ififfy

30M. About to drop the hottest scat captions of 2024.

[ififfy @ififfy1]

@PureScat I'd probably die of happiness

[ififfy @ififfy1]

@toilet_doll @RubberUrinal Corn starch!

[ififfy @ififfy1]

Made this a while ago! #scat #gooning #shiteating

[ififfy @ififfy1]

I should finally post those scat captions I've been promising, huh?

#scat #gooning

[ififfy @ififfy1]

@maxxietrap2 This is hot but also you should know water intoxication is a thing, someone died after a radio show contest once

https://t.co/WoJwinf9ei

[ififfy @ififfy1]

@s0phiasprinkle I highly recommend taking to a therapist about this stuff. Totally changed my perspective.

[ififfy @ififfy1]

@s0phiasprinkle this is the high-quality aerospace content I came here for

[ififfy @ififfy1]

@s0phiasprinkle Sick toot breh

[ififfy @ififfy1]

@s0phiasprinkle Yo

[ififfy @ififfy1]

@gooncultist They have you install some Python stuff, but it's only to covert models from one form to another. Once the 4-bit quantized formats are distributed widely, it'll just be (1) download program, (2) move model to folder, (3) write prompt, (4) goon until death

[ififfy @ififfy1]

@gooncultist The key is that llama.cpp is a C++ implementation that's distributed as a binary, so you don't need a bunch of pytorch bullshit. Macs have integrated GPUs these days, so there's no separate driver, and LLM folks seem to also be Mac folks, so support is there by default.

[ififfy @ififfy1]

@femmesimp sorry, too broken