Loading Dataset with custom splits

🀗Datasets
1

Is there any way to load datasets with custom splits ?
For example : load_dataset(‘csv’,data_files=[‘myfile1.csv’,‘myfile2.csv’],splits=[‘csv1’,‘csv2’]))
Instead of “train, test, and validation,” we can have multiple splits with custom names.

2

Hello :wave:

Something like that should work

data_files = {"csv1": "myfile1.csv", "csv2": "myfile2.csv"}
dataset = load_dataset("namespace/your_dataset_name", data_files=data_files)