2acj2ek.txt https://pomf2.lain.la/f/7t1j10a.png
@bakariso cap
Hi, I'm trying to train my own style lora, and I find your style is one of the best lora I've used. Would you mind sharing the settings and parameters you used when training the lora?
I don't know if it will be really helpful, because multi concept lora params should be different, but sure https://litter.catbox.moe/d5f80vut5slrci5z.toml
@bakariso thank you so much!!!
Hi,really like ur Lora,very impressive.And i want know if i use this lora with the lora i trained(stylized adjust),which one should i set the highest weights,thx
do I have to use the "by" prefix in rouwei for the artists in your lora? Intuitively I'd say no because you didn't probably train with it but maybe the specific training of Rouwei still requires it?
No, "by" is not required for rouwei versions, was trained without it
@bakariso stupid question but what is edm2?
@deitychaser Small network that is trained along and determines timestep weights
@bakariso would consider training a new noob eps version with this network?
@deitychaser Why? I don't really like epsilon model version
@bakariso well for those who do like it of course :D
@deitychaser I can train one for 0.5 epsilon, it was latest stable version afaik
@bakariso did you have stability issues with v5 of your lora? Cuz you trained it on noobai eps 1.1
@deitychaser Yes, it was failed dataset and base model itself are too fried
@bakariso an easy workaround for noob compability is to train on IL 0.1 . LoRas trained on IL 0.1 work just as if trained on noob eps. But noob 0.5 is probably just fine too.
@deitychaser It's not always gonna be better than training from model you are using or closer one, especially with vpreds
@bakariso thank you for the updated eps lora!!