Image biases
Pluto has access to thousands of datasets, and downloading these datasets is as simple as replacing the URL. In particular, he will download the following datasets:
- The State Farm distracted drivers detection (SFDDD) dataset
- The Nike shoes dataset
- The Grapevine leaves dataset
Let’s start with the SFDDD dataset.
State Farm distracted drivers detection
To start, Pluto will slow down and explain every step in downloading the real-world datasets, even though he will use a wrapper function, which seems deceptively simple. Pluto will not write any Python code for programmatically computing the bias fairness matrix values. He relies on your observation to spot the biases in the dataset.
Give Pluto a command to fetch, and he will download and unzip or untar the data to your local disk space. For example, in retrieving data from a competition, ask Pluto to fetch it with the following command:
# fetch real-world data pluto.fetch_kaggle_comp_data...
