run

python runner_datasets.py --configs configs/knn_clsf.json 

to generate datasets specified in config files in the config json.

This only run the program fit once without prediction. We use it for debugging.

The scripts may take one argument that specify how many cores is used for low-level parallelization. Use

kdtree.py 112

to call the kd-tree knn with 112 cores.