Skip to content

Conversation

@MengzhangLI
Copy link
Contributor

@MengzhangLI MengzhangLI commented Aug 29, 2022

Motivation

Adding new loss function named Tversky Loss. https://arxiv.org/abs/1706.05721

Old PR: #634, #919

Modification

Datasets loss_ce (from repo config) loss_ce + loss_tversky (1:1) loss_ce + loss_tversky (1:3) loss_tversky
DRIVE 78.67 79.39 79.81 79.72
STARE 81.02 82.3 82.62 82.03
CHASE_DB1 80.24 80.37 79.74 77.95
HRF 79.45 81.03 80.81 --

@codecov
Copy link

codecov bot commented Aug 29, 2022

Codecov Report

Merging #1986 (c2c7d00) into master (13d4c39) will increase coverage by 0.03%.
The diff coverage is 94.64%.

@@            Coverage Diff             @@
##           master    #1986      +/-   ##
==========================================
+ Coverage   89.04%   89.08%   +0.03%     
==========================================
  Files         144      145       +1     
  Lines        8636     8691      +55     
  Branches     1458     1463       +5     
==========================================
+ Hits         7690     7742      +52     
- Misses        706      707       +1     
- Partials      240      242       +2     
Flag Coverage Δ
unittests 89.08% <94.64%> (+0.03%) ⬆️

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

Impacted Files Coverage Δ
mmseg/models/losses/tversky_loss.py 94.44% <94.44%> (ø)
mmseg/models/losses/__init__.py 100.00% <100.00%> (ø)
mmseg/version.py 58.33% <100.00%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

@MeowZheng MeowZheng merged commit acff839 into open-mmlab:master Aug 30, 2022
@MengzhangLI MengzhangLI deleted the new_tversky branch August 31, 2022 06:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants