[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-XPixelGroup--BasicSR":3,"tool-XPixelGroup--BasicSR":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",138956,2,"2026-04-05T11:33:21",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":75,"owner_avatar_url":76,"owner_bio":77,"owner_company":78,"owner_location":78,"owner_email":79,"owner_twitter":78,"owner_website":80,"owner_url":81,"languages":82,"stars":103,"forks":104,"last_commit_at":105,"license":106,"difficulty_score":10,"env_os":107,"env_gpu":107,"env_ram":107,"env_deps":108,"category_tags":112,"github_topics":113,"view_count":10,"oss_zip_url":78,"oss_zip_packed_at":78,"status":16,"created_at":129,"updated_at":130,"faqs":131,"releases":160},1133,"XPixelGroup\u002FBasicSR","BasicSR","Open Source Image and Video Restoration Toolbox for Super-resolution, Denoise, Deblurring, etc. Currently, it includes EDSR, RCAN, SRResNet, SRGAN, ESRGAN, EDVR, BasicVSR, SwinIR, ECBSR, etc. Also support StyleGAN2, DFDNet.","BasicSR 是一个基于 PyTorch 构建的开源图像与视频复原工具箱，致力于提供高质量的视觉内容修复解决方案。它主要解决低分辨率、模糊、噪点过多或存在压缩伪影的图片与视频质量问题，支持超分辨率、去噪、去模糊等多种核心任务。\n\n对于计算机视觉研究人员和开发者而言，BasicSR 是理想的实验平台。它集成了大量业界领先的模型，包括 EDSR、ESRGAN、SwinIR、BasicVSR 以及针对移动端优化的 ECBSR 等。用户不仅可以直接使用预训练模型进行推理，还能利用其完善的训练脚本和数据集准备工具，轻松复现论文或定制自己的算法。\n\n该项目注重工程化落地，提供了从数据准备、模型训练到结果评估的全流程支持，并包含绘图脚本以便直观分析性能。随着加入 XPixel 社区，BasicSR 持续更新维护，拥有活跃的中文交流群体和详细的文档指引。无论你是想深入探究复原算法原理，还是需要为产品集成画质增强功能，BasicSR 都能提供稳定可靠的技术基础。","\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FXPixelGroup_BasicSR_readme_c1310e056805.png\" height=120>\n\u003C\u002Fp>\n\n## \u003Cdiv align=\"center\">\u003Cb>\u003Ca href=\"README.md\">English\u003C\u002Fa> | \u003Ca href=\"README_CN.md\">简体中文\u003C\u002Fa>\u003C\u002Fb>\u003C\u002Fdiv>\n\n\u003Cdiv align=\"center\">\n\n[![LICENSE](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fxinntao\u002Fbasicsr.svg)](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002FLICENSE.txt)\n[![PyPI](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fbasicsr)](https:\u002F\u002Fpypi.org\u002Fproject\u002Fbasicsr\u002F)\n[![Language grade: Python](https:\u002F\u002Fimg.shields.io\u002Flgtm\u002Fgrade\u002Fpython\u002Fg\u002Fxinntao\u002FBasicSR.svg?logo=lgtm&logoWidth=18)](https:\u002F\u002Flgtm.com\u002Fprojects\u002Fg\u002Fxinntao\u002FBasicSR\u002Fcontext:python)\n[![python lint](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Factions\u002Fworkflows\u002Fpylint.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002F.github\u002Fworkflows\u002Fpylint.yml)\n[![Publish-pip](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Factions\u002Fworkflows\u002Fpublish-pip.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002F.github\u002Fworkflows\u002Fpublish-pip.yml)\n[![gitee mirror](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Factions\u002Fworkflows\u002Fgitee-mirror.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002F.github\u002Fworkflows\u002Fgitee-mirror.yml)\n\n\u003C\u002Fdiv>\n\n\u003Cdiv align=\"center\">\n\n⚡[**HowTo**](#-HOWTOs) **|** 🔧[**Installation**](docs\u002FINSTALL.md) **|** 💻[**Training Commands**](docs\u002FTrainTest.md) **|** 🐢[**DatasetPrepare**](docs\u002FDatasetPreparation.md) **|** 🏰[**Model Zoo**](docs\u002FModelZoo.md)\n\n📕[**中文解读文档**](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR-docs) **|** 📊 [**Plot scripts**](scripts\u002Fplot) **|** 📝[Introduction](docs\u002Fintroduction.md) **|** \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Ftree\u002Fmaster\u002Fcolab\">\u003Cimg src=\"https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg\" height=\"18\" alt=\"google colab logo\">\u003C\u002Fa> **|** ⏳[TODO List](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fprojects) **|** ❓[FAQ](docs\u002FFAQ.md)\n\u003C\u002Fdiv>\n\n🚀 We add [BasicSR-Examples](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR-examples), which provides guidance and templates of using BasicSR as a python package. 🚀 \u003Cbr>\n📢 **技术交流QQ群**：**320960100**   入群答案：**互帮互助共同进步** \u003Cbr>\n🧭 [入群二维码](#-contact) (QQ、微信)    [入群指南 (腾讯文档)](https:\u002F\u002Fdocs.qq.com\u002Fdoc\u002FDYXBSUmxOT0xBZ05u) \u003Cbr>\n\n---\n\nBasicSR (**Basic** **S**uper **R**estoration) is an open-source **image and video restoration** toolbox based on PyTorch, such as super-resolution, denoise, deblurring, JPEG artifacts removal, *etc*.\u003Cbr>\nBasicSR (**Basic** **S**uper **R**estoration) 是一个基于 PyTorch 的开源 图像视频复原工具箱, 比如 超分辨率, 去噪, 去模糊, 去 JPEG 压缩噪声等.\n\n🚩 **New Features\u002FUpdates**\n\n- ✅ July 26, 2022. Add plot scripts 📊[Plot](scripts\u002Fplot).\n- ✅ May 9, 2022. BasicSR joins [XPixel](http:\u002F\u002Fxpixel.group\u002F).\n- ✅ Oct 5, 2021. Add **ECBSR training and testing** codes: [ECBSR](https:\u002F\u002Fgithub.com\u002Fxindongzhang\u002FECBSR).\n  > ACMMM21: Edge-oriented Convolution Block for Real-time Super Resolution on Mobile Devices\n- ✅ Sep 2, 2021. Add **SwinIR training and testing** codes: [SwinIR](https:\u002F\u002Fgithub.com\u002FJingyunLiang\u002FSwinIR) by [Jingyun Liang](https:\u002F\u002Fgithub.com\u002FJingyunLiang). More details are in [HOWTOs.md](docs\u002FHOWTOs.md#how-to-train-swinir-sr)\n- ✅ Aug 5, 2021. Add NIQE, which produces the same results as MATLAB (both are 5.7296 for tests\u002Fdata\u002Fbaboon.png).