Open-Source Machine Learning Repos to Inspire Your Next Project | Hacker News Reader

Open-Source Machine Learning Repos to Inspire Your Next Project

heartbeat.fritz.ai

135 pointsonmyway1337 comments

I'm curious why I don't see any projects applying new ML techniques to text recognition (OCR) in those "advertisements" popping up recently; is this because:

- too boring/mundane?

- "solved problem"/too easy? (then where are the [open source] solutions? is Tesseract-ocr advanced enough already? what's its status?)

- too hard? (is Tesseract-ocr using the state-of-the-art techniques?)

Nice wrap up of ideas, but I wonder why people don't get attention on ML project with structured data, which is the most common btw.

I know that predicting sales is not as fun as make Horse a Zebra but I think they're underrated.

pd: If you have nice ML repos with structured data I'll appreciated :)

Great point! I haven’t tried it yet, but Sales Force just opensourced Transmogrifai, a platform that does just this:

https://engineering.salesforce.com/open-sourcing-transmogrif...

Thanks for the list, I've been thinking of contributing to an OS project recently. These all look extremely interesting, hopefully I can be helpful

Any of these work completely in iOS? I'm new in the field and want to embed everything in mobile

Core ML is going to be your best bet. Most training is still done server side using frameworks like TensorFlow, Keras, and PyTorch. Once you've trained your model, you can convert it to Core ML with coremltools or export it to Core ML directly if the platform supports it.

Apple has a couple tools, Turi Create and Create ML, to train ML models specifically for mobile use, but their not nearly as fully featured or widely used.

If you're interested specifically in mobile ML, check out https://heartbeat.fritz.ai. We've got a bunch of resources for mobile machine learning. If you're looking for ready-to-use models or tools to manage them in your app, check out Fritz (https://fritz.ai). Disclaimer, I'm a founder at Fritz which sponsors Heartbeat. Happy to answer any questions!

I'm not sure yet whether i want to do the training on mobile, but i want to try both before i decide so thanks for the resources i'll dig into them. One important question is, are there tools to decrease the size of a model so it'd better fit on mobile? Or is that part of the training?

Reply on news.ycombinator.com