Skip to main content

A lightweight tool to perform reproducible machine learning experiment using Dask.

Project description

https://img.shields.io/pypi/v/daskperiment.svg Latest Docs https://travis-ci.org/sinhrks/daskperiment.svg?branch=master https://coveralls.io/repos/sinhrks/daskperiment/badge.svg?branch=master&service=github

Overview

A lightweight tool to perform reproducible machine learning experiment using Dask.

The package is EXTREMELY unstable.

Benefits

  • User-intuitive.

    • Minimizing modifications of existing codes.

    • Performing experiments using Dask compatible API.

    • Easily handle experiments history (with pandas basic operations).

    • Requires less work to manage with Git (no need to make branch per trials).

  • Tracking experiment result and its (hyper) parameters.

  • Tracking environment.

    • OS Info

    • Python version

    • Installed package and its version

  • Tracking code context.

  • Auto saving and loading previous experiment history.

  • Parallel execution of experiment steps.

Future Scope

  • Web Dashboard

  • Reproducibility check (function purity check).

  • Sharing experiments.

  • More efficient execution.

    • Omit execution if depending parameters are the same

    • Distributed execution

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

daskperiment-0.1.0.tar.gz (13.8 kB view details)

Uploaded Source

Built Distribution

daskperiment-0.1.0-py3.6.egg (46.4 kB view details)

Uploaded Source

File details

Details for the file daskperiment-0.1.0.tar.gz.

File metadata

  • Download URL: daskperiment-0.1.0.tar.gz
  • Upload date:
  • Size: 13.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.12.1 pkginfo/1.5.0.1 requests/2.21.0 setuptools/40.6.3 requests-toolbelt/0.8.0 tqdm/4.19.5 CPython/3.6.6

File hashes

Hashes for daskperiment-0.1.0.tar.gz
Algorithm Hash digest
SHA256 c4efd6afb145457a2abcd91ef8035dfbede2aa3738c3c3fa1626e382551d6a29
MD5 d205ed7ef371d59dbfd2358516d416a4
BLAKE2b-256 1e2788e03c84fcc903fb5ca1c31f010f2ccd3d373bb3b1c1992364326ce6e9d3

See more details on using hashes here.

File details

Details for the file daskperiment-0.1.0-py3.6.egg.

File metadata

  • Download URL: daskperiment-0.1.0-py3.6.egg
  • Upload date:
  • Size: 46.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.12.1 pkginfo/1.5.0.1 requests/2.21.0 setuptools/40.6.3 requests-toolbelt/0.8.0 tqdm/4.19.5 CPython/3.6.6

File hashes

Hashes for daskperiment-0.1.0-py3.6.egg
Algorithm Hash digest
SHA256 cd61e0eacbdcec6f5f767d713566d9b4fe734a940a0293fc700b60f8dec300e3
MD5 cea422b7149535b38622d610af9af829
BLAKE2b-256 f822d1181b234919226f2871115ec3b4bc5088a2cbc46c479e78daf3196aa911

See more details on using hashes here.

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