\n- ✅ July 31, 2021. Add **bi-directional video super-resolution** codes: [**BasicVSR** and IconVSR](https:\u002F\u002Farxiv.org\u002Fabs\u002F2012.02181).\n  > CVPR21: BasicVSR: The Search for Essential Components in Video Super-Resolution and Beyond\n- **[More](docs\u002Fhistory_updates.md)**\n\n---\n\nIf BasicSR helps your research or work, please help to ⭐ this repo or recommend it to your friends. Thanks😊 \u003Cbr>\nOther recommended projects:\u003Cbr>\n▶️ [Real-ESRGAN](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FReal-ESRGAN): A practical algorithm for general image restoration\u003Cbr>\n▶️ [GFPGAN](https:\u002F\u002Fgithub.com\u002FTencentARC\u002FGFPGAN): A practical algorithm for real-world face restoration \u003Cbr>\n▶️ [facexlib](https:\u002F\u002Fgithub.com\u002Fxinntao\u002Ffacexlib): A collection that provides useful face-relation functions.\u003Cbr>\n▶️ [HandyView](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FHandyView): A PyQt5-based image viewer that is handy for view and comparison. \u003Cbr>\n▶️ [HandyFigure](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FHandyFigure): Open source of paper figures \u003Cbr>\n\u003Csub>([ESRGAN](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FESRGAN), [EDVR](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FEDVR), [DNI](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FDNI), [SFTGAN](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FSFTGAN))\u003C\u002Fsub>\n\u003Csub>([HandyCrawler](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FHandyCrawler), [HandyWriting](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FHandyWriting))\u003C\u002Fsub>\n\n---\n\n## ⚡ HOWTOs\n\nWe provide simple pipelines to train\u002Ftest\u002Finference models for a quick start.\nThese pipelines\u002Fcommands cannot cover all the cases and more details are in the following sections.\n\n| GAN                  |                                                |                                                        |          |                                                |                                                        |\n| :------------------- | :--------------------------------------------: | :----------------------------------------------------: | :------- | :--------------------------------------------: | :----------------------------------------------------: |\n| StyleGAN2            | [Train](docs\u002FHOWTOs.md#How-to-train-StyleGAN2) | [Inference](docs\u002FHOWTOs.md#How-to-inference-StyleGAN2) |          |                                                |                                                        |\n| **Face Restoration** |                                                |                                                        |          |                                                |                                                        |\n| DFDNet               |                       -                        |  [Inference](docs\u002FHOWTOs.md#How-to-inference-DFDNet)   |          |                                                |                                                        |\n| **Super Resolution** |                                                |                                                        |          |                                                |                                                        |\n| ESRGAN               |                     *TODO*                     |                         *TODO*                         | SRGAN    |                     *TODO*                     |                         *TODO*                         |\n| EDSR                 |                     *TODO*                     |                         *TODO*                         | SRResNet |                     *TODO*                     |                         *TODO*                         |\n| RCAN                 |                     *TODO*                     |                         *TODO*                         | SwinIR   | [Train](docs\u002FHOWTOs.md#how-to-train-swinir-sr) | [Inference](docs\u002FHOWTOs.md#how-to-inference-swinir-sr) |\n| EDVR                 |                     *TODO*                     |                         *TODO*                         | DUF      |                       -                        |                         *TODO*                         |\n| BasicVSR             |                     *TODO*                     |                         *TODO*                         | TOF      |                       -                        |                         *TODO*                         |\n| **Deblurring**       |                                                |                                                        |          |                                                |                                                        |\n| DeblurGANv2          |                       -                        |                         *TODO*                         |          |                                                |                                                        |\n| **Denoise**          |                                                |                                                        |          |                                                |                                                        |\n| RIDNet               |                       -                        |                         *TODO*                         | CBDNet   |                       -                        |                         *TODO*                         |\n\n## ✨ **Projects that use BasicSR**\n\n- [**Real-ESRGAN**](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FReal-ESRGAN): A practical algorithm for general image restoration\n- [**GFPGAN**](https:\u002F\u002Fgithub.com\u002FTencentARC\u002FGFPGAN): A practical algorithm for real-world face restoration\n\nIf you use `BasicSR` in your open-source projects, welcome to contact me (by [email](#-contact) or opening an issue\u002Fpull request). I will add your projects to the above list 😊\n\n## 📜 License and Acknowledgement\n\nThis project is released under the [Apache 2.0 license](LICENSE.txt).\u003Cbr>\nMore details about **license** and **acknowledgement** are in [LICENSE](LICENSE\u002FREADME.md).\n\n## 🌏 Citations\n\nIf BasicSR helps your research or work, please cite BasicSR.\u003Cbr>\nThe following is a BibTeX reference. The BibTeX entry requires the `url` LaTeX package.\n\n``` latex\n@misc{basicsr,\n  author =       {Xintao Wang and Liangbin Xie and Ke Yu and Kelvin C.K. Chan and Chen Change Loy and Chao Dong},\n  title =        {{BasicSR}: Open Source Image and Video Restoration Toolbox},\n  howpublished = {\\url{https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR}},\n  year =         {2022}\n}\n```\n\n> Xintao Wang, Liangbin Xie, Ke Yu, Kelvin C.K. Chan, Chen Change Loy and Chao Dong. BasicSR: Open Source Image and Video Restoration Toolbox. \u003Chttps:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR>, 2022.\n\n## 📧 Contact\n\nIf you have any questions, please email `xintao.alpha@gmail.com`, `xintao.wang@outlook.com`.\n\n\u003Cbr>\n\n- **QQ群**: 扫描左边二维码 或者 搜索QQ群号: 320960100   入群答案：互帮互助共同进步\n- **微信群**: 我们的一群已经满500人啦，二群也超过200人了；进群可以添加 Liangbin 的个人微信 (右边二维码)，他会在空闲的时候拉大家入群~\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FXPixelGroup_BasicSR_readme_8f870ded4b5a.jpg\"  height=\"300\">   \n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FXPixelGroup_BasicSR_readme_ffad80acb2c3.png\"  height=\"300\">\n\u003C\u002Fp>\n\n![visitors](https:\u002F\u002Fvisitor-badge.glitch.me\u002Fbadge?page_id=XPixelGroup\u002FBasicSR) (start from 2022-11-06)\n","\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FXPixelGroup_BasicSR_readme_c1310e056805.png\" height=120>\n\u003C\u002Fp>\n\n## \u003Cdiv align=\"center\">\u003Cb>\u003Ca href=\"README.md\">English\u003C\u002Fa> | \u003Ca href=\"README_CN.md\">简体中文\u003C\u002Fa>\u003C\u002Fb>\u003C\u002Fdiv>\n\n\u003Cdiv align=\"center\">\n\n[![LICENSE](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fxinntao\u002Fbasicsr.svg)](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002FLICENSE.txt)\n[![PyPI](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fbasicsr)](https:\u002F\u002Fpypi.org\u002Fproject\u002Fbasicsr\u002F)\n[![Language grade: Python](https:\u002F\u002Fimg.shields.io\u002Flgtm\u002Fgrade\u002Fpython\u002Fg\u002Fxinntao\u002FBasicSR.svg?logo=lgtm&logoWidth=18)](https:\u002F\u002Flgtm.com\u002Fprojects\u002Fg\u002Fxinntao\u002FBasicSR\u002Fcontext:python)\n[![python lint](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Factions\u002Fworkflows\u002Fpylint.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002F.github\u002Fworkflows\u002Fpylint.yml)\n[![Publish-pip](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Factions\u002Fworkflows\u002Fpublish-pip.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002F.github\u002Fworkflows\u002Fpublish-pip.yml)\n[![gitee mirror](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Factions\u002Fworkflows\u002Fgitee-mirror.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002F.github\u002Fworkflows\u002Fgitee-mirror.yml)\n\n\u003C\u002Fdiv>\n\n\u003Cdiv align=\"center\">\n\n⚡[**HowTo**](#-HOWTOs) **|** 🔧[**Installation**](docs\u002FINSTALL.md) **|** 💻[**Training Commands**](docs\u002FTrainTest.md) **|** 🐢[**DatasetPrepare**](docs\u002FDatasetPreparation.md) **|** 🏰[**Model Zoo**](docs\u002FModelZoo.md)\n\n📕[**中文解读文档**](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR-docs) **|** 📊 [**Plot scripts**](scripts\u002Fplot) **|** 📝[Introduction](docs\u002Fintroduction.md) **|** \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Ftree\u002Fmaster\u002Fcolab\">\u003Cimg src=\"https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg\" height=\"18\" alt=\"google colab logo\">\u003C\u002Fa> **|** ⏳[TODO List](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fprojects) **|** ❓[FAQ](docs\u002FFAQ.md)\n\u003C\u002Fdiv>\n\n🚀 我们新增了 [BasicSR-Examples](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR-examples)，其中提供了将 BasicSR 作为 Python 包使用的指导和模板。🚀 \u003Cbr>\n📢 **技术交流QQ群**：**320960100**   入群答案：**互帮互助共同进步** \u003Cbr>\n🧭 [入群二维码](#-contact) (QQ、微信)    [入群指南 (腾讯文档)](https:\u002F\u002Fdocs.qq.com\u002Fdoc\u002FDYXBSUmxOT0xBZ05u) \u003Cbr>\n\n---\n\nBasicSR (**Basic** **S**uper **R**estoration) 是一个基于 PyTorch 的开源 图像视频复原工具箱, 比如 超分辨率, 去噪, 去模糊, 去 JPEG 压缩噪声等.\u003Cbr>\nBasicSR (**Basic** **S**uper **R**estoration) 是一个基于 PyTorch 的开源 **图像和视频复原** 工具箱，例如超分辨率、去噪、去模糊、去除 JPEG 压缩伪影等。\n\n🚩 **新功能\u002F更新**\n\n- ✅ 2022年7月26日。