HenryNdubuaku

Great work, thanks!

Thanks!

Yes, thanks!

Exactly the goal!

The base model is a Simple Attention Network, a foundation model family we’ve been experimenting on at Cactus.

Haha, yeah I’m here, mostly quiet though

Bullseye!

Yes, let us know how it goes!

Yeah, but we wanted Gemini

We have so far released the dataset generation code

Give it a go and let us know!

Thanks, keep me posted!

You can think of “phone use” for instance, what Siri is supposed to be.

No question is stupid! 1. Distilled means taking the intelligence of a big model and compacting into a tiny model. 2. Google already does so with FunctionGemma, but Needle argues that better performan…

Music to our ears!

Thanks, let me know how it goes!

Gemini is pretty solid for 1-shot tool call and affordable as well.

It is for building agentic capabilities into very small devices like phones, glasses, watches and more. Does that make sense?

Sounds cool, play with it and let uk know what you think!

Deploying AI on tiny devices like watches, earphones, glasses etc.

Let me know what you think!

Cheaper APIs

The codebase is fully open, feel free to play around!

Not yet, for now. But it’s in the works!

So it’s a tiny model capable of function calling that could run locally on cheap devices.

Yes, a demo might be a good idea.

Thanks!

We include the dataset pipeline in the codebase so far, might release dataset.

Sounds interesting, would love to see it too!

Thanks, Needle doesn’t compete with those tools though and the distillation process did not access the weights.