[Fix] Fix correct `num_classes` of HRNet in `LoveDA` dataset #1136

MengzhangLI · 2021-12-14T09:56:17Z

In original settings, fcn_hr18_512x512_80k_loveda.py directly inherit num_classes=19 from '../_base_/models/fcn_hr18.py':

mmsegmentation/configs/_base_/models/fcn_hr18.py

Line 45 in bdfbaf0

num_classes=19,

However, the number of classes in LoveDA is 7, so it can run normally but it is not correct. Here is results before and after changing num_classes from 19 to 7:

Backbone	Mem (GB)	Inf time (fps)	Val mIoU	Val mIoU(ms+flip)	Test set mIoU on Codelab	download
HRNetV2p-W18-Small	~~1.72~~ 1.59	~~30.07~~ 24.87	~~49.3~~ 49.28	~~49.23~~ 49.42	~~47.57~~ 47.78	model \| log
HRNetV2p-W18	~~2.90~~ 2.76	~~16.77~~ 12.92	~~50.87~~ 50.81	~~51.24~~ 50.95	~~49.25~~ 48.83	model \| log
HRNetV2p-W48	~~6.25~~ 6.20	~~9.09~~ 9.61	~~51.04~~ 51.42	~~51.12~~ 51.64	~~50.38~~ 50.39	model \| log

codecov · 2021-12-14T10:21:51Z

Codecov Report

Merging #1136 (94159fa) into master (bdfbaf0) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1136   +/-   ##
=======================================
  Coverage   90.11%   90.11%           
=======================================
  Files         125      125           
  Lines        7262     7262           
  Branches     1206     1206           
=======================================
  Hits         6544     6544           
  Misses        515      515           
  Partials      203      203

Flag	Coverage Δ
unittests	`90.11% <ø> (ø)`

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

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 bdfbaf0...94159fa. Read the comment docs.

* segmenter: add model * update * readme: update * config: update * segmenter: update readme * segmenter: update * segmenter: update * segmenter: update * configs: set checkpoint path to pretrain folder * segmenter: modify vit-s/lin, remove data config * rreadme: update * configs: transfer from _base_ to segmenter * configs: add 8x1 suffix * configs: remove redundant lines * configs: cleanup * first attempt * swipe CI error * Update mmseg/models/decode_heads/__init__.py Co-authored-by: Junjun2016 <hejunjun@sjtu.edu.cn> * segmenter_linear: use fcn backbone * segmenter_mask: update * models: add segmenter vit * decoders: yapf+remove unused imports * apply precommit * segmenter/linear_head: fix * segmenter/linear_header: fix * segmenter: fix mask transformer * fix error * segmenter/mask_head: use trunc_normal init * refactor segmenter head * Fetch upstream (#1) * [Feature] Change options to cfg-option (#1129) * [Feature] Change option to cfg-option * add expire date and fix the docs * modify docstring * [Fix] Add  in metafile #1127 * [Fix] Fix correct num_classes of HRNet in LoveDA dataset #1136 * Bump to v0.20.1 (#1138) * bump version 0.20.1 * bump version 0.20.1 * [Fix] revise --option to --options #1140 Co-authored-by: Rockey <41846794+RockeyCoss@users.noreply.github.com> Co-authored-by: MengzhangLI <mcmong@pku.edu.cn> * decode_head: switch from linear to fcn * fix init list formatting * configs: remove variants, keep only vit-s on ade * align inference metric of vit-s-mask * configs: add vit t/b/l * Update mmseg/models/decode_heads/segmenter_mask_head.py Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com> * Update mmseg/models/decode_heads/segmenter_mask_head.py Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com> * Update mmseg/models/decode_heads/segmenter_mask_head.py Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com> * Update mmseg/models/decode_heads/segmenter_mask_head.py Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com> * Update mmseg/models/decode_heads/segmenter_mask_head.py Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com> * model_converters: use torch instead of einops * setup: remove einops * segmenter_mask: fix missing imports * add necessary imported init funtion * segmenter/seg-l: set resolution to 640 * segmenter/seg-l: fix test size * fix vitjax2mmseg * add README and unittest * fix unittest * add docstring * refactor config and add pretrained link * fix typo * add paper name in readme * change segmenter config names * fix typo in readme * fix typos in readme * fix segmenter typo * fix segmenter typo * delete redundant comma in config files * delete redundant comma in config files * fix convert script * update lateset master version Co-authored-by: MengzhangLI <mcmong@pku.edu.cn> Co-authored-by: Junjun2016 <hejunjun@sjtu.edu.cn> Co-authored-by: Rockey <41846794+RockeyCoss@users.noreply.github.com> Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com>

