Hacker News Reader

Comment by jpeter | Hacker News Reader

Which consumer gpu runs llama 70B?

A single 3090, or any 24GB GPU. Just barely.

Yi 34B is a much better fit. I can cram 75K context onto 24GB without brutalizing the model with <3bpw quantization, like you have to do with 70B for 4K context.

Can it produce any meaningful outputs with such an extreme quantisation?

Yeah, quite good actually, especially if you quantize it on text close to what you are trying to output.

Llama 70B is a huge compromise at 2.65bpw... This does make the much "dumber." Yi 34B is much better, as you can quantize it at ~4bpw and still have a huge context.

How would you compare mistral-7b-instruct 16fp (or similar 7b/13b model like llama2 etc) to Yi-34b quantized?

34B is better. Quantization hurts some, especially in "pro" non chat use cases like RAG, but the increased parameter count makes models so much smarter in comparison.

The perplexity graph here is a pretty good illustration: https://github.com/ggerganov/llama.cpp/pull/1684

YMMV, as Mistral and Yi are not necessarily comparable like different sizes of llama, and it depends on the task.

Prosumer gear.

MacBook Pro M3 Max.

A Mac with a lot of unified RAM can do it, or a dual 3090/4090 setup gets you 48gb of VRAM.

Does this actually work? I had thought that you can't use SLI to increase your net memory for the modal?

It works. I use ollama these days, with litellm for the api compatibility, and it seems to use both 24GB GPUs on the server.

I’ve got a 64gb Mac M2. All of the openllm models seem to hang on startup or on API calls. I got them working through GCP colab. Not sure if it’s a configuration issue or if the hardware just isn’t up to it?

Valiant et al work great on my 64Gb Studio at Q4_K_M. Happy to answer questions.

Try llama.cpp with Metal (critical) and GGUF models from TheBloke

Or wait another month or so for https://ChatOnMac.com

Reply on news.ycombinator.com