Building and sharing experiments quickly is an important requirement when working with a global team consisting of different technical capabilities and expertise.
There are various tools for both running and sharing ML/DL research work. In this post I’ll review the process of creating and sharing a DL model using Corlab & Binder.
Corlab, a tool created by Google that allows spinning off a free Jupiter notebook which can run on a CPU or a GPU. No need for a strong local machine nor a local setup, it just works smoothly with a web browser.
Corlab is also linkable from Github so it’s easy to make small changes / iterations directly from the repository.
Once I built my model and ran a few training sessions, I was ready to share the model with a colleague for inference validation. For that, I had two requirements: First I don’t want to share my credentials and second I want my colleague to be able to run the inference on demand.
For that I chose to use Binder, a platform to create a Github repo into a executable Jupiter notebook. Basically it warps the repo code and dependencies into docker container and deploys it, all free and fast.
And a couple of minutes later the inference is ‘production’ ready: