Animation, creating the background separately and animate it with svd xt, then creating a starting frame of character and animate it with animate anyone or animatediff while use of layer diffusion to create the "sprites" animation and insert it manually in DaVinci resolve, or inserting them with inpaint controlnet.....maybe a lot of work, and bugs and tears but all are the steps for ai movies
I am just waiting for a completely Layer Based workflow now. If this is possible and inpaint and outpaint are already present. nothing is stopping someone from making a full fledged layer based UI like Photoshop. Image composition has been made so much easier by this.
I had been trying for hours to get this to work. After manually downloading all the models and trying different things I finally shut down Forge and tried the "Update" batch file. That did the trick! I'm posting this in case anyone else have had trouble getting this to work for them and haven't tried launching the Update.bat file.
woah super amazing, until now I just had been using some chroma green key for the backs and maybe some online background removal tool, but this is just what i needed :)
Yeah that pull request has been sitting for a week now. I wound up just checking out the PR commit, works well for those of us that like to generate forever overnight and see what generates.
Nice plugin but a pain to upscale, in forge hiresfix doesn't work if you try to use it after a lowres generation and if you tick the hiresfix box and upscale all at one time, it messes with the transparency, utlimate SD upscale in img2img remove the transparency too
On the one hand this is really nice, on the other the images dont really fit together as well as if it was generated in one go and its another 6+ gb in models.
I love how if I stop following advancements with stable diffusion for a few days I miss out on multiple new features. I wonder when Adobe are going to join the game 🤔
Possible to get an install guide for Comfyui's Insightface? I'm trying to do the Face ID in the Krita plugin & it seems very complicated for a Custom Comfyui (non-portable) version. 🐭