[Bug] SWINIR Denoising train error: No such file or directory:'data/DFWB/meta_info_DFWB8550sub_GT.txt'

[fnwkqwan]

Prerequisite

• I have searched Issues and Discussions but cannot get the expected help.
• I have read the FAQ documentation but cannot get the expected help.
• The bug has not been fixed in the latest version (main) or latest version (0.x).

Task

I'm using the official example scripts/configs for the officially supported tasks/models/datasets.

Branch

main branch https://github.com/open-mmlab/mmagic

Environment

TorchVision:0.12.0+cu113
OpenCV:4.8.0
MMEngine:0.8.1
MMCV:2.0.0rc4
MMCV Compiler: MSVC 192829924
MMCV CUDA Compiler:11.3
MMagic: 1.0.1+b058b41

Reproduces the problem - code sample

python tools/train.py configs/swinir/swinir_s128w8d6e180_8xb1-lr2e-4-1600k_dfwb-grayDN15.py

Reproduces the problem - command or script

/

Reproduces the problem - error message

FileNotFoundError: No such file or directory:'data/DFWB/meta_info_DFWB8550sub_GT.txt'

Additional information

I have prepared datasets by tools/dataset_converters/denoising 'Preparing Denoising Dataset' as The folder structure looks like:

mmagic
├── mmagic
├── tools
├── configs
├── data
|   ├── denoising_gaussian_test
|   |   ├── Set12
|   |   ├── BSD68
|   |   ├── CBSD68
|   |   ├── Kodak
|   |   ├── McMaster
|   |   ├── Urban100
But there is no DFWB dataset mentioned in the configuration file, how can I get it?

[zengyh1900]

hi @fnwkqwan ,

DFWB denotes the combination of DIV2K (800 training images) + Flickr2K (2650 images) + WED(4744 images) + BSD500 (400 training&testing images).

These four datasets are usually combined and used as the training data of color/grayscale image denoising and greyscale/color JPEG compression artifact reduction. You can simply download and put them together and generate the annotation file so that the DataLoader can sample images from these four datasets during training.