Skip to content
New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

Bump version to v0.3.0 #260

Merged
merged 4 commits into from
Apr 29, 2022
Merged

Bump version to v0.3.0 #260

merged 4 commits into from
Apr 29, 2022

Conversation

zytx121
Copy link
Collaborator

@zytx121 zytx121 commented Apr 28, 2022

v0.3.0 (29/4/2022)

Highlight

New Features

  • Update performance of ReDet on HRSC2016. ([Reimplementation] Update performance of ReDet on HRSC2016 #203)

  • Upgrade visualization to custom colors of different classes ([Enhance] Refactor imshow_det_rbboxes #187)

  • Update Stable KLD, which solves the Nan issue of KLD training. ([Algorithm] Update Stable KLD (NeurIPS'21) #183)

  • Support setting dataloader arguments in config and add functions to handle config compatibility. ([Feature] Add dataloader arguments in config #215)
    The comparison between the old and new usages is as below.

    Before v0.2.0 Since v0.3.0
    data = dict(
        samples_per_gpu=2, workers_per_gpu=2,
        train=dict(type='xxx', ...),
        val=dict(type='xxx', samples_per_gpu=4, ...),
        test=dict(type='xxx', ...),
    )
    # A recommended config that is clear
    data = dict(
        train=dict(type='xxx', ...),
        val=dict(type='xxx', ...),
        test=dict(type='xxx', ...),
        # Use different batch size during inference.
        train_dataloader=dict(samples_per_gpu=2, workers_per_gpu=2),
        val_dataloader=dict(samples_per_gpu=4, workers_per_gpu=4),
        test_dataloader=dict(samples_per_gpu=4, workers_per_gpu=4),
    )
    
    # Old style still works but allows to set more arguments about data loaders
    data = dict(
        samples_per_gpu=2,  # only works for train_dataloader
        workers_per_gpu=2,  # only works for train_dataloader
        train=dict(type='xxx', ...),
        val=dict(type='xxx', ...),
        test=dict(type='xxx', ...),
        # Use different batch size during inference.
        val_dataloader=dict(samples_per_gpu=4, workers_per_gpu=4),
        test_dataloader=dict(samples_per_gpu=4, workers_per_gpu=4),
    )
  • Add get_flops tool ([Feature] Support get flops tool #176)

Bug Fixes

Improvements

Contributors

A total of 7 developers contributed to this release.
Thanks @nijkah @GamblerZSY @liuyanyi @yangxue0827 @jbwang1997 @zytx121 @ZwwWayne

@codecov
Copy link

codecov bot commented Apr 28, 2022

Codecov Report

Merging #260 (427c2b2) into dev (e555fa2) will decrease coverage by 0.17%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##              dev     #260      +/-   ##
==========================================
- Coverage   29.30%   29.13%   -0.18%     
==========================================
  Files         113      113              
  Lines        7298     7298              
  Branches     1110     1110              
==========================================
- Hits         2139     2126      -13     
- Misses       5083     5094      +11     
- Partials       76       78       +2     
Flag Coverage Δ
unittests 29.13% <100.00%> (-0.16%) ⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files Coverage Δ
mmrotate/version.py 61.53% <100.00%> (ø)
mmrotate/datasets/builder.py 25.92% <0.00%> (-25.93%) ⬇️
mmrotate/utils/setup_env.py 68.00% <0.00%> (-24.00%) ⬇️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e555fa2...427c2b2. Read the comment docs.

Copy link
Collaborator

@yangxue0827 yangxue0827 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@ZwwWayne ZwwWayne merged commit c0c286e into open-mmlab:dev Apr 29, 2022
# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants