Recent content by Kimonos

If your talking about the search feature, it seems to just use google. Googling the same query gives the same sources (minus sponsored results) that are seen in OpenAI's search sources tab. I see alot of reddit threads in there. Maybe try using various "search" prompts in ChatGPT related to your...

Because of its niche or because of filtering Blackhat specifically does?

I've had success using Smiling Wolfs ViT models in the past to match people. The weights vary in and around 350-400mb. If your dead set on <300mb look around for other models trained for the purpose of tagging stable diffusion images.

if you have articles written in a style / voice you wanna finetune on, create fake RAG context which contains necessary details to write some article and then train on those pairings. You could also just finetune on the raw text itself but it wont be nearly as effective as using a constrained...

if you go on the github there are collab notebooks with step by step instructions. Only confusing part may be the dataset but you can look at the huggingface websites docs for learning about Datasets / finding datasets as well

Its a "complicated server setup" only if you want to host it on your computer, but there's services which charge per use like openai, azure, openrouter, etc. So its openai, azure, and openrouters job to host it on their computers then they give you an API key to use. Openrouter is somewhat...

Long time subscriber, I wouldnt subscribe rn when there's llama3 out which you could use for ridiculously cheaper + a somewhat better result. Although I'm still subscribed since if they release something really cool like a search engine or GPT5 I dont want to be stuck in a waitlist and I'll...

You can try claude, anthropics competitor to chatgpt. Its supposed to do better with larger contexts. Otherwise there are numerous projects on github that specialize in helping read academic papers and you could hook it up to huggingfaces api which (i believe) offers free calls. If you dont want...

@FatBee and whoever interested in botting anything This post was originally asking a few questions but slowly turned into a adderall enduced conspiracy theory rant on the business relations of SMM Panel resellers and actual botters. The questions I'm most interested about if you'd like to skip...

some amount of dependencies are good, but at some point your 80,000mg intake starts looking akin to importing an "is_even" function

Used to drink ~1000mg of caffeine a day and found it barely provided any boost and just raised my baseline "awakeness" up. Stopped drinking it 2 weeks ago and im about as productive as I was at 1000mg per day without the issues of sleep or it being near impossible to work without caffeine

Username: somethingtap

How many separate fanvue accounts do you have currently? How do you plan on handling chatting? Manual, chatters or eventually automated? Fanvue has the benefit of being able to send locked messages but I've had most of my success just sticking with what social media im on.

Review Copy? Username on dripfeed is: applesaucecross last message appeared to slip through the cracks

test balance? Username: applesaucecross