新增绘图脚本 📊[Plot](scripts\u002Fplot)。\n- ✅ 2022年5月9日。BasicSR 加入 [XPixel](http:\u002F\u002Fxpixel.group\u002F)。\n- ✅ 2021年10月5日。新增 **ECBSR 训练与测试** 代码：[ECBSR](https:\u002F\u002Fgithub.com\u002Fxindongzhang\u002FECBSR)。\n  > ACMMM21：面向边缘的卷积块，用于移动设备上的实时超分辨率\n- ✅ 2021年9月2日。新增 **SwinIR 训练与测试** 代码：[SwinIR](https:\u002F\u002Fgithub.com\u002FJingyunLiang\u002FSwinIR) 由 [Jingyun Liang](https:\u002F\u002Fgithub.com\u002FJingyunLiang) 开发。更多详情请参见 [HOWTOs.md](docs\u002FHOWTOs.md#how-to-train-swinir-sr)。\n- ✅ 2021年8月5日。新增 NIQE，其结果与 MATLAB 完全一致（对于 tests\u002Fdata\u002Fbaboon.png，两者均为 5.7296）。\n- ✅ 2021年7月31日。新增 **双向视频超分辨率** 代码：[**BasicVSR** 和 IconVSR](https:\u002F\u002Farxiv.org\u002Fabs\u002F2012.02181)。\n  > CVPR21：BasicVSR：寻找视频超分辨率及其他领域的关键组件\n- **[更多](docs\u002Fhistory_updates.md)**\n\n---\n\n如果 BasicSR 对您的研究或工作有所帮助，请为本仓库点赞或推荐给您的朋友。谢谢😊 \u003Cbr>\n其他推荐项目：\u003Cbr>\n▶️ [Real-ESRGAN](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FReal-ESRGAN)：一种用于通用图像修复的实用算法\u003Cbr>\n▶️ [GFPGAN](https:\u002F\u002Fgithub.com\u002FTencentARC\u002FGFPGAN)：一种用于真实场景下人脸修复的实用算法\u003Cbr>\n▶️ [facexlib](https:\u002F\u002Fgithub.com\u002Fxinntao\u002Ffacexlib)：提供多种人脸相关功能的集合。\u003Cbr>\n▶️ [HandyView](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FHandyView)：基于 PyQt5 的图像查看器，便于查看和对比。\u003Cbr>\n▶️ [HandyFigure](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FHandyFigure)：论文图表的开源资源。\u003Cbr>\n\u003Csub>([ESRGAN](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FESRGAN), [EDVR](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FEDVR), [DNI](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FDNI), [SFTGAN](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FSFTGAN))\u003C\u002Fsub>\n\u003Csub>([HandyCrawler](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FHandyCrawler), [HandyWriting](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FHandyWriting))\u003C\u002Fsub>\n\n---\n\n## ⚡ 使用指南\n\n我们提供了简单的流程来训练\u002F测试\u002F推理模型，以便快速上手。\n这些流程\u002F命令并不能涵盖所有情况，更多细节请参阅后续章节。\n\n| GAN                  |                                                |                                                        |          |                                                |                                                        |\n| :------------------- | :--------------------------------------------: | :----------------------------------------------------: | :------- | :--------------------------------------------: | :----------------------------------------------------: |\n| StyleGAN2            | [训练](docs\u002FHOWTOs.md#如何训练StyleGAN2)       | [推理](docs\u002FHOWTOs.md#如何推理StyleGAN2)               |          |                                                |                                                        |\n| **人脸修复**         |                                                |                                                        |          |                                                |                                                        |\n| DFDNet               |                       -                        |  [推理](docs\u002FHOWTOs.md#如何推理DFDNet)                 |          |                                                |                                                        |\n| **超分辨率**         |                                                |                                                        |          |                                                |                                                        |\n| ESRGAN               |                     *待办*                     |                         *待办*                         | SRGAN    |                     *待办*                     |                         *待办*                         |\n| EDSR                 |                     *待办*                     |                         *待办*                         | SRResNet |                     *待办*                     |                         *待办*                         |\n| RCAN                 |                     *待办*                     |                         *待办*                         | SwinIR   | [训练](docs\u002FHOWTOs.md#如何训练SwinIR超分辨率) | [推理](docs\u002FHOWTOs.md#如何推理SwinIR超分辨率)         |\n| EDVR                 |                     *待办*                     |                         *待办*                         | DUF      |                       -                        |                         *待办*                         |\n| BasicVSR             |                     *待办*                     |                         *待办*                         | TOF      |                       -                        |                         *待办*                         |\n| **去模糊**           |                                                |                                                        |          |                                                |                                                        |\n| DeblurGANv2          |                       -                        |                         *待办*                         |          |                                                |                                                        |\n| **去噪**             |                                                |                                                        |          |                                                |                                                        |\n| RIDNet               |                       -                        |                         *待办*                         | CBDNet   |                       -                        |                         *待办*                         |\n\n## ✨ **使用BasicSR的项目**\n\n- [**Real-ESRGAN**](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FReal-ESRGAN): 一种用于通用图像修复的实用算法\n- [**GFPGAN**](https:\u002F\u002Fgithub.com\u002FTencentARC\u002FGFPGAN): 一种用于真实场景下人脸修复的实用算法\n\n如果您在开源项目中使用了`BasicSR`，欢迎通过[邮件](#-联系)或提交问题\u002F拉取请求与我联系。