News for hackers. Stuff that matters.
Re:Teachable Machine by tsunamifury on Saturday January 06, 2024 @11:44PM (#38898521)
Isn’t this basically what multimodal LLM does as well… it can do anything on the fly it can understand.

What’s different here?

Re:Teachable Machine by dbish on Sunday January 07, 2024 @12:30AM (#38898759)
The teaching part is what matters, it’s training (tuning in this case) a model, not just using a model already trained for inference (which is what I assume you mean). You’re providing new data that is used to update the model. Inference across an existing multimodal model doesn’t change how it classifies in any way.
Re:Teachable Machine by navanchauhan on Saturday January 06, 2024 @11:55PM (#38898574)
These are smaller scale model that you can export and run anywhere.

Even the smallest multimodal LLM would be wayyyyyyy bigger than an exported model from this

Re:Teachable Machine by jatins on Sunday January 07, 2024 @12:10AM (#38898655)
How small are these models? can I export a model here and embed it in an Android/ios app?
Re:Teachable Machine by jzombie on Sunday January 07, 2024 @12:18AM (#38898701)
The website says they can be embedded in a web app, and export to a format called TensorFlow Lite. I am sure you could embed it.

https://www.tensorflow.org/lite

Re:Teachable Machine by pona-a on Sunday January 07, 2024 @07:55AM (#38900824)
https://www.tensorflow.org/lite/android

TF Lite has first-class Android support with hardware acceleration if I'm not mistaken.

Re:Teachable Machine by vineyardmike on Thursday January 11, 2024 @04:19AM (#38949757)
You can supposedly embed them for Arduinos, so an app should be no problem.
Re:Teachable Machine by jzombie on Saturday January 06, 2024 @11:57PM (#38898585)
I think this is more like fine tuning an existing model to recognize features you specifically intend it to, and be light enough to run locally in a browser.
Re:Teachable Machine by tsunamifury on Sunday January 07, 2024 @12:04AM (#38898629)
Thanks!
Re:Teachable Machine by jerbear4328 on Sunday January 07, 2024 @02:13AM (#38899216)
It's not even fine tuning, it's creating a model from scratch. This isn't like our modern huge models either, these tiny single-purpose models have been around for ages and are quite versatile. They're so small you can't just easily run them in the browser, but train them effectively, which is what this project lets you play around with! Super cool stuff.