/vt/ - Virtual Youtubers

>>46633526
kek, delightful. I'd mess with it this evening to see if I can get it in there, but BBW anon should make a hugginface repo or something, I can't easily download MEGA models in the setup cell without forcing it to install a bunch of currently optional limited to the MEGA cell dependencies

you just want to prompt {{{bigger}}} on the Colab, don't you anon?

>>46633520
nyooooooooooooom
you think I downloaded all those hll4-p3 merges for grids one by one? or even fed them into the terminal or in a shell script? hell no. I just let this thing do it. the "Errors" are because it's a localtunnel tab to a dead Colab, but TL;DR
1. Optionally click Stretch UI to "fix" it the same way I make "wide" checkpoint merger screenshots
2. Click "Turn On Logging"
3. Pick aria2c, it's fastest for downloading even if it does have to install aria2c - alternatively use wget
4. Input a fuckton of links formatted the way the transparent example is, so e.g.
#model
somehuggingfacelinks
somecivitailinks
#VAE
some VAE links
etc.
5. Let it rip

I think it's capable of parsing links, it's been a while since I read through "Readme Page, Example, Syntax, Valid Hashtags, Here's how you can get the direct links" but I only ever feed it a block of huggingface.co/.../resolve/...safetensors links. Works great. I start it up, prompt an image or two while I wait, then it makes a loud PIN PON when it finishes. "Turn Logging Off" since it captures a bunch of extra stuff to the console output. Click the little model refresh icon in the top left. Ta-da. Everyone's here.