我会将您的项目添加到上述列表中 😊\n\n## 📜 许可证与致谢\n\n本项目采用[Apache 2.0许可证](LICENSE.txt)发布。\u003Cbr>\n关于**许可证**和**致谢**的更多详情，请参阅[LICENSE](LICENSE\u002FREADME.md)。\n\n## 🌏 引用\n\n如果BasicSR对您的研究或工作有所帮助，请引用BasicSR。\u003Cbr>\n以下是BibTeX参考文献。该BibTeX条目需要`url` LaTeX包。\n\n``` latex\n@misc{basicsr,\n  author =       {Xintao Wang and Liangbin Xie and Ke Yu and Kelvin C.K. Chan and Chen Change Loy and Chao Dong},\n  title =        {{BasicSR}: 开源图像与视频修复工具箱},\n  howpublished = {\\url{https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR}},\n  year =         {2022}\n}\n```\n\n> Xintao Wang, Liangbin Xie, Ke Yu, Kelvin C.K. Chan, Chen Change Loy 和 Chao Dong. BasicSR：开源图像与视频修复工具箱。 \u003Chttps:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR>, 2022.\n\n## 📧 联系方式\n\n如有任何问题，请发送邮件至 `xintao.alpha@gmail.com` 或 `xintao.wang@outlook.com`。\n\n\u003Cbr>\n\n- **QQ群**: 扫描左侧二维码 或者 搜索QQ群号：320960100   入群答案：互帮互助共同进步\n- **微信群**: 我们的一群已满500人，二群也超过200人；进群可添加Liangbin的个人微信（右侧二维码），他会在空闲时将大家拉入群~\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FXPixelGroup_BasicSR_readme_8f870ded4b5a.jpg\"  height=\"300\">   \n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FXPixelGroup_BasicSR_readme_ffad80acb2c3.png\"  height=\"300\">\n\u003C\u002Fp>\n\n![访问者](https:\u002F\u002Fvisitor-badge.glitch.me\u002Fbadge?page_id=XPixelGroup\u002FBasicSR) (自2022年11月6日起)","# BasicSR 快速上手指南\n\n## 简介\nBasicSR (Basic Super Restoration) 是一个基于 PyTorch 的开源图像和视频复原工具箱，支持超分辨率、去噪、去模糊、JPEG  artifacts 移除等多种任务。\n\n## 环境准备\n在开始之前，请确保您的开发环境满足以下要求：\n- **操作系统**：Linux \u002F Windows \u002F macOS\n- **编程语言**：Python 3.7+\n- **深度学习框架**：PyTorch (建议配合 CUDA 使用以加速训练和推理)\n- **依赖库**：OpenCV, LMDB, Pillow 等（安装时会自动处理）\n\n## 安装步骤\n\n### 方式一：通过 PyPI 安装（推荐）\n适用于直接使用预构建包的用户。建议使用国内镜像源加速下载：\n\n```bash\npip install basicsr -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n```\n\n### 方式二：从源码安装（适合开发者）\n如果您需要修改代码或进行二次开发，建议克隆仓库。若 GitHub 访问缓慢，可使用 Gitee 镜像：\n\n```bash\n# 使用 Gitee 镜像克隆\ngit clone https:\u002F\u002Fgitee.com\u002Fmirrors\u002FBasicSR.git\ncd BasicSR\n\n# 安装依赖\npip install -r requirements.txt\npip install -e .\n```\n\n## 基本使用\n\nBasicSR 提供了标准化的训练、测试和推理流程。具体模型的使用命令请参考官方文档中的 [HOWTOs](docs\u002FHOWTOs.md)。\n\n### 1. 训练模型\n使用配置文件启动训练任务：\n\n```bash\npython scripts\u002Ftrain.py -opt options\u002FModelName\u002Ftrain_ModelName.yml\n```\n\n### 2. 测试与推理\n加载预训练模型进行图像或视频复原：\n\n```bash\npython scripts\u002Ftest.py -opt options\u002FModelName\u002Ftest_ModelName.yml\n```\n\n### 3. 作为 Python 包使用\n如果您希望在项目中直接调用 BasicSR 的功能，可参考 [BasicSR-Examples](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR-examples) 获取模板和指引。\n\n---\n> **提示**：更多详细文档（如数据集准备、模型列表、常见问题）请访问项目根目录下的 `docs` 文件夹。","某数字媒体工作室接到客户委托，急需修复一批分辨率低、噪点多且模糊的 90 年代家庭录像带，以便制作高清纪念纪录片。\n\n### 没有 BasicSR 时\n- 需分别寻找超分、去噪和去模糊的独立代码库，各库依赖冲突导致环境配置频繁报错。\n- 缺乏统一接口，不同算法模型难以串联成自动化处理流水线，人工干预成本高。\n- 从头训练模型耗时过长，且缺乏基准对比，无法在客户要求的期限内交付成果。\n- 旧视频帧间闪烁严重，传统单帧处理方法难以保持时间维度的一致性与流畅度。\n\n### 使用 BasicSR 后\n- 基于 PyTorch 的统一框架直接调用 ESRGAN、BasicVSR 等成熟模型，无需重复造轮子。\n- 提供标准化的训练与测试命令及配置文件，轻松构建端到端的视频增强工作流。\n- 内置大量预训练权重支持快速推理，显著缩短项目周期并满足严格的交付时效。\n- 利用双向视频超分技术有效抑制闪烁，输出画质稳定且细节丰富的高清影像，提升客户满意度。\n\nBasicSR 通过模块化设计与丰富的模型库，让专业级的图像视频复原变得高效且可落地。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FXPixelGroup_BasicSR_0455eaad.png","XPixelGroup","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FXPixelGroup_f828c3b4.jpg","Our mission is to make the world look clearer and better!",null,"xpixelgroup@outlook.com","http:\u002F\u002Fxpixel.group\u002F","https:\u002F\u002Fgithub.com\u002FXPixelGroup",[83,87,91,95,99],{"name":84,"color":85,"percentage":86},"Python","#3572A5",88.7,{"name":88,"color":89,"percentage":90},"Cuda","#3A4E3A",6.3,{"name":92,"color":93,"percentage":94},"C++","#f34b7d",4.2,{"name":96,"color":97,"percentage":98},"MATLAB","#e16737",0.6,{"name":100,"color":101,"percentage":102},"Shell","#89e051",0.1,8196,1402,"2026-04-03T08:06:11","Apache-2.0","未说明",{"notes":109,"python":107,"dependencies":110},"README 中未明确列出具体的系统环境参数，详细信息请参考 docs\u002FINSTALL.md 文档。该项目是一个基于 PyTorch 的图像和视频复原工具箱。",[111],"PyTorch",[13,14],[114,115,116,117,118,119,120,121,122,123,124,125,126,127,128],"basicsr","esrgan","edsr","rcan","edvr","srresnet","srgan","super-resolution","restoration","pytorch","stylegan2","dfdnet","basicvsr","swinir","ecbsr","2026-03-27T02:49:30.150509","2026-04-06T05:15:35.919708",[132,137,141,146,151,156],{"id":133,"question_zh":134,"answer_zh":135,"source_url":136},5107,"如何配置训练过程中自动保存检查点的频率？","在 yml 配置文件中使用 `save_checkpoint_freq` 键。例如设置为 `!!float 5e3` 表示每 5000 次迭代保存一次模型和训练状态。设置该参数后，程序会自动保存训练状态，无需额外手动操作。","https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fissues\u002F64",{"id":138,"question_zh":139,"answer_zh":140,"source_url":136},5108,"如何正确恢复中断的训练任务而不从头开始？","应使用 `resume_state` 参数指定之前保存的 `.state` 文件路径（例如 `\"resume_state\": \"..