BasicSR

🚀 See you again 😸

✨ Highlights
✅ Add torch to setup_requires & dynamic import to prevent import error
✅ fix some codespell
✅ Add plot, you can find the scripts of drawing the paper figure. Welcome to contribute your figure scripts. Here is the example

🚀 See you again 😸

📢📢📢
✅ Add Chinses docs: https://github.com/XPixelGroup/BasicSR-docs. Download the PDF here: https://github.com/XPixelGroup/BasicSR-docs/releases
✅ BasicSR 的中文解读文档来啦: https://github.com/XPixelGroup/BasicSR-docs. 你可以从以下链接下载完整 PDF： https://github.com/XPixelGroup/BasicSR-docs/releases

✨ Highlights
✅ fix bgr2ycbcr bug in color conversion

🚀 Long time no see ☄️

This is the first time that BasicSR is released under the XPixelGroup :-)

✨ Highlights
✅ Add training codes of BasicVSR++
✅ Add inference code of BasicVSR++
✅ add psnr ssim pytorch version
✅ add LDL loss
✅ Add online document: basicsr.readthedocs.io/en/latest/
✅ update loss registry
✅ add realesrgan to basicsr
✅ support registry with basicsr suffix

📢📢📢
我们正在准备 BasicSR 的📕中文解读文档。将在近期 release :-)

Have a nice day 😸 and happy everyday 😃

I am happy to add a simple logo to BasicSR 😋 (designed by myself! and inspired by the 4K logo)
So I release a new version~

You can see it on ReadMe

(I know that there are still a lot of things to improve in BasicSR =-= I will put more time on it~ )

🚀 See you again ☄️

This is a minor update. It is mainly for the release of Real-ESRGAN 😄 , in which we release small models for anime videos.

✨✨✨ Highlight

• fix bug in AvgTimer: https://github.com/xinntao/BasicSR/commit/c05b52798fe29dbb45c839780c10e8a9cd2eebe5
• add pytest unittest framework

🚀 See you again ☄️

✨✨✨ Highlight

✅ [New method] Add ECBSR training and testing codes. (#478)

ACMMM21: Edge-oriented Convolution Block for Real-time Super Resolution on Mobile Devices
Xindong Zhang, Hui Zeng, Lei Zhang

✅ [Features] Support multiple validation datasets and also print the best metric value results.

Example:

✅ [Bug fix] metric_data is not initialized when no metric is used during validation. See Here

🚀 Long time no see 😹

✨✨✨ Highlight

✅ [Features] Support multiple inputs for metrics during validation (#467)
✅ [Bug fix] Fix bug in option: force_yml sometimes cannot be correctly modified
✅ [Bug fix] Fix bug in redsrecurrentdataset: support interval argument (#463)
✅ [Enhancement] add codespell hook, fix typos discovered by codespell

📢 📢 📢

建立了 BasicSR交流讨论的 QQ群和微信群：

技术交流QQ群：320960100 入群答案：互帮互助共同进步

📢 入群指南 (腾讯文档)

🧭 入群二维码

🚀 Have a nice day 🐶

This is a minor release~

✨ Highlight

✅ [Enhancement] Support the official torchvision.ops.deform_conv2d for torchvision>=0.9.0
✅ [Enhancement] Add force_yml option. You can force to change the option yml options in the command line. Examples: python basicsr/train.py -opt options/train/SRResNet_SRGAN/train_MSRResNet_x4.yml -train:ema_decay=0.999
✅ [Enhancement] Copy the option yml file to the experiment folder
✅ [Enhancement] Add AvgTimer. The data time and iteration time in the logging are averaged in 200 iterations.
✅ [Enhancement] Add persistent_workers option to dataloader
✅ [Enhancement] Add vscode settings.
✅ [Enhancement] Improve some formats

🚀 See you again 😸

✨ Highlight

• ✅ Add SwinIR training and testing codes
  • SwinIR by Jingyun Liang 👍 Thanks for their great work~
  • More details are in HOWTOs.md
• ✅ All models in BasicSR support EMA (Exponential Moving Average).

🚀

A lot of things have been improved since the last release note.

✨ Highlight

1. We add the official training and testing codes of BasicVSR - a video SR method
2. We improve NIQE metric. Now, the python version of NIQE could generate almost the same results as MATLAB.
3. Add HiFaceGAN codes by Lotayou.
4. Add more degradation
5. Add BasicSR-Examples, which provides guidance and templates of using BasicSR as a python package.
6. Fixed many bugs: such as no logging in pt18, resume bugs, etc
   ......

We will spend more time on solving issues and reviewing the pull request.
Welcome to your contributions 😄

🚀🚀🚀

🚀

✨ Highlight

A minor version for quick bug fixed.
#396
#397

🚀

✨ Highlights

We reorganize the BasicSR codes. It may be incompatible with the previous v1.2.0

• Add registry mechanism
• Support pip install
• Support JIT CUDA ops
• Can be easily used as an external package to develop your own project (example project is coming soon)
• Update format: change max line length to 120
• Add degradations (data utils)

This major change may introduce bugs. If you encounter bugs, please let me know. Thanks!

🚀

✨ Highlights

• Add ESRGAN and DFDNet colab
• Add FID and LPIPS metrics
• Add matlab imresize bicubic (#317)
• README add datasets download links (#318)

🐛 Bug Fixes

• PSNR and SSIM calculation on uint8 type
• Fix metrics bug in video_base_model.py (#314)

🌴 Improvements

• Reorganize code structure and remove unnessary packages
• tensor2img support gray images
• Refactor DFDNet codes

🚀

✨ Highlights

• Add Baidu Drive (百度网盘) download links
• Add funny emoji ☺️

🐛 Bug Fixes

• bgr2rgb type conversion in stylegan2 model
• Supporting training w/o validation

🌴 Improvements

• download_pretrained_models.py script supports downloading all the models
• Refactor define_network functions

Hope all is well 🚀

Highlights

• Add DFDNet inference codes (ECCV20: Blind Face Restoration via Deep Multi-scale Component Dictionaries)
• Add more official StyleGAN2 pretrained models: Model Zoo
• Add New Feature section in README.

Bug Fixes

• PyTorch 1.6 uses a new serialization for torch.save. The saved model cannot be loaded by the previous PyTorch version. We updated the publish_models.py with _use_new_zipfile_serialization=False. More details.

Hope all is well 🚀

Highlights

• Add StyleGAN2 training and testing codes. Pretrained models are here.
• Fix bug: cuda prefetcher return none twice.
• Add HOWTOs for quick starts.

We will use releases to manage BasicSR 😄

Hope all is well 🚀

This is a brand-new version of BasicSR. We have re-organized all the codes and frameworks.

Highlights

• We use Dynamic Instantiation for creating datasets, architectures, and models. So it is easier and more friendly to develop your own algorithms.
• We provide richer documents. At the same time, we also provide a Chinese version (同时也提供了中文版本的文档说明).
• We provide more pre-trained models, training examples. We also upload the training process and curves to wandb.
• Currently, it supports:
  • Training: EDSR, EDVR, ESRGAN, SRResNet, SRGAN
  • Testing: DUF, EDSR, EDVR, ESRGAN, RCAN, SRResNt, SRGAN, and TOF.
• We also mirror this codebase to Gitee码云 for easy access of Chinese users.

Sorry that this version of BasicSR is not compatible with the previous versions.

We will add more features to this codebase. And welcome contribute, and report bugs! 😆