AirLLM 70B inference with single 4GB GPU

AAnon84 about 4 hours ago 29 commentsRead Article on github.com

ZH version is available. Content is displayed in original English for accuracy.

⚡ Community Insights

Discussion Sentiment

53% Positive

Analyzed from 868 words in the discussion.

Discussion (29 Comments)Read Original on HackerNews

imenani•about 2 hours ago

For anyone wondering “how slow is this?”

IIUC, Kimi K3 on RTX 6000 Ada (48GB) takes 292 s/token

https://github.com/lyogavin/airllm/releases/tag/v3.1.0

bensyverson•about 1 hour ago

I do think these “run a bigger model than will fit in VRAM” projects are necessary steps, but are they functionally useful or helpful to anyone currently? For example, is anyone out there running a big Qwen for coding on a 16-32GB machine with these techniques?

pizza234•about 1 hour ago

> but are they functionally useful or helpful to anyone currently?

Yes and no, depend on your expectations. Some/many like to run local LLMs just for the sake of it, so anything will do.

MoE are useful on PC systems, at the condition of having high enough memory bandwidth (and large amounts of RAM) - that is, Threadripper/Pro.

The advantage of MoE is that only a subset of the model's experts is used for each token, so not all weights need to be present in VRAM at once. The remaining weights can reside in system RAM, although moving and accessing them still carries a substantial performance cost (and that's why high memory bandwidth is needed).

bensyverson•19 minutes ago

Are people getting decent tokens/second throughput? Some of these demos crawl at 1 tok/s or worse, which limits their utility.

ddevnyc•21 minutes ago

Does MoE help with multimodality? Can it in general enable reasoning in imagery (technical drawings, diagrams, schematics) rather than text-based?

nickpsecurity•11 minutes ago

If I could justify wear and tear and electricity, I was willing to do something like this for batch processing. The batches would be a bunch of prompts whose outputs I'd look at the next day. Maybe common operations, like QA or refactoring, on whatever software I wrote.

If so, I could use a larger model than I have real-time hardware for. The largest, well-trained models can often get the output mostly right in one try. I also would be using AI's as a supplement to, not replacement for, my own brain. So, issues with the outputs wouldn't be a problem because I'm just keeping what's helpful.

If I still need to re-generate it all, it might still save money over time by avoiding cloud costs. Also, hardware that's already paid for is a sunk cost that doesn't inflate over time. Glitches in loading or destroying VM's might blow up into a big bill.

xnorswap•about 1 hour ago

I wonder what this measures in J/token.

throwawayffffas•25 minutes ago

Assuming 30% gpu power utilization because of all the loading and unloading 29.2 kJ per token

jackb4040•about 1 hour ago

Ahaha thank you, I naively assumed the unlabeled graph in the readme was tps, not spt!

throwawayffffas•23 minutes ago

Running 292 tps on K3 would make your gpu a money printer.

mring33621•22 minutes ago

How many is that in tokens per Scaramucci?

logicallee•32 minutes ago

that's 0.003 tokens/second. To get an hour's work done that's normally 30 tokens/second (108k output tokens in an hour) will take 416 days at this rate. And if you're using 100 watts, during that time you will spend $124.61 in electricity, as well as not being able to use your device for something else, plus the noise and heat from your device.

For $124, on Moonshot's official Kimi K3 API rates ($0.30 per 1M cached input, $3 per 1M fresh input, $15 per 1M fresh output), you can purchase 42 million fresh-input tokens, or 8.3 million generated output tokens, in whatever mix you want.

So what you get is 80x more expensive and you wait 416 days to get it.

roger_•about 2 hours ago

Seeing a lot of these “run 1TB models with 1GB RAM” projects recently. Most seem vibe coded and probably won’t be maintained.

Hoping a winner emerges with some real momentum behind it.

hiramwen•about 1 hour ago

You don't really need a maintainer when codex or claude code can set it up for you; thats how I got Trellis2 working on windows and tiny VRAM despite Microsoft recommending you have 24GB VRAM and Linux. Models are pretty disposable now.

titularcomment•24 minutes ago

This is complete dependency on LLM tools all the way from development to usage, and is risky as well as prone to failure

hiramwen•9 minutes ago

Local inference without access to internet and tools is pretty safe, worst you get is a slop output, best case it solves your prompt.

akie•about 1 hour ago

No no, don't just say "vibe coded", say "Fable and $500 of credits"

Lalabadie•35 minutes ago

"This one is made with premium vibes"

classified•about 1 hour ago

How a patchwork of Python modules constitutes an API(?) or whatever(?) and how to use it (does anyone?), is beyond me. There is nothing that I would call documentation, let alone concrete usage examples. I know more amusing ways to waste my time. When I want to play with models, I use llama.cpp.

cpfohl•about 2 hours ago

I’m still slightly confused on what this adds.

Let’s say I wanted to run a full size open weight model. I have a 128GB m3 max laptop.

Does this basically load layers in and out on demand? So I still have to download the full model to disk, but the RAM requirements go way down? The readme calls out that one still needs to connect HuggingFace, which leads me to believe that maybe you don’t even need to download the full model?

dofm•about 2 hours ago

If you point it at a huggingface model identifier, it will download it, I assume. No way around that.

It reads like it is keeping only the core and the active layer loaded at any one point, and streams layers from disk; there are several other solutions like this and if my understanding is right, this is probably better than an mmap implementation or just streaming experts in.

pvtmert•about 1 hour ago

Not an expert in this field, but the "expert" is consisting of multiple layers. To keep it small in terms of memory print, this project streams each layer (dividing even further).

It also requires extra space because of decomposition of the layers. Normally the file format optimized for compute intense workloads. But here the bottleneck is the memory capacity.

Also guessing that you need to be able to hold at least 3-layers at once in the memory, given M x N = R operation, M is the previous layer, N is next, and R is the result. on the next "layer", the R (result) becomes M, gets computed against the next layer, N, yielding the further result R'. And so on, until all layers are processed.

I assume it's horribly slow, but can be put in a non-intrusive background task...

cpfohl•about 1 hour ago

Obviously needs downloading eventually :).

It seems like this tool saves on both disk space and RAM, then. Classic trade off: speed vs space.

ilaksh•about 1 hour ago

I guess the use case is something like: you have a slightly obsolete Mac or PC or a whole bunch of them, and just need to compose one or more convincing spam emails, but it's fine if it takes a full week to do it?

speedgoose•about 1 hour ago

And you don’t pay for the electricity.

myshapeprotocol•41 minutes ago

Running 70B on a 4GB GPU is wild. Really impressive engineering feat for resource-constrained environments.

book_mike•about 1 hour ago

We will see if this project has legs. This is the kind of efficiency we desperately need. Now if we can address efficiency with llm training.

hna8hjbqzy•about 2 hours ago

Same problem, different day