/Sentience

An AutoML framework to define custom models based on custom datasets.

Primary LanguagePythonMIT LicenseMIT

Sentience

A framework to automatically search for the best models using AutoML philosophies. Just define a dataset and other optional hyperparameters and wait for the framework to search the best model that can surpass some human defined models as well. The workflow definition is done using a .stnt file which the compiler sentienc parses and creates a workspace. We can define minimal hyperparameters like dataset and epochs per model to get the system started, or we can opt for more granular controls.