[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-Fafa-DL--Awesome-Backbones":3,"tool-Fafa-DL--Awesome-Backbones":61},[4,18,26,36,44,53],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},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,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},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",143909,2,"2026-04-07T11:33:18",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":32,"last_commit_at":42,"category_tags":43,"status":17},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。",107888,"2026-04-06T11:32:50",[14,15,13],{"id":45,"name":46,"github_repo":47,"description_zh":48,"stars":49,"difficulty_score":32,"last_commit_at":50,"category_tags":51,"status":17},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[52,14],"插件",{"id":54,"name":55,"github_repo":56,"description_zh":57,"stars":58,"difficulty_score":10,"last_commit_at":59,"category_tags":60,"status":17},4487,"LLMs-from-scratch","rasbt\u002FLLMs-from-scratch","LLMs-from-scratch 是一个基于 PyTorch 的开源教育项目，旨在引导用户从零开始一步步构建一个类似 ChatGPT 的大型语言模型（LLM）。它不仅是同名技术著作的官方代码库，更提供了一套完整的实践方案，涵盖模型开发、预训练及微调的全过程。\n\n该项目主要解决了大模型领域“黑盒化”的学习痛点。许多开发者虽能调用现成模型，却难以深入理解其内部架构与训练机制。通过亲手编写每一行核心代码，用户能够透彻掌握 Transformer 架构、注意力机制等关键原理，从而真正理解大模型是如何“思考”的。此外，项目还包含了加载大型预训练权重进行微调的代码，帮助用户将理论知识延伸至实际应用。\n\nLLMs-from-scratch 特别适合希望深入底层原理的 AI 开发者、研究人员以及计算机专业的学生。对于不满足于仅使用 API，而是渴望探究模型构建细节的技术人员而言，这是极佳的学习资源。其独特的技术亮点在于“循序渐进”的教学设计：将复杂的系统工程拆解为清晰的步骤，配合详细的图表与示例，让构建一个虽小但功能完备的大模型变得触手可及。无论你是想夯实理论基础，还是为未来研发更大规模的模型做准备",90106,"2026-04-06T11:19:32",[35,15,13,14],{"id":62,"github_repo":63,"name":64,"description_en":65,"description_zh":66,"ai_summary_zh":66,"readme_en":67,"readme_zh":68,"quickstart_zh":69,"use_case_zh":70,"hero_image_url":71,"owner_login":72,"owner_name":73,"owner_avatar_url":74,"owner_bio":75,"owner_company":76,"owner_location":76,"owner_email":76,"owner_twitter":76,"owner_website":76,"owner_url":77,"languages":78,"stars":83,"forks":84,"last_commit_at":85,"license":76,"difficulty_score":32,"env_os":86,"env_gpu":87,"env_ram":86,"env_deps":88,"category_tags":93,"github_topics":94,"view_count":32,"oss_zip_url":76,"oss_zip_packed_at":76,"status":17,"created_at":103,"updated_at":104,"faqs":105,"releases":141},5227,"Fafa-DL\u002FAwesome-Backbones","Awesome-Backbones","Integrate deep learning models for image classification | Backbone learning\u002Fcomparison\u002Fmagic modification project","Awesome-Backbones 是一个专为图像分类任务打造的深度学习模型集成项目，旨在帮助开发者轻松对比、训练和修改各类主流骨干网络（Backbone）。它解决了研究人员在复现经典模型时面临的环境配置繁琐、代码风格不统一以及超参数调整困难等痛点，提供了一个标准化的 PyTorch 实现框架。\n\n该项目非常适合计算机视觉领域的研究人员、算法工程师以及希望深入理解模型原理的开发者使用。无论是需要快速验证新想法，还是希望在特定数据集上微调现有模型，Awesome-Backbones 都能提供高效支持。其独特亮点在于不仅集成了从 MobileNet、EfficientNet 到 ViT、Swin Transformer 等数十种前沿架构，还贴心地提供了详细的训练调优指南，例如针对小数据集如何关闭可能污染数据的图像增强策略。此外，项目持续更新，近期已支持模型转 ONNX 格式、生成类别激活图（CAM）可视化，并能自动输出完整的训练与验证指标（如准确率、损失值等），极大地便利了实验分析与结果复现。通过统一的接口和清晰的文档，Awesome-Backbones 让模型探索变得更加简单高效。","Awesome backbones for image classification\n===========================\n\n\u003Cdiv align=\"center\">\n\n[![BILIBILI](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_readme_c7801df1a6d3.png)](https:\u002F\u002Fspace.bilibili.com\u002F46880349)\n\n![](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FAwesome%20Backbones-v0.6.3-brightgreen)\n![](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPyTorch-%3E%3Dv1.7.1-green)\n![](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPython-%3E%3Dv3.6-yellowgreen)\n[![GitHub forks](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fforks\u002FFafa-DL\u002FAwesome-Backbones)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones)\n[![GitHub stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FFafa-DL\u002FAwesome-Backbones)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones)\n\n\u003C\u002Fdiv>\n\n## 写在前面\n- 若训练效果不佳，首先需要调整学习率和Batch size，这俩超参很大程度上影响收敛。其次，从关闭图像增强手段（尤其小数据集）开始，有的图像增强方法会污染数据，如\n\n![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_readme_54d1539a62e8.jpg) ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_readme_c2cf1354964f.jpg) ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_readme_3e7dd95a32fb.jpg)\n\n  如何去除增强？