Tokens, what goes on yer tabletops and D&D.
Built as a test.
Works the best with drawn checkpoints, I like the comicbabes one.
Forces a style and a topdown perspective upon your luckless subject. Use strength of the lora to control how loosey-goosey you want to get from the standard view.
Somethings just aren't prepared to be viewed this way, so your mileage may vary
Description
Initial release.
FAQ
Comments (12)
looks very useful for indiegames
Gotta try this, Shame hands look worse than usual. Limitations of diffusion I guess.
But as a DnD player this is something that I wish I had 2 years ago.
As a DM/GM, this is my holy grail. Thank you for pioneering this one
Can you tell me which checkpoint to use specifically? I can't reproduce these images
Help create a promt that does not have shadows and unnecessary parts protruding from the body
I'm having trouble using this model. Not able to really use it it seems. Both standalone and merging your model and comicbabes as the secondary. I am using your generation data too, so not sure what I'm doing wrong. But this model looks freaking amazing.
topdown view is the issue, it merging really bad.
suggesting you to use control net with depth (just build any pose in some online editor)
Can't get this model working, with the same model, using PNG Info to just pull the exact same prompts.
Nice work, the token's perspective is good every time.
However, I always have an error that spoils the token like an extra limb, deformed body, horrble face etc...
How can I adjust this? I've tested your prompts with the same parameters (especially the negative prompt).
I use this in Img2Img generation. And that's perfect! Army of tokens in a minute
would love to get an updated version of this. I trained a lora from foundry tokens but i dont love the outputs from my lora. still kinda new to curating datasets