\u002Fexperiments\u002Fmodel\u002Ftraining_state\u002F18000.state\"`）。如果同时保留 `pretrain_model_G` 配置，可能会导致模型从预训练权重重新开始而非继续之前的训练进度。",{"id":142,"question_zh":143,"answer_zh":144,"source_url":145},5109,"加载预训练模型时报错 KeyError: 'params' 怎么办？","此错误通常因模型文件格式不匹配引起。请改用项目提供的 `download_pretrained_models.py` 脚本下载官方模型，避免使用外部链接（如某些 Google Drive 旧版模型），以确保模型字典结构包含正确的 `params` 键。","https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fissues\u002F253",{"id":147,"question_zh":148,"answer_zh":149,"source_url":150},5110,"为什么我的测试结果 PSNR 比论文或官方数据低？","差异主要来自两方面：1) MATLAB 双三次下采样与 Python 实现存在细微数值差异；2) 图像裁剪方式不同（如 mod 8 边界处理）。建议统一测试标准（如下采样方法和裁剪规则）后再对比结果。","https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fissues\u002F28",{"id":152,"question_zh":153,"answer_zh":154,"source_url":155},5111,"找不到 logger.py 文件，如何查看 Tensorboard 训练曲线？","logger.py 已合并至 util.py 中。Tensorboard 是 TensorFlow 提供的独立工具，需参考 TensorFlow 官方文档配置日志监控。确保训练日志已正确输出到指定目录，然后通过 Tensorboard 命令启动即可。","https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fissues\u002F117",{"id":157,"question_zh":158,"answer_zh":159,"source_url":155},5112,"训练 SRGAN\u002FESRGAN 时如何保证稳定性及正确的预训练步骤？","建议先用 L1 Loss 对生成器 G 进行预训练（约 5000 次迭代直至 Loss 不再明显下降），再切换至 GAN 训练。若判别器 D Loss 震荡剧烈或为 0，说明 G 难以欺骗 D，可尝试调整超参数、不使用 D 预训练或更换损失函数（如 WGAN-GP）。",[161,166,171,176,181,186,191,196,201,206,211,216,221,226,231,236,241,246],{"id":162,"version":163,"summary_zh":164,"released_at":165},104632,"v1.4.2","🚀 See you again 😸\r\n\r\n✨ **Highlights**\r\n✅ [Add torch to setup_requires & dynamic import to prevent import error](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002F3974c3fb3701ff2295d5cecaf652c50f563c4b12)\r\n✅ [fix some codespell](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002F44672b0c6a607fc2a5aaeea70b9f565d1f3f095b)\r\n✅ [Add plot](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002Fb4f48db7d3207d7732d03446010ab2f0d1fdec25), you can find the scripts of drawing the paper figure. **Welcome to contribute your figure scripts**. Here is the [example](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Ftree\u002Fmaster\u002Fscripts\u002Fplot)\r\n\r\n\u003Cp align=\"center\">\r\n  \u003Cimg src=\"https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fblob\u002Fmaster\u002Fassets\u002Fplot\u002Fmodel_complexity_cmp_bsrn.png\" height=200>\r\n\u003C\u002Fp>\r\n\r\n\r\n\r\n\u003Cp align=\"center\">\r\n   \u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002FXPixelGroup\u002FBasicSR\u002Fmaster\u002Fassets\u002Fbasicsr_xpixel_logo.png\" height=150>\r\n\u003C\u002Fp>","2022-08-31T09:43:33",{"id":167,"version":168,"summary_zh":169,"released_at":170},104633,"v1.4.1","🚀 See you again 😸\r\n\r\n📢📢📢\r\n✅ Add Chinses docs: https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR-docs. Download the PDF here: https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR-docs\u002Freleases\r\n✅ BasicSR 的中文解读文档来啦: https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR-docs. 你可以从以下链接下载完整 PDF： https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR-docs\u002Freleases\r\n\r\n✨ **Highlights**\r\n✅ [fix bgr2ycbcr bug in color conversion](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002F85c1c0191eaebcc4819938fc9cddd1aca8609622)\r\n\r\n\u003Cp align=\"center\">\r\n   \u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002FXPixelGroup\u002FBasicSR\u002Fmaster\u002Fassets\u002Fbasicsr_xpixel_logo.png\" height=150>\r\n\u003C\u002Fp>","2022-07-18T10:48:46",{"id":172,"version":173,"summary_zh":174,"released_at":175},104634,"v1.4.0","🚀 Long time no see ☄️\r\n\r\nThis is the first time that BasicSR is released under the XPixelGroup :-)\r\n\r\n✨ **Highlights**\r\n✅ [Add training codes of BasicVSR++](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002Fab55f479f26ca28cd03426ef492f83fec4f4c68c)\r\n✅ [Add inference code of BasicVSR++](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002F786c5fd4d090b16d447d22bb5fc9e7fe7971d834)\r\n✅ [add psnr ssim pytorch version](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002F01a75fb9fcdc686b3af9f6602d392e6c6fae0b32)\r\n✅ [add LDL loss](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002F542507534c3e959f7598e6f74687738a930b9b7a)\r\n✅ Add online document: [basicsr.readthedocs.io\u002Fen\u002Flatest\u002F](https:\u002F\u002Fbasicsr.readthedocs.io\u002Fen\u002Flatest\u002F)\r\n✅ [update loss registry](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002Ffc36502cc406f3b23b97c3910483d95ce232625a)\r\n✅ [add realesrgan to basicsr](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002Fdd7d188b4adaf9b2b6db303e01b4de7c28d2d737)\r\n✅ [support registry with basicsr suffix](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR\u002Fcommit\u002F2f0ad00923db3370a9609cb722531da59925c7b5)\r\n\r\n📢📢📢\r\n我们正在准备 BasicSR 的📕[中文解读文档](https:\u002F\u002Fgithub.com\u002FXPixelGroup\u002FBasicSR-docs)。