Upload 112 files

.gitattributes

@@ -32,3 +32,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+figs/ETH_BSRGAN.png filter=lfs diff=lfs merge=lfs -text
+figs/ETH_LR.png filter=lfs diff=lfs merge=lfs -text
+figs/ETH_SwinIR-L.png filter=lfs diff=lfs merge=lfs -text
+figs/ETH_SwinIR.png filter=lfs diff=lfs merge=lfs -text
+figs/OST_009_crop_realESRGAN.png filter=lfs diff=lfs merge=lfs -text

LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [2021] [SwinIR Authors]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

README.md

@@ -1,3 +1,254 @@
+# SwinIR: Image Restoration Using Swin Transformer
+[Jingyun Liang](https://jingyunliang.github.io), [Jiezhang Cao](https://www.jiezhangcao.com/), [Guolei Sun](https://vision.ee.ethz.ch/people-details.MjYzMjMw.TGlzdC8zMjg5LC0xOTcxNDY1MTc4.html), [Kai Zhang](https://cszn.github.io/), [Luc Van Gool](https://scholar.google.com/citations?user=TwMib_QAAAAJ&hl=en), [Radu Timofte](http://people.ee.ethz.ch/~timofter/)
+
+Computer Vision Lab, ETH Zurich
+
 ---
-license: unlicense
+
+[![arXiv](https://img.shields.io/badge/arXiv-Paper-<COLOR>.svg)](https://arxiv.org/abs/2108.10257)
+[![GitHub Stars](https://img.shields.io/github/stars/JingyunLiang/SwinIR?style=social)](https://github.com/JingyunLiang/SwinIR)
+[![download](https://img.shields.io/github/downloads/JingyunLiang/SwinIR/total.svg)](https://github.com/JingyunLiang/SwinIR/releases)
+![visitors](https://visitor-badge.glitch.me/badge?page_id=jingyunliang/SwinIR)
+[ <a href="https://colab.research.google.com/gist/JingyunLiang/a5e3e54bc9ef8d7bf594f6fee8208533/swinir-demo-on-real-world-image-sr.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="google colab logo"></a>](https://colab.research.google.com/gist/JingyunLiang/a5e3e54bc9ef8d7bf594f6fee8208533/swinir-demo-on-real-world-image-sr.ipynb)
+<a href="https://replicate.ai/jingyunliang/swinir"><img src="https://img.shields.io/static/v1?label=Replicate&message=Demo and Docker Image&color=blue"></a>
+[![PlayTorch Demo](https://github.com/facebookresearch/playtorch/blob/main/website/static/assets/playtorch_badge.svg)](https://playtorch.dev/snack/@playtorch/swinir/)
+[Gradio Web Demo](https://huggingface.co/spaces/akhaliq/SwinIR)
+
+This repository is the official PyTorch implementation of SwinIR: Image Restoration Using Shifted Window Transformer
+([arxiv](https://arxiv.org/pdf/2108.10257.pdf), [supp](https://github.com/JingyunLiang/SwinIR/releases), [pretrained models](https://github.com/JingyunLiang/SwinIR/releases), [visual results](https://github.com/JingyunLiang/SwinIR/releases)). SwinIR achieves **state-of-the-art performance** in
+- bicubic/lighweight/real-world image SR
+- grayscale/color image denoising
+- grayscale/color JPEG compression artifact reduction
+
+</br>
+
+:rocket:  :rocket:  :rocket: **News**:
+- **Aug. 16, 2022**: Add PlayTorch Demo on running the real-world image SR model on mobile devices [![PlayTorch Demo](https://github.com/facebookresearch/playtorch/blob/main/website/static/assets/playtorch_badge.svg)](https://playtorch.dev/snack/@playtorch/swinir/).
+- **Aug. 01, 2022**: Add pretrained models and results on JPEG compression artifact reduction for color images. 
+- **Jun. 10, 2022**: See our work on video restoration :fire::fire::fire: [VRT: A Video Restoration Transformer](https://github.com/JingyunLiang/VRT) 
+[![GitHub Stars](https://img.shields.io/github/stars/JingyunLiang/VRT?style=social)](https://github.com/JingyunLiang/VRT)
+[![download](https://img.shields.io/github/downloads/JingyunLiang/VRT/total.svg)](https://github.com/JingyunLiang/VRT/releases)
+and [RVRT: Recurrent Video Restoration Transformer](https://github.com/JingyunLiang/RVRT) 
+[![GitHub Stars](https://img.shields.io/github/stars/JingyunLiang/RVRT?style=social)](https://github.com/JingyunLiang/RVRT)
+[![download](https://img.shields.io/github/downloads/JingyunLiang/RVRT/total.svg)](https://github.com/JingyunLiang/RVRT/releases)
+for video SR, video deblurring, video denoising, video frame interpolation and space-time video SR.
+- **Sep. 07, 2021**: We provide an interactive online Colab demo for real-world image SR <a href="https://colab.research.google.com/gist/JingyunLiang/a5e3e54bc9ef8d7bf594f6fee8208533/swinir-demo-on-real-world-image-sr.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="google colab logo"></a>:fire: for comparison with [the first practical degradation model BSRGAN (ICCV2021) ![GitHub Stars](https://img.shields.io/github/stars/cszn/BSRGAN?style=social)](https://github.com/cszn/BSRGAN) and a recent model RealESRGAN. Try to super-resolve your own images on Colab!
+
+|Real-World Image (x4)|[BSRGAN, ICCV2021](https://github.com/cszn/BSRGAN)|[Real-ESRGAN](https://github.com/xinntao/Real-ESRGAN)|SwinIR (ours)|SwinIR-Large (ours)|
+|       :---       |     :---:        |        :-----:         |        :-----:         |        :-----:         | 
+| <img width="200" src="figs/ETH_LR.png">|<img width="200" src="figs/ETH_BSRGAN.png">|<img width="200" src="figs/ETH_realESRGAN.jpg">|<img width="200" src="figs/ETH_SwinIR.png">|<img width="200" src="figs/ETH_SwinIR-L.png">
+|<img width="200" src="figs/OST_009_crop_LR.png">|<img width="200" src="figs/OST_009_crop_BSRGAN.png">|<img width="200" src="figs/OST_009_crop_realESRGAN.png">|<img width="200" src="figs/OST_009_crop_SwinIR.png">|<img width="200" src="figs/OST_009_crop_SwinIR-L.png">|
+  
+ - ***Aug. 26, 2021**: See our recent work on [real-world image SR: a pratical degrdation model BSRGAN, ICCV2021](https://github.com/cszn/BSRGAN)
+[![GitHub Stars](https://img.shields.io/github/stars/cszn/BSRGAN?style=social)](https://github.com/cszn/BSRGAN)*
+ - ***Aug. 26, 2021**: See our recent work on [generative modelling of image SR and image rescaling: normalizing-flow-based HCFlow, ICCV2021](https://github.com/JingyunLiang/HCFlow)
+[![GitHub Stars](https://img.shields.io/github/stars/JingyunLiang/HCFlow?style=social)](https://github.com/JingyunLiang/HCFlow)[ <a href="https://colab.research.google.com/gist/JingyunLiang/cdb3fef89ebd174eaa43794accb6f59d/hcflow-demo-on-x8-face-image-sr.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="google colab logo"></a>](https://colab.research.google.com/gist/JingyunLiang/cdb3fef89ebd174eaa43794accb6f59d/hcflow-demo-on-x8-face-image-sr.ipynb)*
+ - ***Aug. 26, 2021**: See our recent work on [blind SR: spatially variant kernel estimation (MANet, ICCV2021)](https://github.com/JingyunLiang/MANet) [![GitHub Stars](https://img.shields.io/github/stars/JingyunLiang/MANet?style=social)](https://github.com/JingyunLiang/MANet)
+[ <a href="https://colab.research.google.com/gist/JingyunLiang/4ed2524d6e08343710ee408a4d997e1c/manet-demo-on-spatially-variant-kernel-estimation.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="google colab logo"></a>](https://colab.research.google.com/gist/JingyunLiang/4ed2524d6e08343710ee408a4d997e1c/manet-demo-on-spatially-variant-kernel-estimation.ipynb) and [unsupervised kernel estimation (FKP, CVPR2021)](https://github.com/JingyunLiang/FKP)
+[![GitHub Stars](https://img.shields.io/github/stars/JingyunLiang/FKP?style=social)](https://github.com/JingyunLiang/FKP)*
+
 ---
+
+> Image restoration is a long-standing low-level vision problem that aims to restore high-quality images from low-quality images (e.g., downscaled, noisy and compressed images). While state-of-the-art image restoration methods are based on convolutional neural networks, few attempts have been made with Transformers which show impressive performance on high-level vision tasks. In this paper, we propose a strong baseline model SwinIR for image restoration based on the Swin Transformer. SwinIR consists of three parts: shallow feature extraction, deep feature extraction and high-quality image reconstruction. In particular, the deep feature extraction module is composed of several residual Swin Transformer blocks (RSTB), each of which has several Swin Transformer layers together with a residual connection. We conduct experiments on three representative tasks: image super-resolution (including classical, lightweight and real-world image super-resolution), image denoising (including grayscale and color image denoising) and JPEG compression artifact reduction. Experimental results demonstrate that SwinIR outperforms state-of-the-art methods on different tasks by up to 0.14~0.45dB, while the total number of parameters can be reduced by up to 67%.
+><p align="center">
+  <img width="800" src="figs/SwinIR_archi.png">
+</p>
+
+
+
+#### Contents
+
+1. [Training](#Training)
+1. [Testing](#Testing)
+1. [Results](#Results)
+1. [Citation](#Citation)
+1. [License and Acknowledgement](#License-and-Acknowledgement)
+
+
+### Training
+
+
+Used training and testing sets can be downloaded as follows:
+
+| Task                                                |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        Training Set                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | Testing Set|    Visual Results |    
+|:----------------------------------------------------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|     :---:      |   :---:      |
+| classical/lightweight image SR                      |                                                                                                                                                                                                                                                                                                                                                                                                                                                               [DIV2K](https://cv.snu.ac.kr/research/EDSR/DIV2K.tar) (800 training images) or DIV2K +[Flickr2K](https://cv.snu.ac.kr/research/EDSR/Flickr2K.tar) (2650 images)                                                                                                                                                                                                                                                                                                                                                                                                                                                               | Set5 + Set14 + BSD100 + Urban100 + Manga109 [download all](https://drive.google.com/drive/folders/1B3DJGQKB6eNdwuQIhdskA64qUuVKLZ9u) | [here](https://github.com/JingyunLiang/SwinIR/releases) |
+| real-world image SR                                 | SwinIR-M (middle size): [DIV2K](https://cv.snu.ac.kr/research/EDSR/DIV2K.tar) (800 training images) +[Flickr2K](https://cv.snu.ac.kr/research/EDSR/Flickr2K.tar) (2650 images) + [OST](https://openmmlab.oss-cn-hangzhou.aliyuncs.com/datasets/OST_dataset.zip) ([alternative link](https://drive.google.com/drive/folders/1iZfzAxAwOpeutz27HC56_y5RNqnsPPKr), 10324 images for sky,water,grass,mountain,building,plant,animal) <br /> SwinIR-L (large size): DIV2K + Flickr2K + OST + [WED](http://ivc.uwaterloo.ca/database/WaterlooExploration/exploration_database_and_code.rar)(4744 images) + [FFHQ](https://drive.google.com/drive/folders/1tZUcXDBeOibC6jcMCtgRRz67pzrAHeHL) (first 2000 images, face) + Manga109 (manga) + [SCUT-CTW1500](https://universityofadelaide.box.com/shared/static/py5uwlfyyytbb2pxzq9czvu6fuqbjdh8.zip) (first 100 training images, texts) <br /><br />  ***We use the pionnerring practical degradation model from [BSRGAN, ICCV2021  ![GitHub Stars](https://img.shields.io/github/stars/cszn/BSRGAN?style=social)](https://github.com/cszn/BSRGAN)** | [RealSRSet+5images](https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/RealSRSet+5images.zip) |  [here](https://github.com/JingyunLiang/SwinIR/releases) |
+| color/grayscale image denoising                     |                                                                                                                                                                                                                                                                                                             [DIV2K](https://cv.snu.ac.kr/research/EDSR/DIV2K.tar) (800 training images) + [Flickr2K](https://cv.snu.ac.kr/research/EDSR/Flickr2K.tar) (2650 images) + [BSD500](http://www.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/BSR/BSR_bsds500.tgz) (400 training&testing images) + [WED](http://ivc.uwaterloo.ca/database/WaterlooExploration/exploration_database_and_code.rar)(4744 images)  <br /><br />  *BSD68/BSD100 images are not used in training.                                                                                                                                                                                                                                                                                                              |  grayscale: Set12 + BSD68 + Urban100 <br />  color: CBSD68 + Kodak24 + McMaster + Urban100 [download all](https://github.com/cszn/FFDNet/tree/master/testsets) |  [here](https://github.com/JingyunLiang/SwinIR/releases) |
+| grayscale/color JPEG compression artifact reduction |                                                                                                                                                                                                                                                                                                                                            [DIV2K](https://cv.snu.ac.kr/research/EDSR/DIV2K.tar) (800 training images) + [Flickr2K](https://cv.snu.ac.kr/research/EDSR/Flickr2K.tar) (2650 images) + [BSD500](http://www.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/BSR/BSR_bsds500.tgz) (400 training&testing images) + [WED](http://ivc.uwaterloo.ca/database/WaterlooExploration/exploration_database_and_code.rar)(4744 images)                                                                                                                                                                                                                                                                                                                                             |  grayscale: Classic5 +LIVE1 [download all](https://github.com/cszn/DnCNN/tree/master/testsets) | [here](https://github.com/JingyunLiang/SwinIR/releases) |
