💥💯💡A new open source library for working with data has appeared on the Internet
Cleanlab is a library that helps clean data and labels by automatically detecting problems in a machine learning dataset. To make machine learning easier on messy data, this data-centric II package uses additional models to evaluate problems in data sets that can be corrected to train even better models.
As a result, the AI library performs the following functions:
1. Detection of data problems (mislabeling, omissions, duplicates, drift)
2. Setting up and testing the training model.
3. Conduct active training of models