Hyperdash is a machine learning monitoring system that provides remote push notification, visualization, and updates about training data through a smartphone app. The service can monitor completion, failure, hyperparameters, and detect output from machine learning experiments for data scientists. The hyperdash SDK is available for Python and R programming languages.
The Hyperdash SDK for Python and R language are available on GitHub. Hyperdash needs an API key to perform its monitoring. The signup process for an API key is done on the terminal after installation of python SDK. The API key is installed after the signup process completes and is stored in a hyperdash.json file.
The python SDK can be installed through pip package installer. The hyperdash python SDK can run alongside Tensorflow, Scikit-Learn, and other machine learning modeling libraries. Multiple API keys can be used to represent different machine learning projects. After the SDK is running with an API key, a library can be included onto python scripts or Jupyter to monitor hyperparameters across different model experiments. The hyperdash library is compatible with Python version 2.7 through 3.6.
Hyperdash can be used for pure logging and notification purpose or monitor training models hyperparameters and performance metrics. Experiment module from Hyperdash is used to monitor for hyperparameters and performance metrics. The data metric reports are created in real time to be viewed remotely through the smartphone app.
The Hyperdash R SDK is installed through the install.packages() function. New API creation still needs python SDK installation. Once the hyperdash library is imported onto R code, a function to be monitored is passed through the Monitor function.
The Hyperdash app for monitoring machine learning model’s real-time data is available for iOS and Android devices.
Documentaries, videos and podcasts