+
+
+<!--
+| Task                 | Training Set | Testing Set|        Pretrained Model and Visual Results of SwinIR     | 
+| :---                 | :---:        |     :---:      |:---:      |
+| image denoising (real)      | [SIDD-Medium-sRGB](https://www.eecs.yorku.ca/~kamel/sidd/dataset.php) (320 images, [preprocess]()) + [RENOIR](http://ani.stat.fsu.edu/~abarbu/Renoir.html) (221 images, [preprocess](https://github.com/zsyOAOA/DANet/blob/master/datasets/preparedata/Renoir_big2small_all.py)) + [Poly](https://github.com/csjunxu/PolyU-Real-World-Noisy-Images-Dataset) (40 images in ./OriginalImages) |    [SIDD validation set](https://drive.google.com/drive/folders/1S44fHXaVxAYW3KLNxK41NYCnyX9S79su) (1280 patches, identical to official [.mat](https://www.eecs.yorku.ca/~kamel/sidd/benchmark.php) version) +  [DND](https://noise.visinf.tu-darmstadt.de/downloads/) (pre-defined 100 patches of 50 images, [online eval](https://noise.visinf.tu-darmstadt.de/submit/)) + [Nam](https://www.dropbox.com/s/24kds7c436i5i11/real_image_noise_dataset.zip?dl=0) (random 100 patches of 17 images, [preprocess](https://github.com/zsyOAOA/DANet/blob/master/datasets/preparedata/Nam_patch_prepare.py))|[download model]() [download results]() |
+| image deblurring (synthetic)   | [GoPro](https://drive.google.com/drive/folders/1AsgIP9_X0bg0olu2-1N6karm2x15cJWE) (2103 training images)  |  [GoPro](https://drive.google.com/drive/folders/1a2qKfXWpNuTGOm2-Jex8kfNSzYJLbqkf) (1111 images) + [HIDE](https://drive.google.com/drive/folders/1nRsTXj4iTUkTvBhTcGg8cySK8nd3vlhK) (2050 images) + [RealBlur_J](https://drive.google.com/drive/folders/1KYtzeKCiDRX9DSvC-upHrCqvC4sPAiJ1) (real blur, 980 images) + [RealBlur_R](https://drive.google.com/drive/folders/1EwDoajf5nStPIAcU4s9rdc8SPzfm3tW1) (real blur, 980 images) | [download model]() [download results]()|
+| image deraining (synthetic)  | [Multiple datasets](https://drive.google.com/drive/folders/1Hnnlc5kI0v9_BtfMytC2LR5VpLAFZtVe) (13711 training images, see Table 1 of [MPRNet](https://github.com/swz30/MPRNet) for details.)  |  Rain100H (100 images) + Rain100L (100 images) + Test100 (100 images) + Test2800 (2800 images) + Test1200 (1200 images), [download all](https://drive.google.com/drive/folders/1PDWggNh8ylevFmrjo-JEvlmqsDlWWvZs)  | [download model]() [download results]()|
+
+Note: above datasets may come from the official release or some awesome collections ([BasicSR](https://github.com/xinntao/BasicSR), [MPRNet](https://github.com/swz30/MPRNet)).
+
+-->
+
+The training code is at [KAIR](https://github.com/cszn/KAIR/blob/master/docs/README_SwinIR.md).
+
+## Testing (without preparing datasets)
+For your convience, we provide some example datasets (~20Mb) in `/testsets`. 
+If you just want codes, downloading `models/network_swinir.py`, `utils/util_calculate_psnr_ssim.py` and `main_test_swinir.py` is enough.
+Following commands will download [pretrained models](https://github.com/JingyunLiang/SwinIR/releases) **automatically** and put them in `model_zoo/swinir`. 
+**[All visual results of SwinIR can be downloaded here](https://github.com/JingyunLiang/SwinIR/releases)**. 
+
+We also provide an [online Colab demo for real-world image SR  <a href="https://colab.research.google.com/gist/JingyunLiang/a5e3e54bc9ef8d7bf594f6fee8208533/swinir-demo-on-real-world-image-sr.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="google colab logo"></a>](https://colab.research.google.com/gist/JingyunLiang/a5e3e54bc9ef8d7bf594f6fee8208533/swinir-demo-on-real-world-image-sr.ipynb) for comparison with [the first practical degradation model BSRGAN (ICCV2021)  ![GitHub Stars](https://img.shields.io/github/stars/cszn/BSRGAN?style=social)](https://github.com/cszn/BSRGAN) and a recent model [RealESRGAN](https://github.com/xinntao/Real-ESRGAN). Try to test your own images on Colab!
+
+We provide a PlayTorch demo [![PlayTorch Demo](https://github.com/facebookresearch/playtorch/blob/main/website/static/assets/playtorch_badge.svg)](https://playtorch.dev/snack/@playtorch/swinir/) for real-world image SR to showcase how to run the SwinIR model in mobile application built with React Native.
+
+```bash
+# 001 Classical Image Super-Resolution (middle size)
+# Note that --training_patch_size is just used to differentiate two different settings in Table 2 of the paper. Images are NOT tested patch by patch.
+# (setting1: when model is trained on DIV2K and with training_patch_size=48)
+python main_test_swinir.py --task classical_sr --scale 2 --training_patch_size 48 --model_path model_zoo/swinir/001_classicalSR_DIV2K_s48w8_SwinIR-M_x2.pth --folder_lq testsets/Set5/LR_bicubic/X2 --folder_gt testsets/Set5/HR
+python main_test_swinir.py --task classical_sr --scale 3 --training_patch_size 48 --model_path model_zoo/swinir/001_classicalSR_DIV2K_s48w8_SwinIR-M_x3.pth --folder_lq testsets/Set5/LR_bicubic/X3 --folder_gt testsets/Set5/HR
+python main_test_swinir.py --task classical_sr --scale 4 --training_patch_size 48 --model_path model_zoo/swinir/001_classicalSR_DIV2K_s48w8_SwinIR-M_x4.pth --folder_lq testsets/Set5/LR_bicubic/X4 --folder_gt testsets/Set5/HR
+python main_test_swinir.py --task classical_sr --scale 8 --training_patch_size 48 --model_path model_zoo/swinir/001_classicalSR_DIV2K_s48w8_SwinIR-M_x8.pth --folder_lq testsets/Set5/LR_bicubic/X8 --folder_gt testsets/Set5/HR
+
+# (setting2: when model is trained on DIV2K+Flickr2K and with training_patch_size=64)
+python main_test_swinir.py --task classical_sr --scale 2 --training_patch_size 64 --model_path model_zoo/swinir/001_classicalSR_DF2K_s64w8_SwinIR-M_x2.pth --folder_lq testsets/Set5/LR_bicubic/X2 --folder_gt testsets/Set5/HR
+python main_test_swinir.py --task classical_sr --scale 3 --training_patch_size 64 --model_path model_zoo/swinir/001_classicalSR_DF2K_s64w8_SwinIR-M_x3.pth --folder_lq testsets/Set5/LR_bicubic/X3 --folder_gt testsets/Set5/HR
+python main_test_swinir.py --task classical_sr --scale 4 --training_patch_size 64 --model_path model_zoo/swinir/001_classicalSR_DF2K_s64w8_SwinIR-M_x4.pth --folder_lq testsets/Set5/LR_bicubic/X4 --folder_gt testsets/Set5/HR
+python main_test_swinir.py --task classical_sr --scale 8 --training_patch_size 64 --model_path model_zoo/swinir/001_classicalSR_DF2K_s64w8_SwinIR-M_x8.pth --folder_lq testsets/Set5/LR_bicubic/X8 --folder_gt testsets/Set5/HR
+
+
+# 002 Lightweight Image Super-Resolution (small size)
+python main_test_swinir.py --task lightweight_sr --scale 2 --model_path model_zoo/swinir/002_lightweightSR_DIV2K_s64w8_SwinIR-S_x2.pth --folder_lq testsets/Set5/LR_bicubic/X2 --folder_gt testsets/Set5/HR
+python main_test_swinir.py --task lightweight_sr --scale 3 --model_path model_zoo/swinir/002_lightweightSR_DIV2K_s64w8_SwinIR-S_x3.pth --folder_lq testsets/Set5/LR_bicubic/X3 --folder_gt testsets/Set5/HR
+python main_test_swinir.py --task lightweight_sr --scale 4 --model_path model_zoo/swinir/002_lightweightSR_DIV2K_s64w8_SwinIR-S_x4.pth --folder_lq testsets/Set5/LR_bicubic/X4 --folder_gt testsets/Set5/HR
+
+
+# 003 Real-World Image Super-Resolution (use --tile 400 if you run out-of-memory)
+# (middle size)
+python main_test_swinir.py --task real_sr --scale 4 --model_path model_zoo/swinir/003_realSR_BSRGAN_DFO_s64w8_SwinIR-M_x4_GAN.pth --folder_lq testsets/RealSRSet+5images --tile
+
+# (larger size + trained on more datasets)
+python main_test_swinir.py --task real_sr --scale 4 --large_model --model_path model_zoo/swinir/003_realSR_BSRGAN_DFOWMFC_s64w8_SwinIR-L_x4_GAN.pth --folder_lq testsets/RealSRSet+5images
+
+
+# 004 Grayscale Image Deoising (middle size)
+python main_test_swinir.py --task gray_dn --noise 15 --model_path model_zoo/swinir/004_grayDN_DFWB_s128w8_SwinIR-M_noise15.pth --folder_gt testsets/Set12
+python main_test_swinir.py --task gray_dn --noise 25 --model_path model_zoo/swinir/004_grayDN_DFWB_s128w8_SwinIR-M_noise25.pth --folder_gt testsets/Set12
+python main_test_swinir.py --task gray_dn --noise 50 --model_path model_zoo/swinir/004_grayDN_DFWB_s128w8_SwinIR-M_noise50.pth --folder_gt testsets/Set12
+
+
+# 005 Color Image Deoising (middle size)
+python main_test_swinir.py --task color_dn --noise 15 --model_path model_zoo/swinir/005_colorDN_DFWB_s128w8_SwinIR-M_noise15.pth --folder_gt testsets/McMaster
+python main_test_swinir.py --task color_dn --noise 25 --model_path model_zoo/swinir/005_colorDN_DFWB_s128w8_SwinIR-M_noise25.pth --folder_gt testsets/McMaster
+python main_test_swinir.py --task color_dn --noise 50 --model_path model_zoo/swinir/005_colorDN_DFWB_s128w8_SwinIR-M_noise50.pth --folder_gt testsets/McMaster
+
+
+# 006 JPEG Compression Artifact Reduction (middle size, using window_size=7 because JPEG encoding uses 8x8 blocks)
+# grayscale
+python main_test_swinir.py --task jpeg_car --jpeg 10 --model_path model_zoo/swinir/006_CAR_DFWB_s126w7_SwinIR-M_jpeg10.pth --folder_gt testsets/classic5
+python main_test_swinir.py --task jpeg_car --jpeg 20 --model_path model_zoo/swinir/006_CAR_DFWB_s126w7_SwinIR-M_jpeg20.pth --folder_gt testsets/classic5
+python main_test_swinir.py --task jpeg_car --jpeg 30 --model_path model_zoo/swinir/006_CAR_DFWB_s126w7_SwinIR-M_jpeg30.pth --folder_gt testsets/classic5
+python main_test_swinir.py --task jpeg_car --jpeg 40 --model_path model_zoo/swinir/006_CAR_DFWB_s126w7_SwinIR-M_jpeg40.pth --folder_gt testsets/classic5
+
+# color
+python main_test_swinir.py --task color_jpeg_car --jpeg 10 --model_path model_zoo/swinir/006_colorCAR_DFWB_s126w7_SwinIR-M_jpeg10.pth --folder_gt testsets/LIVE1
+python main_test_swinir.py --task color_jpeg_car --jpeg 20 --model_path model_zoo/swinir/006_colorCAR_DFWB_s126w7_SwinIR-M_jpeg20.pth --folder_gt testsets/LIVE1
+python main_test_swinir.py --task color_jpeg_car --jpeg 30 --model_path model_zoo/swinir/006_colorCAR_DFWB_s126w7_SwinIR-M_jpeg30.pth --folder_gt testsets/LIVE1
+python main_test_swinir.py --task color_jpeg_car --jpeg 40 --model_path model_zoo/swinir/006_colorCAR_DFWB_s126w7_SwinIR-M_jpeg40.pth --folder_gt testsets/LIVE1
+
+```
+
+---
+
+## Results
+We achieved state-of-the-art performance on classical/lightweight/real-world image SR, grayscale/color image denoising and JPEG compression artifact reduction. Detailed results can be found in the [paper](https://arxiv.org/abs/2108.10257). All visual results of SwinIR can be downloaded [here](https://github.com/JingyunLiang/SwinIR/releases). 
+
+<details>
+<summary>Classical Image Super-Resolution (click me)</summary>
+<p align="center">
+  <img width="900" src="figs/classic_image_sr.png">
+  <img width="900" src="figs/classic_image_sr_visual.png">
+</p>
+  
+- More detailed comparison between SwinIR and a representative CNN-based model RCAN (classical image SR, X4)
+
+| Method             | Training Set    |  Training time  <br /> (8GeForceRTX2080Ti <br /> batch=32, iter=500k) |Y-PSNR/Y-SSIM <br /> on Manga109 | Run time  <br /> (1GeForceRTX2080Ti,<br /> on 256x256 LR image)* |  #Params   | #FLOPs |  Testing memory |
+| :---      | :---:        |        :-----:         |     :---:      |     :---:      |     :---:      |   :---:      |  :---:      |
+| RCAN | DIV2K | 1.6 days | 31.22/0.9173 | 0.180s | 15.6M | 850.6G | 593.1M | 
+| SwinIR | DIV2K | 1.8 days |31.67/0.9226 | 0.539s | 11.9M | 788.6G | 986.8M | 
+
+\* We re-test the runtime when the GPU is idle. We refer to the evluation code [here](https://github.com/cszn/KAIR/blob/master/main_challenge_sr.py).
+
+  
+- Results on DIV2K-validation (100 images)
+  
+|  Training Set | scale factor | PSNR (RGB) | PSNR (Y) | SSIM (RGB)  | SSIM (Y) |
+| :--- | :---: | :---:        |     :---:      | :---: | :---:        |
+|  DIV2K (800 images) | 2 | 35.25 | 36.77 | 0.9423 | 0.9500 |
+|  DIV2K+Flickr2K (2650 images) | 2 | 35.34 | 36.86 | 0.9430 |0.9507 |
+|  DIV2K (800 images) | 3 | 31.50 | 32.97 | 0.8832 |0.8965 |
+|  DIV2K+Flickr2K (2650 images) | 3 | 31.63 | 33.10 | 0.8854 |0.8985 |
+|  DIV2K (800 images) | 4 | 29.48 | 30.94 | 0.8311|0.8492 |
+|  DIV2K+Flickr2K (2650 images) | 4 | 29.63 | 31.08 | 0.8347|0.8523 |
+
+</details>
+
+<details>
+<summary>Lightweight Image Super-Resolution</summary>
+<p align="center">
+  <img width="900" src="figs/lightweight_image_sr.png">
+</p>
+</details>
+
+<details>
+<summary>Real-World Image Super-Resolution</summary>
+<p align="center">
+  <img width="900" src="figs/real_world_image_sr.png">
+</p>
+</details>
+
+<details>
+<summary>Grayscale Image Deoising</summary>
+<p align="center">
+  <img width="900" src="figs/gray_image_denoising.png">
+</p>
+</details>
+
+<details>
+<summary>Color Image Deoising</summary>
+<p align="center">
+  <img width="900" src="figs/color_image_denoising.png">
+</p>
+</details>
+
+<details>
+<summary>JPEG Compression Artifact Reduction</summary>
+
+on grayscale images
+<p align="center">
+  <img width="900" src="figs/jepg_compress_artfact_reduction.png">
+</p>
+
+on color images
+
+| Training Set | quality factor | PSNR (RGB) | PSNR-B (RGB) | SSIM (RGB) |
+|:-------------|:--------------:|:----------:|:------------:|:----------:|
+| LIVE1        |       10       |   28.06    |    27.76     |   0.8089   |
+| LIVE1        |       20       |   30.45    |    29.97     |   0.8741   |
+| LIVE1        |       30       |   31.82    |    31.24     |   0.9018   |
+| LIVE1        |       40       |   32.75    |    32.12     |   0.9174   |
+</details>
+
+
+
+## Citation
+    @article{liang2021swinir,
+      title={SwinIR: Image Restoration Using Swin Transformer},
+      author={Liang, Jingyun and Cao, Jiezhang and Sun, Guolei and Zhang, Kai and Van Gool, Luc and Timofte, Radu},
+      journal={arXiv preprint arXiv:2108.10257},
+      year={2021}
+    }
+
+
+## License and Acknowledgement
+This project is released under the Apache 2.0 license. The codes are based on [Swin Transformer](https://github.com/microsoft/Swin-Transformer) and [KAIR](https://github.com/cszn/KAIR). Please also follow their licenses. Thanks for their awesome works.

