Skip to main content

Package for evaluating object detection models

Project description

D.E. Bobo - Detection Evaluation for Bounding Boxes

DEBobo is a library providing an easy-to-use evaluation code for object detection models. It's main motivation was to replace the part of pycocotools responsible for evaluation as it doesn't work well with custom datasets. Additionally, I found the workflow of pycocotools doesn't work well with high-level training libraries such as ignite.

Installation

Most end users should be able to get away with

pip install debobo

For development, install the library with symlinks and with the additional test requirements using

pip install -e .[test]

To run the tests, first download the test-data using ./fetch_testdata.sh. The test is run via

pytest tests/

It compares the results obtained from debobo to the result obtained using pycocotools.

Usage

We provide ready-to-use metrics for ignite in debobo.adapters.ignite. Feel free to request other adapters as an issue. Also, check the tests for how to use the low-level routines.

Thanks

Thanks to @martiningram for the header image.

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

debobo-0.1.6.tar.gz (6.8 kB view hashes)

Uploaded Source

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