将在近期 release :-)\r\n\r\n\u003Cp align=\"center\">\r\n   \u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002FXPixelGroup\u002FBasicSR\u002Fmaster\u002Fassets\u002Fbasicsr_xpixel_logo.png\" height=150>\r\n\u003C\u002Fp>","2022-07-12T11:04:36",{"id":177,"version":178,"summary_zh":179,"released_at":180},104635,"v1.3.5","Have a nice day 😸 and happy everyday 😃\r\n\r\nI am happy to add a simple logo to BasicSR 😋 (designed by myself! and inspired by the [4K logo](https:\u002F\u002Fwww.google.com\u002Fsearch?q=4k+logo))\r\nSo I release a new version~\r\n\r\nYou can see it on [ReadMe](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR)\r\n\r\n(I know that there are still a lot of things to improve in BasicSR =-= I will put more time on it~ )\r\n\r\n\u003Cp align=\"center\">\r\n  \u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fxinntao\u002FBasicSR\u002Fmaster\u002Fassets\u002Fbasicsr_logo.png\" height=150>\r\n\u003C\u002Fp>","2022-02-15T16:30:52",{"id":182,"version":183,"summary_zh":184,"released_at":185},104636,"v1.3.4.9","🚀 See you again ☄️\r\n\r\nThis is a minor update. It is mainly for the [release of Real-ESRGAN](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FReal-ESRGAN\u002Freleases\u002Ftag\u002Fv0.2.3.0) 😄 , in which we release small models for anime videos.\r\n\r\n\r\n✨✨✨ Highlight\r\n\r\n- fix bug in AvgTimer: https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fcommit\u002Fc05b52798fe29dbb45c839780c10e8a9cd2eebe5\r\n- add pytest unittest framework","2021-12-12T12:38:31",{"id":187,"version":188,"summary_zh":189,"released_at":190},104637,"v1.3.4.4","🚀 See you again ☄️ \r\n\r\n\r\n✨✨✨ Highlight\r\n\r\n\r\n✅ [New method] Add ECBSR training and testing codes. (#478)\r\n> ACMMM21: Edge-oriented Convolution Block for Real-time Super Resolution on Mobile Devices \u003Cbe>\r\n> Xindong Zhang, Hui Zeng, Lei Zhang \r\n\r\n✅ [Features] Support **multiple validation** datasets and also print the **best metric value** results. \r\n\r\n> Example: \r\n\u003Cp align=\"center\">\r\n   \u003Cimg src=\"https:\u002F\u002Fuser-images.githubusercontent.com\u002F17445847\u002F136012453-43a4a571-131c-4a67-a3e7-66e13f58718b.png\" height=\"280\">\u003C\u002Fp>\r\n\r\n\r\n✅ [Bug fix] `metric_data` is not initialized when no metric is used during validation. See [Here](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fcommit\u002F2a590e92c01be5fe007ff0ac10a1b968e401839d)\r\n","2021-10-05T11:17:16",{"id":192,"version":193,"summary_zh":194,"released_at":195},104638,"v1.3.4.3","🚀 Long time no see 😹 \r\n\r\n\r\n✨✨✨ **Highlight**\r\n\r\n✅ [Features] Support multiple inputs for **metrics during validation** (#467)\r\n✅ [Bug fix] Fix bug in option: `force_yml` sometimes cannot be correctly modified \r\n✅ [Bug fix] Fix bug in redsrecurrentdataset: support interval argument (#463)\r\n✅ [Enhancement] add codespell hook, fix typos discovered by codespell\r\n\r\n\r\n📢 📢 📢 \r\n\r\n建立了 BasicSR交流讨论的 QQ群和微信群：\r\n\r\n技术交流QQ群：**320960100**    入群答案：**互帮互助共同进步**\r\n\r\n:loudspeaker: [入群指南 (腾讯文档)](https:\u002F\u002Fdocs.qq.com\u002Fdoc\u002FDYXBSUmxOT0xBZ05u)\r\n\r\n:compass: 入群二维码\r\n\u003Cp align=\"center\">\r\n  \u003Cimg src=\"https:\u002F\u002Fuser-images.githubusercontent.com\u002F17445847\u002F134879983-6f2d663b-16e7-49f2-97e1-7c53c8a5f71a.jpg\"  height=\"300\">      \r\n  \u003Cimg src=\"https:\u002F\u002Fuser-images.githubusercontent.com\u002F17445847\u002F134880057-f08e3d3b-2ab1-4ae8-966d-5753fe1f402a.png\"  height=\"300\">\r\n\u003C\u002Fp>\r\n","2021-09-27T11:15:26",{"id":197,"version":198,"summary_zh":199,"released_at":200},104639,"v1.3.4.2","🚀 Have a nice day 🐶 \r\n\r\nThis is a _minor_ release~\r\n\r\n✨ Highlight\r\n\r\n✅ [Enhancement] Support  the official **torchvision.ops.deform_conv2d** for torchvision>=0.9.0\r\n✅ [Enhancement] Add **`force_yml`** option. You can force to change the option yml options in the command line.  Examples: `python basicsr\u002Ftrain.py -opt options\u002Ftrain\u002FSRResNet_SRGAN\u002Ftrain_MSRResNet_x4.yml -train:ema_decay=0.999`\r\n✅ [Enhancement] **Copy the option yml** file to the experiment folder\r\n✅ [Enhancement] Add AvgTimer. The data time and iteration time in the logging are averaged in 200 iterations. \r\n✅ [Enhancement] Add `persistent_workers` option to dataloader\r\n✅ [Enhancement] Add vscode settings. \r\n✅ [Enhancement] Improve some formats","2021-09-12T16:51:47",{"id":202,"version":203,"summary_zh":204,"released_at":205},104640,"v1.3.4.1","🚀  See you again 😸 \r\n\r\n✨ Highlight\r\n\r\n- ✅ Add **SwinIR training and testing** codes\r\n     - [SwinIR](https:\u002F\u002Fgithub.com\u002FJingyunLiang\u002FSwinIR) by [Jingyun Liang](https:\u002F\u002Fgithub.com\u002FJingyunLiang) 👍 Thanks for their great work~\r\n     - More details are in [HOWTOs.md](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002Fdocs\u002FHOWTOs.md#how-to-train-swinir-sr)\r\n- ✅ All models in BasicSR support `EMA (Exponential Moving Average)`.","2021-09-02T12:42:09",{"id":207,"version":208,"summary_zh":209,"released_at":210},104641,"v1.3.4.0","🚀\r\n\r\nA lot of things have been improved since the last release note.\r\n\r\n✨ Highlight\r\n1. We add the official training and testing codes of [BasicVSR](https:\u002F\u002Farxiv.org\u002Fabs\u002F2012.02181) - a video SR method \r\n2. We improve NIQE metric. Now, the python version of NIQE could generate almost the same results as MATLAB.\r\n3. Add [HiFaceGAN](https:\u002F\u002Fgithub.com\u002FLotayou\u002FFace-Renovation) codes by [Lotayou](https:\u002F\u002Flotayou.github.io\u002F).