cog.yaml

@@ -0,0 +1,17 @@
+build:
+  gpu: true
+  python_version: "3.8"
+  system_packages:
+    - "libgl1-mesa-glx"
+    - "libglib2.0-0"
+  python_packages:
+    - "torchvision==0.9.0"
+    - "torch==1.8.0"
+    - "numpy==1.19.4"
+    - "opencv-python==4.4.0.46"
+    - "tqdm==4.62.2"
+    - "Pillow==8.3.2"
+    - "timm==0.4.12"
+    - "ipython==7.19.0"
+
+predict: "predict.py:Predictor"

download-weights.sh

@@ -0,0 +1,13 @@
+#!/bin/sh
+
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/003_realSR_BSRGAN_DFO_s64w8_SwinIR-M_x4_GAN.pth -P experiments/pretrained_models
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/004_grayDN_DFWB_s128w8_SwinIR-M_noise15.pth -P experiments/pretrained_models
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/004_grayDN_DFWB_s128w8_SwinIR-M_noise25.pth -P experiments/pretrained_models
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/004_grayDN_DFWB_s128w8_SwinIR-M_noise50.pth -P experiments/pretrained_models
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/005_colorDN_DFWB_s128w8_SwinIR-M_noise15.pth -P experiments/pretrained_models
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/005_colorDN_DFWB_s128w8_SwinIR-M_noise25.pth -P experiments/pretrained_models
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/005_colorDN_DFWB_s128w8_SwinIR-M_noise50.pth -P experiments/pretrained_models
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/006_CAR_DFWB_s126w7_SwinIR-M_jpeg10.pth -P experiments/pretrained_models
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/006_CAR_DFWB_s126w7_SwinIR-M_jpeg20.pth -P experiments/pretrained_models
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/006_CAR_DFWB_s126w7_SwinIR-M_jpeg30.pth -P experiments/pretrained_models
+wget https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/006_CAR_DFWB_s126w7_SwinIR-M_jpeg40.pth -P experiments/pretrained_models