如[efficientnetv2-b0](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fmodels\u002Fefficientnetv2\u002Fefficientnetv2_b0.py)配置文件中train_pipeline可更改为如下\n```yaml\ntrain_pipeline = [\n    dict(type='LoadImageFromFile'),\n    dict(\n        type='RandomResizedCrop',\n        size=192,\n        efficientnet_style=True,\n        interpolation='bicubic'),\n    dict(type='Normalize', **img_norm_cfg),\n    dict(type='ImageToTensor', keys=['img']),\n    dict(type='ToTensor', keys=['gt_label']),\n    dict(type='Collect', keys=['img', 'gt_label'])\n]\n```\n  若你的数据集提前已经将shape更改为网络要求的尺寸，那么`Resize`操作也可以去除。\n\n## 更新日志\n\n**`2025.01.17`** \n- 支持转ONNX[#136](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fpull\u002F136) @PurpleSky-NS\n- 类别激活图相关脚本[#114](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fpull\u002F114) @jackyjinjing\n\n**`2024.09.06`** \n- 修复高频反馈的**评估时结果浮动大**的问题\n\n**`2023.12.02`** \n- 新增Issue中多人提及的输出**Train Acc**与**Val loss**\n    - `metrics_outputs.csv`保存每周期`train_loss, train_acc, train_precision, train_recall, train_f1-score, val_loss, val_acc, val_precision, val_recall, val_f1-score`方便各位绘图\n    - 终端由原先仅输出**Val**相关metrics升级为Train与Val都输出\n\n    ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_readme_ba74ba9fe021.jpg)\n\n**`2023.08.05`** \n- 新增**TinyViT**(预训练权重不匹配)、**DeiT3**、**EdgeNeXt**、**RevVisionTransformer**\n\n**`2023.03.07`** \n- 新增**MobileViT**、**DaViT**、**RepLKNet**、**BEiT**、**EVA**、**MixMIM**、**EfficientNetV2**\n\n\n## 测试环境\n\n- Pytorch      1.7.1+\n- Python       3.6+\n\n## 资料\n|数据集|视频教程|人工智能技术探讨群|\n|---|---|---|\n|[`花卉数据集` 提取码：0zat](https:\u002F\u002Fpan.baidu.com\u002Fs\u002F1137y4l-J3AgyCiC_cXqIqw)|[点我跳转](https:\u002F\u002Fwww.bilibili.com\u002Fvideo\u002FBV1SY411P7Nd)|[1群：78174903](https:\u002F\u002Fjq.qq.com\u002F?_wv=1027&k=lY5KVICA)\u003Cbr\u002F>[3群：584723646](https:\u002F\u002Fjq.qq.com\u002F?_wv=1027&k=bakez5Yz)\n\n## 快速开始\n\n- 遵循[环境搭建](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FEnvironment_setting.md)完成配置\n- 下载[MobileNetV3-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilenet_v3\u002Fconvert\u002Fmobilenet_v3_small-8427ecf0.pth)权重至**datas**下\n- **Awesome-Backbones**文件夹下终端输入\n```bash\npython tools\u002Fsingle_test.py datas\u002Fcat-dog.png models\u002Fmobilenet\u002Fmobilenet_v3_small.py --classes-map datas\u002FimageNet1kAnnotation.txt\n```\n\n## 教程\n- [环境搭建](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FEnvironment_setting.md)\n- [数据集准备](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FData_preparing.md)\n- [配置文件解释](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FConfigs_description.md)\n- [训练](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FHow_to_train.md)\n- [模型评估](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FHow_to_eval.md)\n- [计算Flops&Params](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FCalculate_Flops.md)\n- [添加新的模型组件](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FAdd_modules.md)\n- [类别激活图可视化](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FCAM_visualization.md)\n- [学习率策略可视化](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FLr_visualization.md)\n- [数据增强策略可视化](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FPipeline_visualization.md)\n\n## 模型\n- [x] [LeNet5](https:\u002F\u002Fieeexplore.ieee.org\u002Fdocument\u002F6795724)\n- [x] [AlexNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124461111)\n- [x] [VGG](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124477080)\n- [x] [DenseNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124630832)\n- [x] [ResNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124477575)\n- [x] [Wide-ResNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124754437)\n- [x] [ResNeXt](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124477919)\n- [x] [SEResNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124478157)\n- [x] [SEResNeXt](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124478347)\n- [x] [RegNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124478426)\n- [x] [MobileNetV2](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124478681)\n- [x] [MobileNetV3](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124478770)\n- [x] [ShuffleNetV1](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