Machine Learning

1 readers

1 users here now

Community Rules:

Be nice. No offensive behavior, insults or attacks: we encourage a diverse community in which members feel safe and have a voice.
Make your post clear and comprehensive: posts that lack insight or effort will be removed. (ex: questions which are easily googled)
Beginner or career related questions go elsewhere. This community is focused in discussion of research and new projects that advance the state-of-the-art.
Limit self-promotion. Comments and posts should be first and foremost about topics of interest to ML observers and practitioners. Limited self-promotion is tolerated, but the sub is not here as merely a source for free advertisement. Such posts will be removed at the discretion of the mods.

founded 1 year ago

MODERATORS

communick@academy.garden

[D] Best value GPU for running AI models? (alien.top)

submitted 11 months ago by IfImhappyyourehappy@alien.top to c/machinelearning@academy.garden

4 comments fedilink hide all child comments

Looking to spend between $300 to $800 max for a gpu that will run ai models efficiently. The 1080 TI has 11 GB of ram, but no tensor cores, so it seems like not the best choice.

Just got a new rig, with a 3080 super, which I thought would be good, but it only has 8 GB of ram, big bummer, so I want to replace it with something that will do a better job. What should I get?

top 4 comments

sorted by: hot top controversial new old

[–] lonelyStupidTree@alien.top 1 points 11 months ago

If you have a good PSU( above 850w and recent ), a used 3090 is probably your best bet.

[–] gevorgter@alien.top 1 points 11 months ago

I am not sure what you call "run ai models"

I train on GPU but then i move to prod and it runs on CPU. The inference is not near as computation intensive as training so CPU might be enough.

I converted my model to onnx format, wrote code in C# and packaged it like AWS lambda function. My AI model is not called constantly and is only a small part of my project. So my AWS bill is literally couple dollars a month.

[–] Oak_fields@alien.top 1 points 11 months ago

I got a membership on google colab and I think it’s really worth it compared to buying your own GPU. The GPUs that you can use there are so much faster than what you can buy with that budget.

[–] DoctorFuu@alien.top 1 points 11 months ago

Depends on what your requirements for AI models are, both now and in the future.