main_test_swinir.py

@@ -0,0 +1,309 @@
+import argparse
+import cv2
+import glob
+import numpy as np
+from collections import OrderedDict
+import os
+import torch
+import requests
+
+from models.network_swinir import SwinIR as net
+from utils import util_calculate_psnr_ssim as util
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--task', type=str, default='color_dn', help='classical_sr, lightweight_sr, real_sr, '
+                                                                     'gray_dn, color_dn, jpeg_car, color_jpeg_car')
+    parser.add_argument('--scale', type=int, default=1, help='scale factor: 1, 2, 3, 4, 8') # 1 for dn and jpeg car
+    parser.add_argument('--noise', type=int, default=15, help='noise level: 15, 25, 50')
+    parser.add_argument('--jpeg', type=int, default=40, help='scale factor: 10, 20, 30, 40')
+    parser.add_argument('--training_patch_size', type=int, default=128, help='patch size used in training SwinIR. '
+                                       'Just used to differentiate two different settings in Table 2 of the paper. '
+                                       'Images are NOT tested patch by patch.')
+    parser.add_argument('--large_model', action='store_true', help='use large model, only provided for real image sr')
+    parser.add_argument('--model_path', type=str,
+                        default='model_zoo/swinir/001_classicalSR_DIV2K_s48w8_SwinIR-M_x2.pth')
+    parser.add_argument('--folder_lq', type=str, default=None, help='input low-quality test image folder')
+    parser.add_argument('--folder_gt', type=str, default=None, help='input ground-truth test image folder')
+    parser.add_argument('--tile', type=int, default=None, help='Tile size, None for no tile during testing (testing as a whole)')
+    parser.add_argument('--tile_overlap', type=int, default=32, help='Overlapping of different tiles')
+    args = parser.parse_args()
+
+    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+    # set up model
+    if os.path.exists(args.model_path):
+        print(f'loading model from {args.model_path}')
+    else:
+        os.makedirs(os.path.dirname(args.model_path), exist_ok=True)
+        url = 'https://github.com/JingyunLiang/SwinIR/releases/download/v0.0/{}'.format(os.path.basename(args.model_path))
+        r = requests.get(url, allow_redirects=True)
+        print(f'downloading model {args.model_path}')
+        open(args.model_path, 'wb').write(r.content)
+
+    model = define_model(args)
+    model.eval()
+    model = model.to(device)
+
+    # setup folder and path
+    folder, save_dir, border, window_size = setup(args)
+    os.makedirs(save_dir, exist_ok=True)
+    test_results = OrderedDict()
+    test_results['psnr'] = []
+    test_results['ssim'] = []
+    test_results['psnr_y'] = []
+    test_results['ssim_y'] = []
+    test_results['psnrb'] = []
+    test_results['psnrb_y'] = []
+    psnr, ssim, psnr_y, ssim_y, psnrb, psnrb_y = 0, 0, 0, 0, 0, 0
+
+    for idx, path in enumerate(sorted(glob.glob(os.path.join(folder, '*')))):
+        # read image
+        imgname, img_lq, img_gt = get_image_pair(args, path)  # image to HWC-BGR, float32
+        img_lq = np.transpose(img_lq if img_lq.shape[2] == 1 else img_lq[:, :, [2, 1, 0]], (2, 0, 1))  # HCW-BGR to CHW-RGB
+        img_lq = torch.from_numpy(img_lq).float().unsqueeze(0).to(device)  # CHW-RGB to NCHW-RGB
+
+        # inference
+        with torch.no_grad():
+            # pad input image to be a multiple of window_size
+            _, _, h_old, w_old = img_lq.size()
+            h_pad = (h_old // window_size + 1) * window_size - h_old
+            w_pad = (w_old // window_size + 1) * window_size - w_old
+            img_lq = torch.cat([img_lq, torch.flip(img_lq, [2])], 2)[:, :, :h_old + h_pad, :]
+            img_lq = torch.cat([img_lq, torch.flip(img_lq, [3])], 3)[:, :, :, :w_old + w_pad]
+            output = test(img_lq, model, args, window_size)
+            output = output[..., :h_old * args.scale, :w_old * args.scale]
+
+        # save image
+        output = output.data.squeeze().float().cpu().clamp_(0, 1).numpy()
+        if output.ndim == 3:
+            output = np.transpose(output[[2, 1, 0], :, :], (1, 2, 0))  # CHW-RGB to HCW-BGR
+        output = (output * 255.0).round().astype(np.uint8)  # float32 to uint8
+        cv2.imwrite(f'{save_dir}/{imgname}_SwinIR.png', output)
+
+        # evaluate psnr/ssim/psnr_b
+        if img_gt is not None:
+            img_gt = (img_gt * 255.0).round().astype(np.uint8)  # float32 to uint8
+            img_gt = img_gt[:h_old * args.scale, :w_old * args.scale, ...]  # crop gt
+            img_gt = np.squeeze(img_gt)
+
+            psnr = util.calculate_psnr(output, img_gt, crop_border=border)
+            ssim = util.calculate_ssim(output, img_gt, crop_border=border)
+            test_results['psnr'].append(psnr)
+            test_results['ssim'].append(ssim)
+            if img_gt.ndim == 3:  # RGB image
+                psnr_y = util.calculate_psnr(output, img_gt, crop_border=border, test_y_channel=True)
+                ssim_y = util.calculate_ssim(output, img_gt, crop_border=border, test_y_channel=True)
+                test_results['psnr_y'].append(psnr_y)
+                test_results['ssim_y'].append(ssim_y)
+            if args.task in ['jpeg_car', 'color_jpeg_car']:
+                psnrb = util.calculate_psnrb(output, img_gt, crop_border=border, test_y_channel=False)
+                test_results['psnrb'].append(psnrb)
+                if args.task in ['color_jpeg_car']:
+                    psnrb_y = util.calculate_psnrb(output, img_gt, crop_border=border, test_y_channel=True)
+                    test_results['psnrb_y'].append(psnrb_y)
+            print('Testing {:d} {:20s} - PSNR: {:.2f} dB; SSIM: {:.4f}; PSNRB: {:.2f} dB;'
+                  'PSNR_Y: {:.2f} dB; SSIM_Y: {:.4f}; PSNRB_Y: {:.2f} dB.'.
+                  format(idx, imgname, psnr, ssim, psnrb, psnr_y, ssim_y, psnrb_y))
+        else:
+            print('Testing {:d} {:20s}'.format(idx, imgname))
+
+    # summarize psnr/ssim
+    if img_gt is not None:
+        ave_psnr = sum(test_results['psnr']) / len(test_results['psnr'])
+        ave_ssim = sum(test_results['ssim']) / len(test_results['ssim'])
+        print('\n{} \n-- Average PSNR/SSIM(RGB): {:.2f} dB; {:.4f}'.format(save_dir, ave_psnr, ave_ssim))
+        if img_gt.ndim == 3:
+            ave_psnr_y = sum(test_results['psnr_y']) / len(test_results['psnr_y'])
+            ave_ssim_y = sum(test_results['ssim_y']) / len(test_results['ssim_y'])
+            print('-- Average PSNR_Y/SSIM_Y: {:.2f} dB; {:.4f}'.format(ave_psnr_y, ave_ssim_y))
+        if args.task in ['jpeg_car', 'color_jpeg_car']:
+            ave_psnrb = sum(test_results['psnrb']) / len(test_results['psnrb'])
+            print('-- Average PSNRB: {:.2f} dB'.format(ave_psnrb))
+            if args.task in ['color_jpeg_car']:
+                ave_psnrb_y = sum(test_results['psnrb_y']) / len(test_results['psnrb_y'])
+                print('-- Average PSNRB_Y: {:.2f} dB'.format(ave_psnrb_y))
+
+
+def define_model(args):
+    # 001 classical image sr
+    if args.task == 'classical_sr':
+        model = net(upscale=args.scale, in_chans=3, img_size=args.training_patch_size, window_size=8,
+                    img_range=1., depths=[6, 6, 6, 6, 6, 6], embed_dim=180, num_heads=[6, 6, 6, 6, 6, 6],
+                    mlp_ratio=2, upsampler='pixelshuffle', resi_connection='1conv')
+        param_key_g = 'params'
+
+    # 002 lightweight image sr
+    # use 'pixelshuffledirect' to save parameters
+    elif args.task == 'lightweight_sr':
+        model = net(upscale=args.scale, in_chans=3, img_size=64, window_size=8,
+                    img_range=1., depths=[6, 6, 6, 6], embed_dim=60, num_heads=[6, 6, 6, 6],
+                    mlp_ratio=2, upsampler='pixelshuffledirect', resi_connection='1conv')
+        param_key_g = 'params'
+
+    # 003 real-world image sr
+    elif args.task == 'real_sr':
+        if not args.large_model:
+            # use 'nearest+conv' to avoid block artifacts
+            model = net(upscale=args.scale, in_chans=3, img_size=64, window_size=8,
+                        img_range=1., depths=[6, 6, 6, 6, 6, 6], embed_dim=180, num_heads=[6, 6, 6, 6, 6, 6],
+                        mlp_ratio=2, upsampler='nearest+conv', resi_connection='1conv')
+        else:
+            # larger model size; use '3conv' to save parameters and memory; use ema for GAN training
+            model = net(upscale=args.scale, in_chans=3, img_size=64, window_size=8,
+                        img_range=1., depths=[6, 6, 6, 6, 6, 6, 6, 6, 6], embed_dim=240,
+                        num_heads=[8, 8, 8, 8, 8, 8, 8, 8, 8],
+                        mlp_ratio=2, upsampler='nearest+conv', resi_connection='3conv')
+        param_key_g = 'params_ema'
+
+    # 004 grayscale image denoising
+    elif args.task == 'gray_dn':
+        model = net(upscale=1, in_chans=1, img_size=128, window_size=8,
+                    img_range=1., depths=[6, 6, 6, 6, 6, 6], embed_dim=180, num_heads=[6, 6, 6, 6, 6, 6],
+                    mlp_ratio=2, upsampler='', resi_connection='1conv')
+        param_key_g = 'params'
+
+    # 005 color image denoising
+    elif args.task == 'color_dn':
+        model = net(upscale=1, in_chans=3, img_size=128, window_size=8,
+                    img_range=1., depths=[6, 6, 6, 6, 6, 6], embed_dim=180, num_heads=[6, 6, 6, 6, 6, 6],
+                    mlp_ratio=2, upsampler='', resi_connection='1conv')
+        param_key_g = 'params'
+
+    # 006 grayscale JPEG compression artifact reduction
+    # use window_size=7 because JPEG encoding uses 8x8; use img_range=255 because it's sligtly better than 1
+    elif args.task == 'jpeg_car':
+        model = net(upscale=1, in_chans=1, img_size=126, window_size=7,
+                    img_range=255., depths=[6, 6, 6, 6, 6, 6], embed_dim=180, num_heads=[6, 6, 6, 6, 6, 6],
+                    mlp_ratio=2, upsampler='', resi_connection='1conv')
+        param_key_g = 'params'
+
+    # 006 color JPEG compression artifact reduction
+    # use window_size=7 because JPEG encoding uses 8x8; use img_range=255 because it's sligtly better than 1
+    elif args.task == 'color_jpeg_car':
+        model = net(upscale=1, in_chans=3, img_size=126, window_size=7,
+                    img_range=255., depths=[6, 6, 6, 6, 6, 6], embed_dim=180, num_heads=[6, 6, 6, 6, 6, 6],
+                    mlp_ratio=2, upsampler='', resi_connection='1conv')
+        param_key_g = 'params'
+
+    pretrained_model = torch.load(args.model_path)
+    model.load_state_dict(pretrained_model[param_key_g] if param_key_g in pretrained_model.keys() else pretrained_model, strict=True)
+
+    return model
+
+
+def setup(args):
+    # 001 classical image sr/ 002 lightweight image sr
+    if args.task in ['classical_sr', 'lightweight_sr']:
+        save_dir = f'results/swinir_{args.task}_x{args.scale}'
+        folder = args.folder_gt
+        border = args.scale
+        window_size = 8
+
+    # 003 real-world image sr
+    elif args.task in ['real_sr']:
+        save_dir = f'results/swinir_{args.task}_x{args.scale}'
+        if args.large_model:
+            save_dir += '_large'
+        folder = args.folder_lq
+        border = 0
+        window_size = 8
+
+    # 004 grayscale image denoising/ 005 color image denoising
+    elif args.task in ['gray_dn', 'color_dn']:
+        save_dir = f'results/swinir_{args.task}_noise{args.noise}'
+        folder = args.folder_gt
+        border = 0
+        window_size = 8
+
+    # 006 JPEG compression artifact reduction
+    elif args.task in ['jpeg_car', 'color_jpeg_car']:
+        save_dir = f'results/swinir_{args.task}_jpeg{args.jpeg}'
+        folder = args.folder_gt
+        border = 0
+        window_size = 7
+
+    return folder, save_dir, border, window_size
+
+
+def get_image_pair(args, path):
+    (imgname, imgext) = os.path.splitext(os.path.basename(path))
+
+    # 001 classical image sr/ 002 lightweight image sr (load lq-gt image pairs)
+    if args.task in ['classical_sr', 'lightweight_sr']:
+        img_gt = cv2.imread(path, cv2.IMREAD_COLOR).astype(np.float32) / 255.
+        img_lq = cv2.imread(f'{args.folder_lq}/{imgname}x{args.scale}{imgext}', cv2.IMREAD_COLOR).astype(
+            np.float32) / 255.
+
+    # 003 real-world image sr (load lq image only)
+    elif args.task in ['real_sr']:
+        img_gt = None
+        img_lq = cv2.imread(path, cv2.IMREAD_COLOR).astype(np.float32) / 255.
+
+    # 004 grayscale image denoising (load gt image and generate lq image on-the-fly)
+    elif args.task in ['gray_dn']:
+        img_gt = cv2.imread(path, cv2.IMREAD_GRAYSCALE).astype(np.float32) / 255.
+        np.random.seed(seed=0)
+        img_lq = img_gt + np.random.normal(0, args.noise / 255., img_gt.shape)
+        img_gt = np.expand_dims(img_gt, axis=2)
+        img_lq = np.expand_dims(img_lq, axis=2)
+
+    # 005 color image denoising (load gt image and generate lq image on-the-fly)
+    elif args.task in ['color_dn']:
+        img_gt = cv2.imread(path, cv2.IMREAD_COLOR).astype(np.float32) / 255.
+        np.random.seed(seed=0)
+        img_lq = img_gt + np.random.normal(0, args.noise / 255., img_gt.shape)
+
+    # 006 grayscale JPEG compression artifact reduction (load gt image and generate lq image on-the-fly)
+    elif args.task in ['jpeg_car']:
+        img_gt = cv2.imread(path, cv2.IMREAD_UNCHANGED)
+        if img_gt.ndim != 2:
+            img_gt = util.bgr2ycbcr(img_gt, y_only=True)
+        result, encimg = cv2.imencode('.jpg', img_gt, [int(cv2.IMWRITE_JPEG_QUALITY), args.jpeg])
+        img_lq = cv2.imdecode(encimg, 0)
+        img_gt = np.expand_dims(img_gt, axis=2).astype(np.float32) / 255.
+        img_lq = np.expand_dims(img_lq, axis=2).astype(np.float32) / 255.
+
+    # 006 JPEG compression artifact reduction (load gt image and generate lq image on-the-fly)
+    elif args.task in ['color_jpeg_car']:
+        img_gt = cv2.imread(path)
+        result, encimg = cv2.imencode('.jpg', img_gt, [int(cv2.IMWRITE_JPEG_QUALITY), args.jpeg])
+        img_lq = cv2.imdecode(encimg, 1)
+        img_gt = img_gt.astype(np.float32)/ 255.
+        img_lq = img_lq.astype(np.float32)/ 255.
+
+    return imgname, img_lq, img_gt
+
+
+def test(img_lq, model, args, window_size):
+    if args.tile is None:
+        # test the image as a whole
+        output = model(img_lq)
+    else:
+        # test the image tile by tile
+        b, c, h, w = img_lq.size()
+        tile = min(args.tile, h, w)
+        assert tile % window_size == 0, "tile size should be a multiple of window_size"
+        tile_overlap = args.tile_overlap
+        sf = args.scale
+
+        stride = tile - tile_overlap
+        h_idx_list = list(range(0, h-tile, stride)) + [h-tile]
+        w_idx_list = list(range(0, w-tile, stride)) + [w-tile]
+        E = torch.zeros(b, c, h*sf, w*sf).type_as(img_lq)
+        W = torch.zeros_like(E)
+
+        for h_idx in h_idx_list:
+            for w_idx in w_idx_list:
+                in_patch = img_lq[..., h_idx:h_idx+tile, w_idx:w_idx+tile]
+                out_patch = model(in_patch)
+                out_patch_mask = torch.ones_like(out_patch)
+
+                E[..., h_idx*sf:(h_idx+tile)*sf, w_idx*sf:(w_idx+tile)*sf].add_(out_patch)
+                W[..., h_idx*sf:(h_idx+tile)*sf, w_idx*sf:(w_idx+tile)*sf].add_(out_patch_mask)
+        output = E.div_(W)
+
+    return output
+
+if __name__ == '__main__':
+    main()