\r\n4. Add more [degradation](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002Fbasicsr\u002Fdata\u002Fdegradations.py)\r\n5. Add [BasicSR-Examples](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR-examples), which provides guidance and templates of using BasicSR as a python package.\r\n6. Fixed many bugs: such as no logging in pt18, resume bugs, etc\r\n......\r\n\r\nWe will spend more time on solving issues and reviewing the pull request.\r\nWelcome to your contributions 😄 \r\n\r\n🚀🚀🚀\r\n","2021-08-29T14:17:17",{"id":212,"version":213,"summary_zh":214,"released_at":215},104642,"v1.3.3.1","🚀\r\n\r\n✨ Highlight\r\n\r\nA minor version for quick bug fixed. \r\n#396\r\n#397","2021-05-25T07:44:00",{"id":217,"version":218,"summary_zh":219,"released_at":220},104643,"v1.3.3","🚀\r\n\r\n:sparkles: **Highlights** \r\n\r\nWe reorganize the BasicSR codes. It may be incompatible with the previous v1.2.0\r\n- Add registry mechanism\r\n- Support pip install\r\n- Support JIT CUDA ops\r\n- Can be easily used as an external package to develop your own project (example project is coming soon)\r\n- Update format: change max line length to 120\r\n- Add degradations (data utils)\r\n\r\nThis major change may introduce bugs. If you encounter bugs, please let me know. Thanks!","2021-05-23T10:44:46",{"id":222,"version":223,"summary_zh":224,"released_at":225},104644,"v1.2.0","🚀\r\n\r\n:sparkles: **Highlights** \r\n\r\n- Add ESRGAN and DFDNet [colab](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Ftree\u002Fmaster\u002Fcolab)\r\n- Add FID and LPIPS metrics\r\n- Add matlab imresize bicubic (#317)\r\n- README add datasets download links (#318)\r\n\r\n:bug: **Bug Fixes**\r\n- PSNR and SSIM calculation on uint8 type\r\n- Fix metrics bug in video_base_model.py (#314)\r\n\r\n:palm_tree: **Improvements**\r\n- Reorganize code structure and remove unnessary packages\r\n- `tensor2img` support gray images\r\n- Refactor DFDNet codes\r\n","2020-11-29T06:14:32",{"id":227,"version":228,"summary_zh":229,"released_at":230},104645,"v1.1.1","🚀\r\n\r\n:sparkles: **Highlights** \r\n\r\n- Add Baidu Drive (百度网盘) download links\r\n- Add funny emoji :relaxed:\r\n\r\n:bug: **Bug Fixes**\r\n- bgr2rgb type conversion in stylegan2 model\r\n- Supporting training w\u002Fo validation\r\n\r\n:palm_tree: **Improvements**\r\n- `download_pretrained_models.py` script supports downloading all the models\r\n- Refactor `define_network` functions\r\n","2020-09-18T16:02:50",{"id":232,"version":233,"summary_zh":234,"released_at":235},104646,"v1.1.0","Hope all is well 🚀 \r\n\r\n**Highlights**\r\n- Add [DFDNet](https:\u002F\u002Fgithub.com\u002Fcsxmli2016\u002FDFDNet) inference codes (ECCV20: Blind Face Restoration via Deep Multi-scale Component Dictionaries)\r\n- Add more official StyleGAN2 pretrained models: [Model Zoo](https:\u002F\u002Fdrive.google.com\u002Fdrive\u002Ffolders\u002F15DgDtfaLASQ3iAPJEVHQF49g9msexECG?usp=sharing)\r\n- Add *New Feature* section in [README](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR).\r\n\r\n**Bug Fixes**\r\n- PyTorch 1.6 uses a new serialization for torch.save. The saved model cannot be loaded by the previous PyTorch version. We updated the `publish_models.py` with `_use_new_zipfile_serialization=False`. [More details](https:\u002F\u002Fpytorch.org\u002Fdocs\u002Fstable\u002Fgenerated\u002Ftorch.save.html?highlight=save#torch.save).\r\n","2020-09-08T13:05:13",{"id":237,"version":238,"summary_zh":239,"released_at":240},104647,"v1.0.1","Hope all is well 🚀 \r\n\r\n**Highlights**\r\n- Add StyleGAN2 training and testing codes. Pretrained models are [here](https:\u002F\u002Fdrive.google.com\u002Fdrive\u002Ffolders\u002F15DgDtfaLASQ3iAPJEVHQF49g9msexECG?usp=sharing).\r\n- Fix bug: cuda prefetcher return none twice.\r\n- Add HOWTOs for quick starts.","2020-08-27T17:31:45",{"id":242,"version":243,"summary_zh":244,"released_at":245},104648,"v1.0.0","We will use `releases` to manage BasicSR :smile: \r\n\r\nHope all is well 🚀 \r\n\r\nThis is a brand-new version of `BasicSR`. We have re-organized all the codes and frameworks. \r\n\r\n**Highlights**\r\n- We use [*Dynamic Instantiation*](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002Fdocs\u002FDesignConvention.md#Dynamic-Instantiation) for creating datasets, architectures, and models. So it is easier and more friendly to develop your own algorithms. \r\n- We provide [richer documents](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Ftree\u002Fmaster\u002Fdocs). At the same time, we also provide a Chinese version (同时也提供了中文版本的文档说明).\r\n- We provide more [pre-trained models, training examples](https:\u002F\u002Fgithub.com\u002Fxinntao\u002FBasicSR\u002Fblob\u002Fmaster\u002Fdocs\u002FModelZoo.md). We also upload the training process and curves to [wandb](https:\u002F\u002Fapp.wandb.ai\u002Fxintao\u002Fbasicsr).\r\n- Currently, it supports: \r\n    - Training: EDSR, EDVR, ESRGAN, SRResNet, SRGAN\r\n    - Testing: DUF, EDSR, EDVR, ESRGAN, RCAN, SRResNt, SRGAN, and TOF.\r\n- We also mirror this codebase to *[Gitee码云](https:\u002F\u002Fgitee.com\u002Fxinntao\u002FBasicSR)* for easy access of Chinese users. \r\n\r\n---\r\n\r\nSorry that this version of BasicSR is not compatible with the previous versions. \r\n\r\nWe will add more features to this codebase. And welcome contribute, and report bugs! 😆 ","2020-08-19T14:56:50",{"id":247,"version":248,"summary_zh":78,"released_at":249},104649,"v0.0","2019-06-13T13:09:28"]