How to use UpTrain? 💻

Integrating UpTrain in existing ML pipelines

UpTrain allows seamless integration into existing ML pipelines (feel free to open a new issue if you think an important integration is missing). To this end, users need to define two key variables:

1. The UpTrain Config: The UpTrain config is a dictionary that contains all the settings and parameters needed to run a monitoring experiment. The users can pass information such as the monitors they want to add (e.g., data drift, edge case, model bias, etc.), training arguments (which can contain the training function, location of training data, etc.), logging arguments, etc.

2. The UpTrain Framework: Framework is an object that takes the config as an input and performs all the tasks, such as monitoring, smart data collection, logging, model refinement, etc., in the background.

Let's discuss the UpTrain config and framework in more detail next.