Chexpert data is in a csv file. Each row has a set of features and has an path to an image and some labels.
Leaf data is in a json file. The data is supposed to be with different users. The top level keys in the json file are - users, num_samples and user_data.
users - a list of users (or fed learn nodes)
num_samples - a list again, number of samples per user. All users have same number of data (initially, i think this is known as an iid distribution).
user_data - todo
(please add some code snippets below)
CHEXPERT DATA ---> INITIAL PREPROCESSING (map to numeric categories and convert image dimensions) ----> Map to Leaf json format
Open questions :-
1) What are our image dimensions?
Chexpert data is in a csv file. Each row has a set of features and has an path to an image and some labels.
Leaf data is in a json file. The data is supposed to be with different users. The top level keys in the json file are - users, num_samples and user_data.
(please add some code snippets below)
CHEXPERT DATA ---> INITIAL PREPROCESSING (map to numeric categories and convert image dimensions) ----> Map to Leaf json format
Open questions :- 1) What are our image dimensions?