This idea depends on: https://ibmwatsondataplatform.ideas.aha.io/ideas/IBMWDP-I-12
I would like to be able to prototype neural networks in notebooks. Realistically this will require GPUs for performance reasons hence the dependency on IBMWDP-I-12. However, just having GPUs will not be enough.
The Deep Learning coding guideline documentation lists other steps that the deep learning program should follow to make your code work with the Watson Machine Learning backend GPU cluster. Not all of these guidelines will work in notebooks without hacky code such as setting environment variables. This will require separate code branch logic for each environment that the code runs on.
Additionally, The Watson Machine Learning service expects the deep learning code to be located in object store. This will require having to export a notebook and then uploading the code manually.
Ideally the process should be made much more seamless so that deep learning engineers can go from rapid prototyping in notebooks to experiments without much overhead.
Why is it useful?
|Who would benefit from this IDEA?||deep learning engineer|
How should it work?
|Submitting Organization||F2F Sales|