model_zoo/README.md

@@ -0,0 +1,3 @@
+model_zoo
+
+The SwinIR models are available at [here](https://github.com/JingyunLiang/SwinIR/releases/tag/v0.0).

models/network_swinir.py

@@ -0,0 +1,867 @@
+# -----------------------------------------------------------------------------------
+# SwinIR: Image Restoration Using Swin Transformer, https://arxiv.org/abs/2108.10257
+# Originally Written by Ze Liu, Modified by Jingyun Liang.
+# -----------------------------------------------------------------------------------
+
+import math
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.utils.checkpoint as checkpoint
+from timm.models.layers import DropPath, to_2tuple, trunc_normal_
+
+
+class Mlp(nn.Module):
+    def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
+        super().__init__()
+        out_features = out_features or in_features
+        hidden_features = hidden_features or in_features
+        self.fc1 = nn.Linear(in_features, hidden_features)
+        self.act = act_layer()
+        self.fc2 = nn.Linear(hidden_features, out_features)
+        self.drop = nn.Dropout(drop)
+
+    def forward(self, x):
+        x = self.fc1(x)
+        x = self.act(x)
+        x = self.drop(x)
+        x = self.fc2(x)
+        x = self.drop(x)
+        return x
+
+
+def window_partition(x, window_size):
+    """
+    Args:
+        x: (B, H, W, C)
+        window_size (int): window size
+
+    Returns:
+        windows: (num_windows*B, window_size, window_size, C)
+    """
+    B, H, W, C = x.shape
+    x = x.view(B, H // window_size, window_size, W // window_size, window_size, C)
+    windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
+    return windows
+
+
+def window_reverse(windows, window_size, H, W):
+    """
+    Args:
+        windows: (num_windows*B, window_size, window_size, C)
+        window_size (int): Window size
+        H (int): Height of image
+        W (int): Width of image
+
+    Returns:
+        x: (B, H, W, C)
+    """
+    B = int(windows.shape[0] / (H * W / window_size / window_size))
+    x = windows.view(B, H // window_size, W // window_size, window_size, window_size, -1)
+    x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, H, W, -1)
+    return x
+
+
+class WindowAttention(nn.Module):
+    r""" Window based multi-head self attention (W-MSA) module with relative position bias.
+    It supports both of shifted and non-shifted window.
+
+    Args:
+        dim (int): Number of input channels.
+        window_size (tuple[int]): The height and width of the window.
+        num_heads (int): Number of attention heads.
+        qkv_bias (bool, optional):  If True, add a learnable bias to query, key, value. Default: True
+        qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set
+        attn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0
+        proj_drop (float, optional): Dropout ratio of output. Default: 0.0
+    """
+
+    def __init__(self, dim, window_size, num_heads, qkv_bias=True, qk_scale=None, attn_drop=0., proj_drop=0.):
+
+        super().__init__()
+        self.dim = dim
+        self.window_size = window_size  # Wh, Ww
+        self.num_heads = num_heads
+        head_dim = dim // num_heads
+        self.scale = qk_scale or head_dim ** -0.5
+
+        # define a parameter table of relative position bias
+        self.relative_position_bias_table = nn.Parameter(
+            torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads))  # 2*Wh-1 * 2*Ww-1, nH
+
+        # get pair-wise relative position index for each token inside the window
+        coords_h = torch.arange(self.window_size[0])
+        coords_w = torch.arange(self.window_size[1])
+        coords = torch.stack(torch.meshgrid([coords_h, coords_w]))  # 2, Wh, Ww
+        coords_flatten = torch.flatten(coords, 1)  # 2, Wh*Ww
+        relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]  # 2, Wh*Ww, Wh*Ww
+        relative_coords = relative_coords.permute(1, 2, 0).contiguous()  # Wh*Ww, Wh*Ww, 2
+        relative_coords[:, :, 0] += self.window_size[0] - 1  # shift to start from 0
+        relative_coords[:, :, 1] += self.window_size[1] - 1
+        relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1
+        relative_position_index = relative_coords.sum(-1)  # Wh*Ww, Wh*Ww
+        self.register_buffer("relative_position_index", relative_position_index)
+
+        self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
+        self.attn_drop = nn.Dropout(attn_drop)
+        self.proj = nn.Linear(dim, dim)
+
+        self.proj_drop = nn.Dropout(proj_drop)
+
+        trunc_normal_(self.relative_position_bias_table, std=.02)
+        self.softmax = nn.Softmax(dim=-1)
+
+    def forward(self, x, mask=None):
+        """
+        Args:
+            x: input features with shape of (num_windows*B, N, C)
+            mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
+        """
+        B_, N, C = x.shape
+        qkv = self.qkv(x).reshape(B_, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
+        q, k, v = qkv[0], qkv[1], qkv[2]  # make torchscript happy (cannot use tensor as tuple)
+
+        q = q * self.scale
+        attn = (q @ k.transpose(-2, -1))
+
+        relative_position_bias = self.relative_position_bias_table[self.relative_position_index.view(-1)].view(
+            self.window_size[0] * self.window_size[1], self.window_size[0] * self.window_size[1], -1)  # Wh*Ww,Wh*Ww,nH
+        relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()  # nH, Wh*Ww, Wh*Ww
+        attn = attn + relative_position_bias.unsqueeze(0)
+
+        if mask is not None:
+            nW = mask.shape[0]
+            attn = attn.view(B_ // nW, nW, self.num_heads, N, N) + mask.unsqueeze(1).unsqueeze(0)
+            attn = attn.view(-1, self.num_heads, N, N)
+            attn = self.softmax(attn)
+        else:
+            attn = self.softmax(attn)
+
+        attn = self.attn_drop(attn)
+
+        x = (attn @ v).transpose(1, 2).reshape(B_, N, C)
+        x = self.proj(x)
+        x = self.proj_drop(x)
+        return x
+
+    def extra_repr(self) -> str:
+        return f'dim={self.dim}, window_size={self.window_size}, num_heads={self.num_heads}'
+
+    def flops(self, N):
+        # calculate flops for 1 window with token length of N
+        flops = 0
+        # qkv = self.qkv(x)
+        flops += N * self.dim * 3 * self.dim
+        # attn = (q @ k.transpose(-2, -1))
+        flops += self.num_heads * N * (self.dim // self.num_heads) * N
+        #  x = (attn @ v)
+        flops += self.num_heads * N * N * (self.dim // self.num_heads)
+        # x = self.proj(x)
+        flops += N * self.dim * self.dim
+        return flops
+
+
+class SwinTransformerBlock(nn.Module):
+    r""" Swin Transformer Block.
+
+    Args:
+        dim (int): Number of input channels.
+        input_resolution (tuple[int]): Input resulotion.
+        num_heads (int): Number of attention heads.
+        window_size (int): Window size.
+        shift_size (int): Shift size for SW-MSA.
+        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+        qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
+        qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
+        drop (float, optional): Dropout rate. Default: 0.0
+        attn_drop (float, optional): Attention dropout rate. Default: 0.0
+        drop_path (float, optional): Stochastic depth rate. Default: 0.0
+        act_layer (nn.Module, optional): Activation layer. Default: nn.GELU
+        norm_layer (nn.Module, optional): Normalization layer.  Default: nn.LayerNorm
+    """
+
+    def __init__(self, dim, input_resolution, num_heads, window_size=7, shift_size=0,
+                 mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0., drop_path=0.,
+                 act_layer=nn.GELU, norm_layer=nn.LayerNorm):
+        super().__init__()
+        self.dim = dim
+        self.input_resolution = input_resolution
+        self.num_heads = num_heads
+        self.window_size = window_size
+        self.shift_size = shift_size
+        self.mlp_ratio = mlp_ratio
+        if min(self.input_resolution) <= self.window_size:
+            # if window size is larger than input resolution, we don't partition windows
+            self.shift_size = 0
+            self.window_size = min(self.input_resolution)
+        assert 0 <= self.shift_size < self.window_size, "shift_size must in 0-window_size"
+
+        self.norm1 = norm_layer(dim)
+        self.attn = WindowAttention(
+            dim, window_size=to_2tuple(self.window_size), num_heads=num_heads,
+            qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop)
+
+        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
+        self.norm2 = norm_layer(dim)
+        mlp_hidden_dim = int(dim * mlp_ratio)
+        self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)
+
+        if self.shift_size > 0:
+            attn_mask = self.calculate_mask(self.input_resolution)
+        else:
+            attn_mask = None
+
+        self.register_buffer("attn_mask", attn_mask)
+
+    def calculate_mask(self, x_size):
+        # calculate attention mask for SW-MSA
+        H, W = x_size
+        img_mask = torch.zeros((1, H, W, 1))  # 1 H W 1
+        h_slices = (slice(0, -self.window_size),
+                    slice(-self.window_size, -self.shift_size),
+                    slice(-self.shift_size, None))
+        w_slices = (slice(0, -self.window_size),
+                    slice(-self.window_size, -self.shift_size),
+                    slice(-self.shift_size, None))
+        cnt = 0
+        for h in h_slices:
+            for w in w_slices:
+                img_mask[:, h, w, :] = cnt
+                cnt += 1
+
+        mask_windows = window_partition(img_mask, self.window_size)  # nW, window_size, window_size, 1
+        mask_windows = mask_windows.view(-1, self.window_size * self.window_size)
+        attn_mask = mask_windows.unsqueeze(1) - mask_windows.unsqueeze(2)
+        attn_mask = attn_mask.masked_fill(attn_mask != 0, float(-100.0)).masked_fill(attn_mask == 0, float(0.0))
+
+        return attn_mask
+
+    def forward(self, x, x_size):
+        H, W = x_size
+        B, L, C = x.shape
+        # assert L == H * W, "input feature has wrong size"
+
+        shortcut = x
+        x = self.norm1(x)
+        x = x.view(B, H, W, C)
+
+        # cyclic shift
+        if self.shift_size > 0:
+            shifted_x = torch.roll(x, shifts=(-self.shift_size, -self.shift_size), dims=(1, 2))
+        else:
+            shifted_x = x
+
+        # partition windows
+        x_windows = window_partition(shifted_x, self.window_size)  # nW*B, window_size, window_size, C
+        x_windows = x_windows.view(-1, self.window_size * self.window_size, C)  # nW*B, window_size*window_size, C
+
+        # W-MSA/SW-MSA (to be compatible for testing on images whose shapes are the multiple of window size
+        if self.input_resolution == x_size:
+            attn_windows = self.attn(x_windows, mask=self.attn_mask)  # nW*B, window_size*window_size, C
+        else:
+            attn_windows = self.attn(x_windows, mask=self.calculate_mask(x_size).to(x.device))
+
+        # merge windows
+        attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C)
+        shifted_x = window_reverse(attn_windows, self.window_size, H, W)  # B H' W' C
+
+        # reverse cyclic shift
+        if self.shift_size > 0:
+            x = torch.roll(shifted_x, shifts=(self.shift_size, self.shift_size), dims=(1, 2))
+        else:
+            x = shifted_x
+        x = x.view(B, H * W, C)
+
+        # FFN
+        x = shortcut + self.drop_path(x)
+        x = x + self.drop_path(self.mlp(self.norm2(x)))
+
+        return x
+
+    def extra_repr(self) -> str:
+        return f"dim={self.dim}, input_resolution={self.input_resolution}, num_heads={self.num_heads}, " \
+               f"window_size={self.window_size}, shift_size={self.shift_size}, mlp_ratio={self.mlp_ratio}"
+
+    def flops(self):
+        flops = 0
+        H, W = self.input_resolution
+        # norm1
+        flops += self.dim * H * W
+        # W-MSA/SW-MSA
+        nW = H * W / self.window_size / self.window_size
+        flops += nW * self.attn.flops(self.window_size * self.window_size)
+        # mlp
+        flops += 2 * H * W * self.dim * self.dim * self.mlp_ratio
