Skip to content
New issue

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

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

Already on GitHub? Sign in to your account

[Feature] Support FastFCN #885

Merged
merged 25 commits into from
Sep 30, 2021
Merged

[Feature] Support FastFCN #885

merged 25 commits into from
Sep 30, 2021

Conversation

MengzhangLI
Copy link
Contributor

@MengzhangLI MengzhangLI commented Sep 18, 2021

Introduction

Support an interesting algorithm which could use smaller feature map getting good results with less computation (FLOPs), FastFCN: Rethinking Dilated Convolution in the Backbone for SemanticSegmentation.

Original Repo: https://github.com/wuhuikai/FastFCN

Its framework is shown as below:
image

And core novelty of it is Joint Pyramid Upsampling (JPU):
image

Modification

  1. Add a new neck modified from https://github.com/wuhuikai/FastFCN
  2. Add config files.

Current results on Cityscapes using MMSegmentation are below:

image

Future work

TO DO:

  • Aligning training & testing metric of JPU for PSPNet, DeepLabV3 and EncNet.
  • Adding config files.
  • Making fair comparison about FPS.
  • Adding README and uploading models & logs.

(Optional) Results on ADE20K and Pascal Context.

@MengzhangLI MengzhangLI added the WIP Work in process label Sep 18, 2021
@codecov
Copy link

codecov bot commented Sep 18, 2021

Codecov Report

Merging #885 (222d4b3) into master (d966f98) will increase coverage by 0.06%.
The diff coverage is 100.00%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master     #885      +/-   ##
==========================================
+ Coverage   89.77%   89.83%   +0.06%     
==========================================
  Files         116      117       +1     
  Lines        6511     6555      +44     
  Branches     1011     1019       +8     
==========================================
+ Hits         5845     5889      +44     
  Misses        463      463              
  Partials      203      203              
Flag Coverage Δ
unittests 89.83% <100.00%> (+0.06%) ⬆️

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

Impacted Files Coverage Δ
mmseg/models/necks/__init__.py 100.00% <100.00%> (ø)
mmseg/models/necks/jpu.py 100.00% <100.00%> (ø)

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 d966f98...222d4b3. Read the comment docs.

mmseg/models/necks/jpu.py Outdated Show resolved Hide resolved
mmseg/models/necks/jpu.py Outdated Show resolved Hide resolved
mmseg/models/necks/jpu.py Outdated Show resolved Hide resolved
Copy link
Collaborator

@Junjun2016 Junjun2016 left a comment

Choose a reason for hiding this comment

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

Please rename all the config names accordingly.

Copy link
Collaborator

@Junjun2016 Junjun2016 left a comment

Choose a reason for hiding this comment

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

LGTM, please also resolve the comment.

configs/_base_/models/fastfcn_r50-d32.py Outdated Show resolved Hide resolved
mmseg/models/necks/jpu.py Show resolved Hide resolved
@xvjiarui xvjiarui removed the WIP Work in process label Sep 30, 2021
mmseg/models/necks/jpu.py Outdated Show resolved Hide resolved
mmseg/models/necks/jpu.py Outdated Show resolved Hide resolved
mmseg/models/necks/jpu.py Outdated Show resolved Hide resolved
Copy link
Collaborator

@xvjiarui xvjiarui left a comment

Choose a reason for hiding this comment

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

LGTM. Please resolve the conflict.

@Junjun2016 Junjun2016 merged commit f910caf into open-mmlab:master Sep 30, 2021
@MengzhangLI MengzhangLI deleted the FastFCN branch February 1, 2022 03:15
bowenroom pushed a commit to bowenroom/mmsegmentation that referenced this pull request Feb 25, 2022
* FastFCN first commit

* FastFCN first commit

* Fixing lint error

* Fixing lint error

* use for loop on JPU

* Use For Loop

* Refactor FastFCN

* FastFCN

* FastFCN

* temp

* Uploading models & logs (4x4)

* Fixing typos

* fix typos

* rename config

* change README.md

* use _delete_=True

* change configs

* change start_level to 0

* change start_level to 0

* jpu

* add unittest for start_level!=0
wjkim81 pushed a commit to wjkim81/mmsegmentation that referenced this pull request Dec 3, 2023
* add default test and train args

* remove partition field from default args to the parser

* fix typo in config

* fix typo in model list
sibozhang pushed a commit to sibozhang/mmsegmentation that referenced this pull request Mar 22, 2024
[fix] Correct the key name of  ‘eval_results’ dictionary for metric 'mmit_mean_average_precision'
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.

3 participants