MachineLabs – Run Machine Learning Code in the Browser
blog.machinelabs.ai
blog.machinelabs.ai
How do you compare to Kaggle kernels?
For uploading data sets we're working on a resolver strategy that lets you specify endpoints from where the data can be downloaded before the lab is executed.
We'll provide containers with all common libraries. In the future you'll be able to configure your own containers which will allow you to run pretty much whatever you want.
Your idea can be very interesting! :)
>>Do you have public code for containers as Floydhub has?
Can you elaborate what you mean with that?
>What machines are you using in your cloud? Can I specify the >machine CPUs and RAM size?
We'll have different hardware type categories. At the end it boils down to VPS on Google Cloud Engine running Ubuntu. That said, thinking long-term we'd love to give people access to things like Nvidia Tesla V100.
>Do you have time/memory limit for job execution?
Your experiments can run for as long as you wish (but you'll pay for everything that goes beyond free hours). Memory limits will be defined by the hardware category you chose when you run your lab.
Do you have specific requirements regarding memory or CPU/GPU power? We'd love to learn from people like you what the community needs.
I don't have any specific requirements regarding memory or CPU - just asking. It will be nice, if I can use your service to run some kaggle kernels which take too long on kaggle.
Would love to get your signup for the private beta to get detailed feedback. Especially given your experience with kaggle.
https://blog.machinelabs.ai/2017/05/11/introducing-machinela...