Ive been trying to learn about these AI things like Llama and I really want to run my own version for a project Im doing for school here in Seattle but honestly Im totally lost. I tried running it on my old laptop and it basically sounded like it was going to explode and then just crashed after typing like two words. I think I need some kind of web hosting but when I look at sites like AWS or Google Cloud its like reading another language with all the talk about vCPUs and instances and stuff I just dont get at all. My budget is pretty tight like maybe 50 dollars a month max since Im a student and I dont want to accidentally get a bill for a thousand dollars because I clicked the wrong button. I heard you need something called a GPU for this to work fast but I have no idea which hosting companies actually give you those without charging a fortune or making you sign a crazy contract. Sorry if this is a really dumb question but is there a place thats easy for someone who has no idea what theyre doing to just upload a model and have it work? I just want to play around with it for a few months without it being a total headache. what is actually the best web hosting for running these large language models if youre a total beginner?
In my experience trying many different hosts, RunPod Community Cloud RTX 3090 is the play. It's way cheaper than AWS and wont murder your bank account like you're worried about.