Skip to content

Releases: open-mmlab/mmeval

MMEval V0.2.1 Release

03 Apr 08:07
5b4ef8d
Compare
Choose a tag to compare

Features

Enhancements

  • Update the test case style of PCKAccuracy by @ice-tong in #68
  • Rename metrics by @ice-tong in #83
  • Make scipy as a runtime dependency by @C1rN09 in #109
  • Refactor single and multi label metrics by @yingfhu in #81
  • Refactor detection dataset metainfo to lowercase and update detection metric logics by @BIGWangYuDong in #98

Bug fixes

  • Fix CI failed caused by OneFlow installation and new version of NumPy by @ice-tong in #73
  • Fix import error in python3.6 by @ytzhao in #77
  • Upgrade the version of isort to fix lint error by @zhouzaida in #85
  • Use official isort by @zhouzaida in #86
  • Fix psnr,snr,ssim,mae and mse fail to compute on videos by @Z-Fran in #89
  • CircleCI fail due to plum-dispatch v2.0.0 by @C1rN09 in #108
  • Fix format by @zhouzaida in #111
  • Add opencv to docs dependency & setup.py to resolve API docs error by @C1rN09 in #113

Others

New Contributors

Full Changelog: v0.2.0...v0.2.1

MMEval V0.2.0 Release

07 Dec 11:51
8a5ed43
Compare
Choose a tag to compare

Since MMEval v0.2.0, we support evaluation with OneFlow!

Features

  • Add oneflow distributed backend (#59)
  • Support seven metrics for OneFlow backend (#58)

Documentations

  • Add mmeval into PYTHONPATH for API doc generation(#45)
  • Update qq group link (#56)
  • Add OneFlow support (#62)

Contributors

A total of 4 developers contributed to this release.
@vansin @ofhwei @zhouzaida @ice-tong

New Contributors

Full Changelog: v0.1.0...v0.2.0

MMEval V0.1.0 Release

31 Oct 09:08
c9e24bd
Compare
Choose a tag to compare

We are excited to introduce our new open-source library: MMEval, a unified and open cross-ML framework evaluation library.

Install MMEval via pip:

pip install mmeval

MMEval is a machine learning evaluation library that supports efficient and accurate distributed evaluation on a variety of machine learning frameworks. Major features:

  • Comprehensive metrics for various computer vision tasks (NLP will be covered soon!)
  • Efficient and accurate distributed evaluation, backed by multiple distributed communication backends
  • Support multiple machine learning frameworks via dynamic input dispatching mechanism