125 points by ml4all 5 months ago flag hide 21 comments
johnsmith 5 months ago next
Great to see more effort towards accessible machine learning! I'm thrilled to try this out.
alice123 5 months ago next
I agree with johnsmith. Exciting progress and hope it democratizes ML further.
hackingislife 5 months ago prev next
Looks promising! But will it scale to handle large datasets?
machinelearningnerd 5 months ago next
It's optimized for scalability and relies on distributed computing. See user guide for more info.
optimusprime 5 months ago prev next
It can efficiently handle terabyte-scale datasets. I'm impressed!
algorithm_wiz 5 months ago next
@optimusprime, can you explain how you tested scalability?
optimusprime 5 months ago next
@algorithm_wiz, we used a cloud-based managed workflow with 10,000 CPU cores and 2 PB RAM.
curiouscoder 5 months ago prev next
How does this compare to other accessible ML tools like TensorFlow's Playground?
megan_d 5 months ago next
This toolkit is more comprehensive, with better support for custom models. TensorFlow Playground is limited to pre-built models.
nlpgeek 5 months ago next
Nice. This toolkit will help non-technical users create custom NLP models, not just image-based ones.
megan_d 5 months ago next
@nlpgeek, absolutely! The toolkit is designed for NLP, image recognition, and even time series analysis.
karpathydev 5 months ago prev next
Will this be compatible with 3rd party libraries like Keras or Fast.ai?
devopsteam 5 months ago next
Yes, this toolkit is designed with extensibility in mind. There's a plugin architecture for adding 3rd party library support.
codewizz 5 months ago prev next
Interesting! Do follow up on the development and thank you for sharing 🙏
artificialmind 5 months ago next
@codewizz, glad to hear you're interested. We'll have bi-weekly updates. Follow us and contribute on GitHub!
intelinsid 5 months ago prev next
What's the story behind this project? Who created it?
devopsteam 5 months ago next
@intelinsid, a team of researchers from Open Research Institute (ORI) started this toolkit for democratizing ML.
deeplearner 5 months ago prev next
I'm concerned about the complexity of the setup. Can it be simplified?
machinelearningnerd 5 months ago next
@deeplearner, a single installer for macOS, Linux, and Windows is in the works. It'll make things easier.
hadoopstar 5 months ago prev next
I like the open source nature, but who will maintain it long term?
devopsteam 5 months ago next
@hadoopstar, a community-driven stewardship & maintenance model is adopted. Anyone with PRs can join!