+        # norm2
+        flops += self.dim * H * W
+        return flops
+
+
+class PatchMerging(nn.Module):
+    r""" Patch Merging Layer.
+
+    Args:
+        input_resolution (tuple[int]): Resolution of input feature.
+        dim (int): Number of input channels.
+        norm_layer (nn.Module, optional): Normalization layer.  Default: nn.LayerNorm
+    """
+
+    def __init__(self, input_resolution, dim, norm_layer=nn.LayerNorm):
+        super().__init__()
+        self.input_resolution = input_resolution
+        self.dim = dim
+        self.reduction = nn.Linear(4 * dim, 2 * dim, bias=False)
+        self.norm = norm_layer(4 * dim)
+
+    def forward(self, x):
+        """
+        x: B, H*W, C
+        """
+        H, W = self.input_resolution
+        B, L, C = x.shape
+        assert L == H * W, "input feature has wrong size"
+        assert H % 2 == 0 and W % 2 == 0, f"x size ({H}*{W}) are not even."
+
+        x = x.view(B, H, W, C)
+
+        x0 = x[:, 0::2, 0::2, :]  # B H/2 W/2 C
+        x1 = x[:, 1::2, 0::2, :]  # B H/2 W/2 C
+        x2 = x[:, 0::2, 1::2, :]  # B H/2 W/2 C
+        x3 = x[:, 1::2, 1::2, :]  # B H/2 W/2 C
+        x = torch.cat([x0, x1, x2, x3], -1)  # B H/2 W/2 4*C
+        x = x.view(B, -1, 4 * C)  # B H/2*W/2 4*C
+
+        x = self.norm(x)
+        x = self.reduction(x)
+
+        return x
+
+    def extra_repr(self) -> str:
+        return f"input_resolution={self.input_resolution}, dim={self.dim}"
+
+    def flops(self):
+        H, W = self.input_resolution
+        flops = H * W * self.dim
+        flops += (H // 2) * (W // 2) * 4 * self.dim * 2 * self.dim
+        return flops
+
+
+class BasicLayer(nn.Module):
+    """ A basic Swin Transformer layer for one stage.
+
+    Args:
+        dim (int): Number of input channels.
+        input_resolution (tuple[int]): Input resolution.
+        depth (int): Number of blocks.
+        num_heads (int): Number of attention heads.
+        window_size (int): Local window size.
+        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+        qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
+        qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
+        drop (float, optional): Dropout rate. Default: 0.0
+        attn_drop (float, optional): Attention dropout rate. Default: 0.0
+        drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0
+        norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
+        downsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: None
+        use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False.
+    """
+
+    def __init__(self, dim, input_resolution, depth, num_heads, window_size,
+                 mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
+                 drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):
+
+        super().__init__()
+        self.dim = dim
+        self.input_resolution = input_resolution
+        self.depth = depth
+        self.use_checkpoint = use_checkpoint
+
+        # build blocks
+        self.blocks = nn.ModuleList([
+            SwinTransformerBlock(dim=dim, input_resolution=input_resolution,
+                                 num_heads=num_heads, window_size=window_size,
+                                 shift_size=0 if (i % 2 == 0) else window_size // 2,
+                                 mlp_ratio=mlp_ratio,
+                                 qkv_bias=qkv_bias, qk_scale=qk_scale,
+                                 drop=drop, attn_drop=attn_drop,
+                                 drop_path=drop_path[i] if isinstance(drop_path, list) else drop_path,
+                                 norm_layer=norm_layer)
+            for i in range(depth)])
+
+        # patch merging layer
+        if downsample is not None:
+            self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)
+        else:
+            self.downsample = None
+
+    def forward(self, x, x_size):
+        for blk in self.blocks:
+            if self.use_checkpoint:
+                x = checkpoint.checkpoint(blk, x, x_size)
+            else:
+                x = blk(x, x_size)
+        if self.downsample is not None:
+            x = self.downsample(x)
+        return x
+
+    def extra_repr(self) -> str:
+        return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"
+
+    def flops(self):
+        flops = 0
+        for blk in self.blocks:
+            flops += blk.flops()
+        if self.downsample is not None:
+            flops += self.downsample.flops()
+        return flops
+
+
+class RSTB(nn.Module):
+    """Residual Swin Transformer Block (RSTB).
+
+    Args:
+        dim (int): Number of input channels.
+        input_resolution (tuple[int]): Input resolution.
+        depth (int): Number of blocks.
+        num_heads (int): Number of attention heads.
+        window_size (int): Local window size.
+        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+        qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
+        qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
+        drop (float, optional): Dropout rate. Default: 0.0
+        attn_drop (float, optional): Attention dropout rate. Default: 0.0
+        drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0
+        norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
+        downsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: None
+        use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False.
+        img_size: Input image size.
+        patch_size: Patch size.
+        resi_connection: The convolutional block before residual connection.
+    """
+
+    def __init__(self, dim, input_resolution, depth, num_heads, window_size,
+                 mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
+                 drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False,
+                 img_size=224, patch_size=4, resi_connection='1conv'):
+        super(RSTB, self).__init__()
+
+        self.dim = dim
+        self.input_resolution = input_resolution
+
+        self.residual_group = BasicLayer(dim=dim,
+                                         input_resolution=input_resolution,
+                                         depth=depth,
+                                         num_heads=num_heads,
+                                         window_size=window_size,
+                                         mlp_ratio=mlp_ratio,
+                                         qkv_bias=qkv_bias, qk_scale=qk_scale,
+                                         drop=drop, attn_drop=attn_drop,
+                                         drop_path=drop_path,
+                                         norm_layer=norm_layer,
+                                         downsample=downsample,
+                                         use_checkpoint=use_checkpoint)
+
+        if resi_connection == '1conv':
+            self.conv = nn.Conv2d(dim, dim, 3, 1, 1)
+        elif resi_connection == '3conv':
+            # to save parameters and memory
+            self.conv = nn.Sequential(nn.Conv2d(dim, dim // 4, 3, 1, 1), nn.LeakyReLU(negative_slope=0.2, inplace=True),
+                                      nn.Conv2d(dim // 4, dim // 4, 1, 1, 0),
+                                      nn.LeakyReLU(negative_slope=0.2, inplace=True),
+                                      nn.Conv2d(dim // 4, dim, 3, 1, 1))
+
+        self.patch_embed = PatchEmbed(
+            img_size=img_size, patch_size=patch_size, in_chans=0, embed_dim=dim,
+            norm_layer=None)
+
+        self.patch_unembed = PatchUnEmbed(
+            img_size=img_size, patch_size=patch_size, in_chans=0, embed_dim=dim,
+            norm_layer=None)
+
+    def forward(self, x, x_size):
+        return self.patch_embed(self.conv(self.patch_unembed(self.residual_group(x, x_size), x_size))) + x
+
+    def flops(self):
+        flops = 0
+        flops += self.residual_group.flops()
+        H, W = self.input_resolution
+        flops += H * W * self.dim * self.dim * 9
+        flops += self.patch_embed.flops()
+        flops += self.patch_unembed.flops()
+
+        return flops
+
+
+class PatchEmbed(nn.Module):
+    r""" Image to Patch Embedding
+
+    Args:
+        img_size (int): Image size.  Default: 224.
+        patch_size (int): Patch token size. Default: 4.
+        in_chans (int): Number of input image channels. Default: 3.
+        embed_dim (int): Number of linear projection output channels. Default: 96.
+        norm_layer (nn.Module, optional): Normalization layer. Default: None
+    """
+
+    def __init__(self, img_size=224, patch_size=4, in_chans=3, embed_dim=96, norm_layer=None):
+        super().__init__()
+        img_size = to_2tuple(img_size)
+        patch_size = to_2tuple(patch_size)
+        patches_resolution = [img_size[0] // patch_size[0], img_size[1] // patch_size[1]]
+        self.img_size = img_size
+        self.patch_size = patch_size
+        self.patches_resolution = patches_resolution
+        self.num_patches = patches_resolution[0] * patches_resolution[1]
+
+        self.in_chans = in_chans
+        self.embed_dim = embed_dim
+
+        if norm_layer is not None:
+            self.norm = norm_layer(embed_dim)
+        else:
+            self.norm = None
+
+    def forward(self, x):
+        x = x.flatten(2).transpose(1, 2)  # B Ph*Pw C
+        if self.norm is not None:
+            x = self.norm(x)
+        return x
+
+    def flops(self):
+        flops = 0
+        H, W = self.img_size
+        if self.norm is not None:
+            flops += H * W * self.embed_dim
+        return flops
+
+
+class PatchUnEmbed(nn.Module):
+    r""" Image to Patch Unembedding
+
+    Args:
+        img_size (int): Image size.  Default: 224.
+        patch_size (int): Patch token size. Default: 4.
+        in_chans (int): Number of input image channels. Default: 3.
+        embed_dim (int): Number of linear projection output channels. Default: 96.
+        norm_layer (nn.Module, optional): Normalization layer. Default: None
+    """
+
+    def __init__(self, img_size=224, patch_size=4, in_chans=3, embed_dim=96, norm_layer=None):
+        super().__init__()
+        img_size = to_2tuple(img_size)
+        patch_size = to_2tuple(patch_size)
+        patches_resolution = [img_size[0] // patch_size[0], img_size[1] // patch_size[1]]
+        self.img_size = img_size
+        self.patch_size = patch_size
+        self.patches_resolution = patches_resolution
+        self.num_patches = patches_resolution[0] * patches_resolution[1]
+
+        self.in_chans = in_chans
+        self.embed_dim = embed_dim
+
+    def forward(self, x, x_size):
+        B, HW, C = x.shape
+        x = x.transpose(1, 2).view(B, self.embed_dim, x_size[0], x_size[1])  # B Ph*Pw C
+        return x
+
+    def flops(self):
+        flops = 0
+        return flops
+
+
+class Upsample(nn.Sequential):
+    """Upsample module.
+
+    Args:
+        scale (int): Scale factor. Supported scales: 2^n and 3.
+        num_feat (int): Channel number of intermediate features.
+    """
+
+    def __init__(self, scale, num_feat):
+        m = []
+        if (scale & (scale - 1)) == 0:  # scale = 2^n
+            for _ in range(int(math.log(scale, 2))):
+                m.append(nn.Conv2d(num_feat, 4 * num_feat, 3, 1, 1))
+                m.append(nn.PixelShuffle(2))
+        elif scale == 3:
+            m.append(nn.Conv2d(num_feat, 9 * num_feat, 3, 1, 1))
+            m.append(nn.PixelShuffle(3))
+        else:
+            raise ValueError(f'scale {scale} is not supported. ' 'Supported scales: 2^n and 3.')
+        super(Upsample, self).__init__(*m)
+
+
+class UpsampleOneStep(nn.Sequential):
+    """UpsampleOneStep module (the difference with Upsample is that it always only has 1conv + 1pixelshuffle)
+       Used in lightweight SR to save parameters.
+
+    Args:
+        scale (int): Scale factor. Supported scales: 2^n and 3.
+        num_feat (int): Channel number of intermediate features.
+
+    """
+
+    def __init__(self, scale, num_feat, num_out_ch, input_resolution=None):
+        self.num_feat = num_feat
+        self.input_resolution = input_resolution
+        m = []
+        m.append(nn.Conv2d(num_feat, (scale ** 2) * num_out_ch, 3, 1, 1))
+        m.append(nn.PixelShuffle(scale))
