Type | |
Stats | 764 255 |
Reviews | (123) |
Published | Apr 19, 2023 |
Base Model | |
Trigger Words | TMNTRaphael red bandana mask with eye-holes holding sai dagger weapons |
Hash | AutoV2 C13C8AA9D4 |
Nothing I've come across so far could do convincing enough TMNT Turtles, so I took it upon myself to make a few models of my own. This is by no means perfect and it basically refuses to work on many models, probably due to the fact I trained on Experience 8.0. I''l be training on a comic/toon based model soon in hopes that I can reproduce the comic book style art as well.
Trained on about 100 HD of the 1990s style TMNT images for about 2000 steps
I recommend 512x512 or 768 and a weight of 0.8-0.9
High res fix is hit or miss, sometime it ruins it
20 - 30 steps seems to be the sweet spot, sometimes lower steps (around 20) produces better results
I recommend putting "toy" and "3D" in the negative prompt
Play around with the setting and weights and prompts and hopefully you can get something you like.
This is the initial alpha release and as such it is very much a trial and error process.
This was trained on a singular character and tested for such, so I'm not sure how it would handle groups.
Not trained for, or tested for NSFW, so be aware it may not work for NSFW.
The other Turtles are also being released as alphas at the time of writing this so be sure to check them out!
As always, if you have any tips for improving my training, please let me know! And please rate & comment & share your generations.