Skip to main content

Assist small-scale machine learning.

Project description

learning-utility

Assist small-scale machine learning.

learning-utility is a package of utilities for small-scale machine learning tasks with scikit-learn.

image image PyPI - Python Version Downloads PyPI

Installation

pip install Lutil

Key Features

Cache Intermediate Results

InlineCheckpoint can cache the computation result in the first call. Since then, if nothing has changed, it retrieves the cache and skips computation.

Suppose you have such a .py file.

from Lutil.checkpoints import InlineCheckpoint

a, b = 1, 2
with InlineCheckpoint(watch=["a", "b"], produce=["c"]):
   print("Heavy computation.")
   c = a + b

print(c)

Run the script, you will get:

Heavy computation.
3

Run this script again, the with-statement will be skipped. You will get:

3

Once a value among watch changes or the code inside the with-statement changes, re-calculation takes place to ensure the correct output.

Save Prediction Result According to the Given Format

Lots of machine learning competitions require a .csv file in a given format. Most of them provide an example file.

In example.csv:

id, pred
1, 0.25
2, 0.45
3, 0.56

Run:

>>> import numpy as np
>>> from Lutil.dataIO import AutoSaver

>>> result = np.array([0.2, 0.4, 0.1, 0.5])
       # Typical output of a scikit-learn predictor

>>> ac = AutoSaver(save_dir="somedir", example_path="path/to/example.csv")
>>> ac.save(result, "some_name.csv")

Then in your somedir/some_name.csv:

id, pred
1, 0.2
2, 0.4
3, 0.1
4, 0.5

It also works if the result is a pandas DataFrame, Series, 2-dim numpy array, etc. Also, the encoding, seperator, header, index of the example.csv will all be recognized.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

Lutil-0.1.10.tar.gz (8.8 kB view hashes)

Uploaded Source

Built Distribution

Lutil-0.1.10-py3-none-any.whl (22.9 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page