/vt/ - Virtual Youtubers

>>41664643
Nope. If webui is still running from whenever you mixed them, the console will show which models exactly you loaded, but it won't show the mixin parameters.
>>41665701
You shouldn't need any command line arguments for that other than "--no-half-vae" depending on your GPU. As long as SD VAE isn't set to "auto", it should be working. Look in your console output, right when you start up, towards the end, you should see an entry that looks like the second to last line here:
>Calculating sha256 for /content/stable-diffusion-webui/models/Stable-diffusion/HoloCreamsicle.safetensors: f01270c5a7d0d638046642dffa5e2f98a5928f156ee94804f8c9eaec1e168532
>Loading weights [f01270c5a7] from /content/stable-diffusion-webui/models/Stable-diffusion/HoloCreamsicle.safetensors
>Loading VAE weights specified in settings: /content/stable-diffusion-webui/models/VAE/nai.vae.pt
>Applying cross attention optimization (Doggettx).
Note the "Loading VAE weights..." - the paths/model names/vae names might differ, obviously. If you do see that and they still look faded, that's just Creamsicle smearing Vaseline over everything. As it does.
>>41666494
I mean, "vibrant and colorful" >>41659703 example wasn't with Creamsicle, nor was it with the NAI vae. I used the SD 1.5 MSE VAE, https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.ckpt - not sure if you need to rename .ckpt to .vae.pt - but I do. There's also the WD anime2/Berrymix VAE: https://huggingface.co/hakurei/waifu-diffusion-v1-4/resolve/main/vae/kl-f8-anime2.ckpt
Both of them are "stronger" than the NAI one. And they can sometimes deepfry images on model mixes that have more NAI/Anything in them.

>>41668392
oh, and there are easter eggs if you don't read instructions. picrel.