124479156)\n- [x] [ShuffleNetV2](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124479336)\n- [x] [EfficientNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124754493)\n- [x] [RepVGG](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124479644)\n- [x] [Res2Net](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124479467)\n- [x] [ConvNeXt](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124481466)\n- [x] [HRNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124481590)\n- [x] [ConvMixer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124481766)\n- [x] [CSPNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124481930)\n- [x] [Swin-Transformer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124538198)\n- [x] [Vision-Transformer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124567953)\n- [x] [Transformer-in-Transformer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596023)\n- [x] [MLP-Mixer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596093)\n- [x] [DeiT](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124591888)\n- [x] [Conformer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596343)\n- [x] [T2T-ViT](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596425)\n- [x] [Twins](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596619)\n- [x] [PoolFormer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596740)\n- [x] [VAN](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124630541)\n- [x] [HorNet](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2207.14284v2.pdf)\n- [x] [EfficientFormer](https:\u002F\u002Farxiv.org\u002Fabs\u002F2206.01191)\n- [x] [Swin Transformer V2](https:\u002F\u002Farxiv.org\u002Fabs\u002F2111.09883.pdf)\n- [x] [MViT V2](http:\u002F\u002Fopenaccess.thecvf.com\u002F\u002Fcontent\u002FCVPR2022\u002Fpapers\u002FLi_MViTv2_Improved_Multiscale_Vision_Transformers_for_Classification_and_Detection_CVPR_2022_paper.pdf)\n- [x] [MobileViT](https:\u002F\u002Farxiv.org\u002Fabs\u002F2110.02178)\n- [x] [DaViT](https:\u002F\u002Farxiv.org\u002Fabs\u002F2204.03645v1)\n- [x] [replknet](https:\u002F\u002Farxiv.org\u002Fabs\u002F2203.06717)\n- [x] [BEiT](https:\u002F\u002Farxiv.org\u002Fabs\u002F2106.08254)\n- [x] [EVA](https:\u002F\u002Farxiv.org\u002Fabs\u002F2211.07636)\n- [x] [MixMIM](https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.13137)\n- [x] [EfficientNetV2](https:\u002F\u002Farxiv.org\u002Fabs\u002F2104.00298)\n## 预训练权重\n\n| 名称 | 权重 | 名称 | 权重 | 名称 | 权重 |\n| :-----: | :-----: | :------: | :------: | :------: | :-----: |\n| **LeNet5** | None | **AlexNet** | None | **VGG** | [VGG-11](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg11_batch256_imagenet_20210208-4271cd6c.pth)\u003Cbr\u002F>[VGG-13](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg13_batch256_imagenet_20210208-4d1d6080.pth)\u003Cbr\u002F>[VGG-16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg16_batch256_imagenet_20210208-db26f1a5.pth)\u003Cbr\u002F>[VGG-19](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg19_batch256_imagenet_20210208-e6920e4a.pth)\u003Cbr\u002F>[VGG-11-BN](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg11_bn_batch256_imagenet_20210207-f244902c.pth)\u003Cbr\u002F>[VGG-13-BN](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg13_bn_batch256_imagenet_20210207-1a8b7864.pth)\u003Cbr\u002F>[VGG-16-BN](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg16_bn_batch256_imagenet_20210208-7e55cd29.pth)\u003Cbr\u002F>[VGG-19-BN](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg19_bn_batch256_imagenet_20210208-da620c4f.pth)|\n| **ResNet** |[ResNet-18](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnet18_8xb32_in1k_20210831-fbbb1da6.pth)\u003Cbr\u002F>[ResNet-34](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnet34_8xb32_in1k_20210831-f257d4e6.pth)\u003Cbr\u002F>[ResNet-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnet50_8xb32_in1k_20210831-ea4938fc.pth)\u003Cbr\u002F>[ResNet-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnet101_8xb32_in1k_20210831-539c63f8.pth)\u003Cbr\u002F>[ResNet-152](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnet152_8xb32_in1k_20210901