Web application for scheduling, monitoring and analyzing cloud-based hyper-parameter optimization runs
At DeepCure, we train and evaluate hundreds of machine learning models for different projects every day. We have developed algorithms that use cloud-based, distributed computing to perform efficient hyper-parameter optimization, i.e. selecting the best parameters for each model. Currently the algorithms are run by command line scripts, and the results are analyzed by manual inspection of a database.
Have a web-based platform that allows the machine learning scientists at DeepCure to schedule, monitor and analyze the results of a hyper-parameter optimization run. We expect the platform to minimize the overhead in problem solving and increase the speed of our hypothesis testing cycle.
- Programming Languages
- Frameworks
- Services
Setup can be found here