+        super(UpsampleOneStep, self).__init__(*m)
+
+    def flops(self):
+        H, W = self.input_resolution
+        flops = H * W * self.num_feat * 3 * 9
+        return flops
+
+
+class SwinIR(nn.Module):
+    r""" SwinIR
+        A PyTorch impl of : `SwinIR: Image Restoration Using Swin Transformer`, based on Swin Transformer.
+
+    Args:
+        img_size (int | tuple(int)): Input image size. Default 64
+        patch_size (int | tuple(int)): Patch size. Default: 1
+        in_chans (int): Number of input image channels. Default: 3
+        embed_dim (int): Patch embedding dimension. Default: 96
+        depths (tuple(int)): Depth of each Swin Transformer layer.
+        num_heads (tuple(int)): Number of attention heads in different layers.
+        window_size (int): Window size. Default: 7
+        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim. Default: 4
+        qkv_bias (bool): If True, add a learnable bias to query, key, value. Default: True
+        qk_scale (float): Override default qk scale of head_dim ** -0.5 if set. Default: None
+        drop_rate (float): Dropout rate. Default: 0
+        attn_drop_rate (float): Attention dropout rate. Default: 0
+        drop_path_rate (float): Stochastic depth rate. Default: 0.1
+        norm_layer (nn.Module): Normalization layer. Default: nn.LayerNorm.
+        ape (bool): If True, add absolute position embedding to the patch embedding. Default: False
+        patch_norm (bool): If True, add normalization after patch embedding. Default: True
+        use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False
+        upscale: Upscale factor. 2/3/4/8 for image SR, 1 for denoising and compress artifact reduction
+        img_range: Image range. 1. or 255.
+        upsampler: The reconstruction reconstruction module. 'pixelshuffle'/'pixelshuffledirect'/'nearest+conv'/None
+        resi_connection: The convolutional block before residual connection. '1conv'/'3conv'
+    """
+
+    def __init__(self, img_size=64, patch_size=1, in_chans=3,
+                 embed_dim=96, depths=[6, 6, 6, 6], num_heads=[6, 6, 6, 6],
+                 window_size=7, mlp_ratio=4., qkv_bias=True, qk_scale=None,
+                 drop_rate=0., attn_drop_rate=0., drop_path_rate=0.1,
+                 norm_layer=nn.LayerNorm, ape=False, patch_norm=True,
+                 use_checkpoint=False, upscale=2, img_range=1., upsampler='', resi_connection='1conv',
+                 **kwargs):
+        super(SwinIR, self).__init__()
+        num_in_ch = in_chans
+        num_out_ch = in_chans
+        num_feat = 64
+        self.img_range = img_range
+        if in_chans == 3:
+            rgb_mean = (0.4488, 0.4371, 0.4040)
+            self.mean = torch.Tensor(rgb_mean).view(1, 3, 1, 1)
+        else:
+            self.mean = torch.zeros(1, 1, 1, 1)
+        self.upscale = upscale
+        self.upsampler = upsampler
+        self.window_size = window_size
+
+        #####################################################################################################
+        ################################### 1, shallow feature extraction ###################################
+        self.conv_first = nn.Conv2d(num_in_ch, embed_dim, 3, 1, 1)
+
+        #####################################################################################################
+        ################################### 2, deep feature extraction ######################################
+        self.num_layers = len(depths)
+        self.embed_dim = embed_dim
+        self.ape = ape
+        self.patch_norm = patch_norm
+        self.num_features = embed_dim
+        self.mlp_ratio = mlp_ratio
+
+        # split image into non-overlapping patches
+        self.patch_embed = PatchEmbed(
+            img_size=img_size, patch_size=patch_size, in_chans=embed_dim, embed_dim=embed_dim,
+            norm_layer=norm_layer if self.patch_norm else None)
+        num_patches = self.patch_embed.num_patches
+        patches_resolution = self.patch_embed.patches_resolution
+        self.patches_resolution = patches_resolution
+
+        # merge non-overlapping patches into image
+        self.patch_unembed = PatchUnEmbed(
+            img_size=img_size, patch_size=patch_size, in_chans=embed_dim, embed_dim=embed_dim,
+            norm_layer=norm_layer if self.patch_norm else None)
+
+        # absolute position embedding
+        if self.ape:
+            self.absolute_pos_embed = nn.Parameter(torch.zeros(1, num_patches, embed_dim))
+            trunc_normal_(self.absolute_pos_embed, std=.02)
+
+        self.pos_drop = nn.Dropout(p=drop_rate)
+
+        # stochastic depth
+        dpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))]  # stochastic depth decay rule
+
+        # build Residual Swin Transformer blocks (RSTB)
+        self.layers = nn.ModuleList()
+        for i_layer in range(self.num_layers):
+            layer = RSTB(dim=embed_dim,
+                         input_resolution=(patches_resolution[0],
+                                           patches_resolution[1]),
+                         depth=depths[i_layer],
+                         num_heads=num_heads[i_layer],
+                         window_size=window_size,
+                         mlp_ratio=self.mlp_ratio,
+                         qkv_bias=qkv_bias, qk_scale=qk_scale,
+                         drop=drop_rate, attn_drop=attn_drop_rate,
+                         drop_path=dpr[sum(depths[:i_layer]):sum(depths[:i_layer + 1])],  # no impact on SR results
+                         norm_layer=norm_layer,
+                         downsample=None,
+                         use_checkpoint=use_checkpoint,
+                         img_size=img_size,
+                         patch_size=patch_size,
+                         resi_connection=resi_connection
+
+                         )
+            self.layers.append(layer)
+        self.norm = norm_layer(self.num_features)
+
+        # build the last conv layer in deep feature extraction
+        if resi_connection == '1conv':
+            self.conv_after_body = nn.Conv2d(embed_dim, embed_dim, 3, 1, 1)
+        elif resi_connection == '3conv':
+            # to save parameters and memory
+            self.conv_after_body = nn.Sequential(nn.Conv2d(embed_dim, embed_dim // 4, 3, 1, 1),
+                                                 nn.LeakyReLU(negative_slope=0.2, inplace=True),
+                                                 nn.Conv2d(embed_dim // 4, embed_dim // 4, 1, 1, 0),
+                                                 nn.LeakyReLU(negative_slope=0.2, inplace=True),
+                                                 nn.Conv2d(embed_dim // 4, embed_dim, 3, 1, 1))
+
+        #####################################################################################################
+        ################################ 3, high quality image reconstruction ################################
+        if self.upsampler == 'pixelshuffle':
+            # for classical SR
+            self.conv_before_upsample = nn.Sequential(nn.Conv2d(embed_dim, num_feat, 3, 1, 1),
+                                                      nn.LeakyReLU(inplace=True))
+            self.upsample = Upsample(upscale, num_feat)
+            self.conv_last = nn.Conv2d(num_feat, num_out_ch, 3, 1, 1)
+        elif self.upsampler == 'pixelshuffledirect':
+            # for lightweight SR (to save parameters)
+            self.upsample = UpsampleOneStep(upscale, embed_dim, num_out_ch,
+                                            (patches_resolution[0], patches_resolution[1]))
+        elif self.upsampler == 'nearest+conv':
+            # for real-world SR (less artifacts)
+            self.conv_before_upsample = nn.Sequential(nn.Conv2d(embed_dim, num_feat, 3, 1, 1),
+                                                      nn.LeakyReLU(inplace=True))
+            self.conv_up1 = nn.Conv2d(num_feat, num_feat, 3, 1, 1)
+            if self.upscale == 4:
+                self.conv_up2 = nn.Conv2d(num_feat, num_feat, 3, 1, 1)
+            self.conv_hr = nn.Conv2d(num_feat, num_feat, 3, 1, 1)
+            self.conv_last = nn.Conv2d(num_feat, num_out_ch, 3, 1, 1)
+            self.lrelu = nn.LeakyReLU(negative_slope=0.2, inplace=True)
+        else:
+            # for image denoising and JPEG compression artifact reduction
+            self.conv_last = nn.Conv2d(embed_dim, num_out_ch, 3, 1, 1)
+
+        self.apply(self._init_weights)
+
+    def _init_weights(self, m):
+        if isinstance(m, nn.Linear):
+            trunc_normal_(m.weight, std=.02)
+            if isinstance(m, nn.Linear) and m.bias is not None:
+                nn.init.constant_(m.bias, 0)
+        elif isinstance(m, nn.LayerNorm):
+            nn.init.constant_(m.bias, 0)
+            nn.init.constant_(m.weight, 1.0)
+
+    @torch.jit.ignore
+    def no_weight_decay(self):
+        return {'absolute_pos_embed'}
+
+    @torch.jit.ignore
+    def no_weight_decay_keywords(self):
+        return {'relative_position_bias_table'}
+
+    def check_image_size(self, x):
+        _, _, h, w = x.size()
+        mod_pad_h = (self.window_size - h % self.window_size) % self.window_size
+        mod_pad_w = (self.window_size - w % self.window_size) % self.window_size
+        x = F.pad(x, (0, mod_pad_w, 0, mod_pad_h), 'reflect')
+        return x
+
+    def forward_features(self, x):
+        x_size = (x.shape[2], x.shape[3])
+        x = self.patch_embed(x)
+        if self.ape:
+            x = x + self.absolute_pos_embed
+        x = self.pos_drop(x)
+
+        for layer in self.layers:
+            x = layer(x, x_size)
+
+        x = self.norm(x)  # B L C
+        x = self.patch_unembed(x, x_size)
+
+        return x
+
+    def forward(self, x):
+        H, W = x.shape[2:]
+        x = self.check_image_size(x)
+        
+        self.mean = self.mean.type_as(x)
+        x = (x - self.mean) * self.img_range
+
+        if self.upsampler == 'pixelshuffle':
+            # for classical SR
+            x = self.conv_first(x)
+            x = self.conv_after_body(self.forward_features(x)) + x
+            x = self.conv_before_upsample(x)
+            x = self.conv_last(self.upsample(x))
+        elif self.upsampler == 'pixelshuffledirect':
+            # for lightweight SR
+            x = self.conv_first(x)
+            x = self.conv_after_body(self.forward_features(x)) + x
+            x = self.upsample(x)
+        elif self.upsampler == 'nearest+conv':
+            # for real-world SR
+            x = self.conv_first(x)
+            x = self.conv_after_body(self.forward_features(x)) + x
+            x = self.conv_before_upsample(x)
+            x = self.lrelu(self.conv_up1(torch.nn.functional.interpolate(x, scale_factor=2, mode='nearest')))
+            if self.upscale == 4:
+                x = self.lrelu(self.conv_up2(torch.nn.functional.interpolate(x, scale_factor=2, mode='nearest')))
+            x = self.conv_last(self.lrelu(self.conv_hr(x)))
+        else:
+            # for image denoising and JPEG compression artifact reduction
+            x_first = self.conv_first(x)
+            res = self.conv_after_body(self.forward_features(x_first)) + x_first
+            x = x + self.conv_last(res)
+
+        x = x / self.img_range + self.mean
+
+        return x[:, :, :H*self.upscale, :W*self.upscale]
+
+    def flops(self):
+        flops = 0
+        H, W = self.patches_resolution
+        flops += H * W * 3 * self.embed_dim * 9
+        flops += self.patch_embed.flops()
+        for i, layer in enumerate(self.layers):
+            flops += layer.flops()
+        flops += H * W * 3 * self.embed_dim * self.embed_dim
+        flops += self.upsample.flops()
+        return flops
+
+
+if __name__ == '__main__':
+    upscale = 4
+    window_size = 8
+    height = (1024 // upscale // window_size + 1) * window_size
+    width = (720 // upscale // window_size + 1) * window_size
+    model = SwinIR(upscale=2, img_size=(height, width),
+                   window_size=window_size, img_range=1., depths=[6, 6, 6, 6],
+                   embed_dim=60, num_heads=[6, 6, 6, 6], mlp_ratio=2, upsampler='pixelshuffledirect')
+    print(model)
+    print(height, width, model.flops() / 1e9)
+
+    x = torch.randn((1, 3, height, width))
+    x = model(x)
+    print(x.shape)

