Insights from the deep learning industry.
When we founded Valohai two years ago, we were lucky to make friends with team leads for Uber’s Michelangelo machine learning platform. Michelangelo has been an inspiration in building Valohai for the other 99.999...% of companies that aren’t Uber but still need to speed up their machine learning through automation.
In part 1, we looked at the theory behind Q-learning using a very simple dungeon game with two strategies: the accountant and the gambler. This second part takes these examples, turns them into Python code and trains them in the cloud, using the Valohai deep learning management platform.
By now you’ve surely heard about Kubeflow, the machine learning platform based out of Google. Kubeflow basically connects TensorFlow’s ML model building with Kubernetes’ scalable infrastructure (thus the name Kube and Flow) so that you can concentrate on building your predictive model logic, without having to worry about the underlying infrastructure. At least in theory.
There’s only one way to grow your deep learning team effectively: by adding new people to it! (We were just as shocked as you are by this revelation!) Filling your team can be done a couple ways: by recruitment, hiring freelancers, or outsourcing to consulting agencies. Finding talented people is hard enough already, so make sure your newly hired team members hit the ground running and don’t slow down the rest of the team.
This is the first part of a tutorial series about reinforcement learning. We will start with some theory and then move on to more practical things in the next part. During this series, you will not only learn how to train your model, but also what is the best workflow for training it in the cloud with full version control using the Valohai deep learning management platform.
This tutorial will demonstrate how to take a single cell in a local Jupyter Notebook and run it in the cloud, using the Valohai platform and its command-line client (CLI).
Valohai now supports random search for hyperparameter optimization (which we call the Tasks feature), which has been proven in the aptly named paper Random search for hyper-parameter optimization to be an efficient way to find “neighborhoods” of likely-to-be-optimal hyperparameter values, which can then be iterated further to find the really good values.
Since the rise of the deep learning revolution, springboarded by the Krizhevsky et al. 2012 ImageNet victory, people have thought that data, processing power and data scientists were the three key ingredients to building AI solutions. The companies with the largest datasets, the most GPUs to train neural networks on, and the smartest data scientists were going to dominate forever.
Watch a recording of the webinar on version control in machine learning that was held on 22th of November 2018. During the webinar we discussed about the topics below and answered multiple questions addressed by the attendees.