Compiling zed makes my PC completely freeze. I have to build it with fewer threads lmao
good question, I'm not actually sure?!?
Honestly, Llama works pretty well for this as a self-hostable solution, but not being able to attach the image as context kinda sucks. ik some people made vision model things for llama, I wonder if they're any good
I CAUGHT THE OTHER SIDE ANF VIDEO (recorded next to two tourists who looked genuinely scared for their life being near it)
When I used to scanlate, at the start of each chapter, I'd have something along the lines or "For the best quality, read this on <site>". Didn't accuse any aggregators of anything tho :p was just bc some aggregators compress to shit
anyone wanna donate a gpt-4o api key i can play with for a bit? Trying to make a auto-translator for manga, and am exploring the best translation options that preserve context. DeepL is underperforming here :p
has science gone too far?
horrors