predict.py

@@ -0,0 +1,159 @@
+import cog
+import tempfile
+from pathlib import Path
+import argparse
+import shutil
+import os
+import cv2
+import glob
+import torch
+from collections import OrderedDict
+import numpy as np
+from main_test_swinir import define_model, setup, get_image_pair
+
+
+class Predictor(cog.Predictor):
+    def setup(self):
+        model_dir = 'experiments/pretrained_models'
+
+        self.model_zoo = {
+            'real_sr': {
+                4: os.path.join(model_dir, '003_realSR_BSRGAN_DFO_s64w8_SwinIR-M_x4_GAN.pth')
+            },
+            'gray_dn': {
+                15: os.path.join(model_dir, '004_grayDN_DFWB_s128w8_SwinIR-M_noise15.pth'),
+                25: os.path.join(model_dir, '004_grayDN_DFWB_s128w8_SwinIR-M_noise25.pth'),
+                50: os.path.join(model_dir, '004_grayDN_DFWB_s128w8_SwinIR-M_noise50.pth')
+            },
+            'color_dn': {
+                15: os.path.join(model_dir, '005_colorDN_DFWB_s128w8_SwinIR-M_noise15.pth'),
+                25: os.path.join(model_dir, '005_colorDN_DFWB_s128w8_SwinIR-M_noise25.pth'),
+                50: os.path.join(model_dir, '005_colorDN_DFWB_s128w8_SwinIR-M_noise50.pth')
+            },
+            'jpeg_car': {
+                10: os.path.join(model_dir, '006_CAR_DFWB_s126w7_SwinIR-M_jpeg10.pth'),
+                20: os.path.join(model_dir, '006_CAR_DFWB_s126w7_SwinIR-M_jpeg20.pth'),
+                30: os.path.join(model_dir, '006_CAR_DFWB_s126w7_SwinIR-M_jpeg30.pth'),
+                40: os.path.join(model_dir, '006_CAR_DFWB_s126w7_SwinIR-M_jpeg40.pth')
+            }
+        }
+
+        parser = argparse.ArgumentParser()
+        parser.add_argument('--task', type=str, default='real_sr', help='classical_sr, lightweight_sr, real_sr, '
+                                                                        'gray_dn, color_dn, jpeg_car')
+        parser.add_argument('--scale', type=int, default=1, help='scale factor: 1, 2, 3, 4, 8')  # 1 for dn and jpeg car
+        parser.add_argument('--noise', type=int, default=15, help='noise level: 15, 25, 50')
+        parser.add_argument('--jpeg', type=int, default=40, help='scale factor: 10, 20, 30, 40')
+        parser.add_argument('--training_patch_size', type=int, default=128, help='patch size used in training SwinIR. '
+                                                                                 'Just used to differentiate two different settings in Table 2 of the paper. '
+                                                                                 'Images are NOT tested patch by patch.')
+        parser.add_argument('--large_model', action='store_true',
+                            help='use large model, only provided for real image sr')
+        parser.add_argument('--model_path', type=str,
+                            default=self.model_zoo['real_sr'][4])
+        parser.add_argument('--folder_lq', type=str, default=None, help='input low-quality test image folder')
+        parser.add_argument('--folder_gt', type=str, default=None, help='input ground-truth test image folder')
+
+        self.args = parser.parse_args('')
+
+        self.device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+        self.tasks = {
+            'Real-World Image Super-Resolution': 'real_sr',
+            'Grayscale Image Denoising': 'gray_dn',
+            'Color Image Denoising': 'color_dn',
+            'JPEG Compression Artifact Reduction': 'jpeg_car'
+        }
+
+    @cog.input("image", type=Path, help="input image")
+    @cog.input("task_type", type=str, default='Real-World Image Super-Resolution',
+               options=['Real-World Image Super-Resolution', 'Grayscale Image Denoising', 'Color Image Denoising',
+                        'JPEG Compression Artifact Reduction'],
+               help="image restoration task type")
+    @cog.input("noise", type=int, default=15, options=[15, 25, 50],
+               help='noise level, activated for Grayscale Image Denoising and Color Image Denoising. '
+                    'Leave it as default or arbitrary if other tasks are selected')
+    @cog.input("jpeg", type=int, default=40, options=[10, 20, 30, 40],
+               help='scale factor, activated for JPEG Compression Artifact Reduction. '
+                    'Leave it as default or arbitrary if other tasks are selected')
+    def predict(self, image, task_type='Real-World Image Super-Resolution', jpeg=40, noise=15):
+
+        self.args.task = self.tasks[task_type]
+        self.args.noise = noise
+        self.args.jpeg = jpeg
+
+        # set model path
+        if self.args.task == 'real_sr':
+            self.args.scale = 4
+            self.args.model_path = self.model_zoo[self.args.task][4]
+        elif self.args.task in ['gray_dn', 'color_dn']:
+            self.args.model_path = self.model_zoo[self.args.task][noise]
+        else:
+            self.args.model_path = self.model_zoo[self.args.task][jpeg]
+
+        try:
+            # set input folder
+            input_dir = 'input_cog_temp'
+            os.makedirs(input_dir, exist_ok=True)
+            input_path = os.path.join(input_dir, os.path.basename(image))
+            shutil.copy(str(image), input_path)
+            if self.args.task == 'real_sr':
+                self.args.folder_lq = input_dir
+            else:
+                self.args.folder_gt = input_dir
+
+            model = define_model(self.args)
+            model.eval()
+            model = model.to(self.device)
+
+            # setup folder and path
+            folder, save_dir, border, window_size = setup(self.args)
+            os.makedirs(save_dir, exist_ok=True)
+            test_results = OrderedDict()
+            test_results['psnr'] = []
+            test_results['ssim'] = []
+            test_results['psnr_y'] = []
+            test_results['ssim_y'] = []
+            test_results['psnr_b'] = []
+            # psnr, ssim, psnr_y, ssim_y, psnr_b = 0, 0, 0, 0, 0
+            out_path = Path(tempfile.mkdtemp()) / "out.png"
+
+            for idx, path in enumerate(sorted(glob.glob(os.path.join(folder, '*')))):
+                # read image
+                imgname, img_lq, img_gt = get_image_pair(self.args, path)  # image to HWC-BGR, float32
+                img_lq = np.transpose(img_lq if img_lq.shape[2] == 1 else img_lq[:, :, [2, 1, 0]],
+                                      (2, 0, 1))  # HCW-BGR to CHW-RGB
+                img_lq = torch.from_numpy(img_lq).float().unsqueeze(0).to(self.device)  # CHW-RGB to NCHW-RGB
+
+                # inference
+                with torch.no_grad():
+                    # pad input image to be a multiple of window_size
+                    _, _, h_old, w_old = img_lq.size()
+                    h_pad = (h_old // window_size + 1) * window_size - h_old
+                    w_pad = (w_old // window_size + 1) * window_size - w_old
+                    img_lq = torch.cat([img_lq, torch.flip(img_lq, [2])], 2)[:, :, :h_old + h_pad, :]
+                    img_lq = torch.cat([img_lq, torch.flip(img_lq, [3])], 3)[:, :, :, :w_old + w_pad]
+                    output = model(img_lq)
+                    output = output[..., :h_old * self.args.scale, :w_old * self.args.scale]
+
+                # save image
+                output = output.data.squeeze().float().cpu().clamp_(0, 1).numpy()
+                if output.ndim == 3:
+                    output = np.transpose(output[[2, 1, 0], :, :], (1, 2, 0))  # CHW-RGB to HCW-BGR
+                output = (output * 255.0).round().astype(np.uint8)  # float32 to uint8
+                cv2.imwrite(str(out_path), output)
+        finally:
+            clean_folder(input_dir)
+        return out_path
+
+
+def clean_folder(folder):
+    for filename in os.listdir(folder):
+        file_path = os.path.join(folder, filename)
+        try:
+            if os.path.isfile(file_path) or os.path.islink(file_path):
+                os.unlink(file_path)
+            elif os.path.isdir(file_path):
+                shutil.rmtree(file_path)
+        except Exception as e:
+            print('Failed to delete %s. Reason: %s' % (file_path, e))