DI-toolkit 0.0.1

A simple tool for automatic parameter tuning.

DI-toolkit

A simple tool for automatic parameter tuning.

Installation

You can simply install it with pip command line from the official PyPI site.

pip install DI-toolkit

For more information about installation, you can refer to Installation.

Quick Start for HPO

Here is a simple example

import random
import sys
import time

from ditk.hpo import hpo, R, M, choice, uniform, randint


@hpo
def opt_func(v):  # this function is still usable after decorating
    x, y = v['x'], v['y']
    time.sleep(0.1)
    print("This [u]time's[/] config:", v)  # stdout will be captured
    print("This is print line in stderr", file=sys.stderr)  # stderr will be captured
    if random.random() < 0.5:  # randomly raise exception
        raise ValueError('Fxxk this shxt')  # retry is supported

    return {
        'result': x * y,
        'sum': x + y,
    }


if __name__ == '__main__':
    print(opt_func.random()  # random algorithm
          .max_steps(30)  # max steps
          .minimize(R['result'])  # the maximize/minimize target you need to optimize,
          .concern(M['time'], 'time_cost')  # extra concerned values (from metrics)
          .concern(R['sum'], 'sum')  # extra concerned values (from return value of function)
          .stop_when(R['result'] <= -800)  # conditional stop is supported
          .spaces(  # search spaces
        {
            'x': uniform(-10, 110),  # continuous space
            'y': randint(-10, 20),  # integer based space
            'z': {
                't': choice(['a', 'b', 'c', 'd', 'e']),  # enumerate space
            },
        }
    ).run())

Contributing

We appreciate all contributions to improve DI-toolkit, both logic and system designs. Please refer to CONTRIBUTING.md for more guides.

License

DI-toolkit released under the Apache 2.0 license.