{"id":13419223,"url":"https://github.com/leverxgroup/esrgan","last_synced_at":"2025-03-15T05:30:30.183Z","repository":{"id":42208759,"uuid":"304548950","full_name":"leverxgroup/esrgan","owner":"leverxgroup","description":"Enhanced SRGAN. Champion PIRM Challenge on Perceptual Super-Resolution","archived":false,"fork":false,"pushed_at":"2022-09-17T10:32:31.000Z","size":11370,"stargazers_count":121,"open_issues_count":1,"forks_count":15,"subscribers_count":5,"default_branch":"master","last_synced_at":"2024-01-14T10:08:19.495Z","etag":null,"topics":["computer-vision","deep-neural-networks","generative-adversarial-network","image-processing","pytorch","super-resolution"],"latest_commit_sha":null,"homepage":"https://esrgan.readthedocs.io","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/leverxgroup.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2020-10-16T07:13:46.000Z","updated_at":"2024-01-03T23:30:40.000Z","dependencies_parsed_at":"2023-01-18T12:00:54.430Z","dependency_job_id":null,"html_url":"https://github.com/leverxgroup/esrgan","commit_stats":null,"previous_names":[],"tags_count":3,"template":null,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/leverxgroup%2Fesrgan","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/leverxgroup%2Fesrgan/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/leverxgroup%2Fesrgan/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/leverxgroup%2Fesrgan/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/leverxgroup","download_url":"https://codeload.github.com/leverxgroup/esrgan/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":243690115,"owners_count":20331725,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["computer-vision","deep-neural-networks","generative-adversarial-network","image-processing","pytorch","super-resolution"],"created_at":"2024-07-30T22:01:12.997Z","updated_at":"2025-03-15T05:30:29.655Z","avatar_url":"https://github.com/leverxgroup.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"# ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks\nPipeine for Image Super-Resolution task that based on a frequently cited paper, [ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks](https://arxiv.org/abs/1809.00219) (Wang Xintao et al.), published in 2018.\n\nIn few words, image super-resolution (SR) techniques reconstruct a higher-resolution (HR) image or sequence\nfrom the observed lower-resolution (LR) images, e.g. upscaling of 720p image into 1080p.\n\nOne of the common approaches to solving this task is to use deep convolutional neural networks\ncapable of recovering HR images from LR ones. And ESRGAN (Enhanced SRGAN) is one of them.\nKey points of ESRGAN:\n\n- SRResNet-based architecture with residual-in-residual blocks;\n- Mixture of context, perceptual, and adversarial losses. Context and perceptual losses are used for proper image upscaling,\n  while adversarial loss pushes neural network to the natural image manifold using a discriminator network\n  that is trained to differentiate between the super-resolved images and original photo-realistic images.\n\n![ESRGAN architecture](docs/_static/architecture.png)\n\n### Technologies\n\n* `Catalyst` as pipeline runner for deep learning tasks. This new and rapidly developing [library](https://github.com/catalyst-team/catalyst).\n  can significantly reduce the amount of boilerplate code. If you are familiar with the TensorFlow ecosystem, you can think of Catalyst\n  as Keras for PyTorch. This framework is integrated with logging systems such as the well-known [TensorBoard](https://www.tensorflow.org/tensorboard);\n* `Pytorch` and `torchvision` as main frameworks for deep learning;\n* `Albumentations` and `PIQ` for data processing.\n\n## Quick Start\n\n### Setup environment\n```bash\npip install git+https://github.com/leverxgroup/esrgan.git\n```\n\n### Run an experiment\n```bash\ncatalyst-dl run -C esrgan/config.yml --benchmark\n```\nwhere `esrgan/config.yml` is a path to the [config](config.yml) file.\n\n## Results\nSome examples of work of ESRGAN model trained on [DIV2K](https://data.vision.ee.ethz.ch/cvl/DIV2K) dataset:\n\n| LR\u003c/br\u003e(low resolution) | ESRGAN\u003c/br\u003e(original) | ESRGAN\u003c/br\u003e(ours) | HR\u003c/br\u003e(high resolution) |\n|:---:|:---:|:---:|:---:|\n| \u003cimg src=\"docs/_static/0853lr.png\" height=\"128\" width=\"128\"/\u003e | \u003cimg src=\"docs/_static/0853sr.png\" height=\"128\" width=\"128\"/\u003e | \u003cimg src=\"docs/_static/0853.png\" height=\"128\" width=\"128\"/\u003e | \u003cimg src=\"docs/_static/0853hr.png\" height=\"128\" width=\"128\"/\u003e |\n| \u003cimg src=\"docs/_static/0857lr.png\" height=\"128\" width=\"128\"/\u003e | \u003cimg src=\"docs/_static/0857sr.png\" height=\"128\" width=\"128\"/\u003e | \u003cimg src=\"docs/_static/0857.png\" height=\"128\" width=\"128\"/\u003e | \u003cimg src=\"docs/_static/0857hr.png\" height=\"128\" width=\"128\"/\u003e |\n| \u003cimg src=\"docs/_static/0887lr.png\" height=\"128\" width=\"128\"/\u003e | \u003cimg src=\"docs/_static/0887sr.png\" height=\"128\" width=\"128\"/\u003e | \u003cimg src=\"docs/_static/0887.png\" height=\"128\" width=\"128\"/\u003e | \u003cimg src=\"docs/_static/0887hr.png\" height=\"128\" width=\"128\"/\u003e |\n\n## Documentation\nFull documentation for the project is available at https://esrgan.readthedocs.io/\n\n## License\n`esrgan` is released under a CC BY-NC-ND 4.0 license. See [LICENSE](LICENSE) for additional details about it.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fleverxgroup%2Fesrgan","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fleverxgroup%2Fesrgan","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fleverxgroup%2Fesrgan/lists"}