* segmenter: add model * update * readme: update * config: update * segmenter: update readme * segmenter: update * segmenter: update * segmenter: update * configs: set checkpoint path to pretrain folder * segmenter: modify vit-s/lin, remove data config * rreadme: update * configs: transfer from _base_ to segmenter * configs: add 8x1 suffix * configs: remove redundant lines * configs: cleanup * first attempt * swipe CI error * Update mmseg/models/decode_heads/__init__.py Co-authored-by: Junjun2016 <hejunjun@sjtu.edu.cn> * segmenter_linear: use fcn backbone * segmenter_mask: update * models: add segmenter vit * decoders: yapf+remove unused imports * apply precommit * segmenter/linear_head: fix * segmenter/linear_header: fix * segmenter: fix mask transformer * fix error * segmenter/mask_head: use trunc_normal init * refactor segmenter head * Fetch upstream (#1) * [Feature] Change options to cfg-option (open-mmlab#1129) * [Feature] Change option to cfg-option * add expire date and fix the docs * modify docstring * [Fix] Add  in metafile open-mmlab#1127 * [Fix] Fix correct num_classes of HRNet in LoveDA dataset open-mmlab#1136 * Bump to v0.20.1 (open-mmlab#1138) * bump version 0.20.1 * bump version 0.20.1 * [Fix] revise --option to --options open-mmlab#1140 Co-authored-by: Rockey <41846794+RockeyCoss@users.noreply.github.com> Co-authored-by: MengzhangLI <mcmong@pku.edu.cn> * decode_head: switch from linear to fcn * fix init list formatting * configs: remove variants, keep only vit-s on ade * align inference metric of vit-s-mask * configs: add vit t/b/l * Update mmseg/models/decode_heads/segmenter_mask_head.py Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com> * Update mmseg/models/decode_heads/segmenter_mask_head.py Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com> * Update mmseg/models/decode_heads/segmenter_mask_head.py Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com> * Update mmseg/models/decode_heads/segmenter_mask_head.py Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com> * Update mmseg/models/decode_heads/segmenter_mask_head.py Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com> * model_converters: use torch instead of einops * setup: remove einops * segmenter_mask: fix missing imports * add necessary imported init funtion * segmenter/seg-l: set resolution to 640 * segmenter/seg-l: fix test size * fix vitjax2mmseg * add README and unittest * fix unittest * add docstring * refactor config and add pretrained link * fix typo * add paper name in readme * change segmenter config names * fix typo in readme * fix typos in readme * fix segmenter typo * fix segmenter typo * delete redundant comma in config files * delete redundant comma in config files * fix convert script * update lateset master version Co-authored-by: MengzhangLI <mcmong@pku.edu.cn> Co-authored-by: Junjun2016 <hejunjun@sjtu.edu.cn> Co-authored-by: Rockey <41846794+RockeyCoss@users.noreply.github.com> Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com>

fix 404 link in README.mb

* fix yml * sort readme_files to avoid conflicts

fix loveda num_classes in LoveDA

94159fa

MengzhangLI self-assigned this Dec 14, 2021

MengzhangLI added the Merging PR waited for merging label Dec 14, 2021

Junjun2016 approved these changes Dec 14, 2021

View reviewed changes

Junjun2016 requested a review from RockeyCoss December 14, 2021 11:24

RockeyCoss approved these changes Dec 14, 2021

View reviewed changes

Junjun2016 merged commit e9c41c3 into open-mmlab:master Dec 14, 2021

MengzhangLI removed the Merging PR waited for merging label Dec 16, 2021

MengzhangLI deleted the fix_loveda_hrnet branch February 1, 2022 03:18

bowenroom pushed a commit to bowenroom/mmsegmentation that referenced this pull request Feb 25, 2022

[Fix] Fix correct num_classes of HRNet in LoveDA dataset open-mmlab#1136

f0d4700

aravind-h-v pushed a commit to aravind-h-v/mmsegmentation that referenced this pull request Mar 27, 2023

fix 404 link in example/README.mb (open-mmlab#1136)

af7b1c3

fix 404 link in README.mb

wjkim81 pushed a commit to wjkim81/mmsegmentation that referenced this pull request Dec 3, 2023

[Fix] fix pre-commit yml (open-mmlab#1136)

d5fd7c6

* fix yml * sort readme_files to avoid conflicts

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[Fix] Fix correct `num_classes` of HRNet in `LoveDA` dataset #1136

[Fix] Fix correct `num_classes` of HRNet in `LoveDA` dataset #1136

MengzhangLI commented Dec 14, 2021 •

edited

Loading

codecov bot commented Dec 14, 2021 •

edited

Loading

[Fix] Fix correct num_classes of HRNet in LoveDA dataset #1136

[Fix] Fix correct num_classes of HRNet in LoveDA dataset #1136

Conversation

MengzhangLI commented Dec 14, 2021 • edited Loading

codecov bot commented Dec 14, 2021 • edited Loading

Codecov Report

[Fix] Fix correct `num_classes` of HRNet in `LoveDA` dataset #1136

[Fix] Fix correct `num_classes` of HRNet in `LoveDA` dataset #1136

MengzhangLI commented Dec 14, 2021 •

edited

Loading

codecov bot commented Dec 14, 2021 •

edited

Loading