-4d7582fa.pth) | **ResNetV1C** | [ResNetV1C-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1c50_8xb32_in1k_20220214-3343eccd.pth)\u003Cbr\u002F>[ResNetV1C-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1c101_8xb32_in1k_20220214-434fe45f.pth)\u003Cbr\u002F>[ResNetV1C-152](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1c152_8xb32_in1k_20220214-c013291f.pth) |**ResNetV1D** | [ResNetV1D-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1d50_b32x8_imagenet_20210531-db14775a.pth)\u003Cbr\u002F>[ResNetV1D-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1d101_b32x8_imagenet_20210531-6e13bcd3.pth)\u003Cbr\u002F>[ResNetV1D-152](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1d152_b32x8_imagenet_20210531-278cf22a.pth) |\n| **ResNeXt** | [ResNeXt-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnext\u002Fresnext50_32x4d_b32x8_imagenet_20210429-56066e27.pth)\u003Cbr\u002F>[ResNeXt-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnext\u002Fresnext101_32x4d_b32x8_imagenet_20210506-e0fa3dd5.pth)\u003Cbr\u002F>[ResNeXt-152](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnext\u002Fresnext152_32x4d_b32x8_imagenet_20210524-927787be.pth) | **SEResNet** | [SEResNet-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fse-resnet\u002Fse-resnet50_batch256_imagenet_20200804-ae206104.pth)\u003Cbr\u002F>[SEResNet-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fse-resnet\u002Fse-resnet101_batch256_imagenet_20200804-ba5b51d4.pth)| **SEResNeXt**| None|\n| **RegNet** |[RegNetX-400MF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-400mf_8xb128_in1k_20211213-89bfc226.pth)\u003Cbr\u002F>[RegNetX-800MF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-800mf_8xb128_in1k_20211213-222b0f11.pth)\u003Cbr\u002F>[RegNetX-1.6GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-1.6gf_8xb128_in1k_20211213-d1b89758.pth)\u003Cbr\u002F>[RegNetX-3.2GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-3.2gf_8xb64_in1k_20211213-1fdd82ae.pth)\u003Cbr\u002F>[RegNetX-4.0GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-4.0gf_8xb64_in1k_20211213-efed675c.pth)\u003Cbr\u002F>[RegNetX-6.4GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-6.4gf_8xb64_in1k_20211215-5c6089da.pth)\u003Cbr\u002F>[RegNetX-8.0GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-8.0gf_8xb64_in1k_20211213-9a9fcc76.pth)\u003Cbr\u002F>[RegNetX-12GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-12gf_8xb64_in1k_20211213-5df8c2f8.pth) | **MobileNetV2** | [MobileNetV2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilenet_v2\u002Fmobilenet_v2_batch256_imagenet_20200708-3b2dc3af.pth) |**MobileNetV3** | [MobileNetV3-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilenet_v3\u002Fconvert\u002Fmobilenet_v3_small-8427ecf0.pth)\u003Cbr\u002F>[MobileNetV3-Large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilenet_v3\u002Fconvert\u002Fmobilenet_v3_large-3ea3c186.pth) |\n| **ShuffleNetV1** |[ShuffleNetV1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fshufflenet_v1\u002Fshufflenet_v1_batch1024_imagenet_20200804-5d6cec73.pth) | **ShuffleNetV2** | [ShuffleNetV2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fshufflenet_v2\u002Fshufflenet_v2_batch1024_imagenet_20200812-5bf4721e.pth) |**EfficientNet** | [EfficientNet-B0](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b0_3rdparty_8xb32_in1k_20220119-a7e2a0b1.pth)\u003Cbr\u002F>[EfficientNet-B1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b1_3rdparty_8xb32_in1k_20220119-002556d9.pth)\u003Cbr\u002F>[EfficientNet-B2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b2_3rdparty_8xb32_in1k_20220119-ea374a30.pth)\u003Cbr\u002F>[EfficientNet-B3](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b3_3rdparty_8xb32_in1k_20220119-4b4d7487.pth)\u003Cbr\u002F>[EfficientNet-B4](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b4_3rdparty_8xb32_in1k_20220119-81fd4077.pth)\u003Cbr\u002F>[EfficientNet-B5](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b5_3rdparty_8xb32_in1k_20220119-e9814430.pth)\u003Cbr\u002F>[EfficientNet-B6](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b6_3rdparty_8xb32-aa_in1k_20220119-45b03310.pth)\u003Cbr\u002F>[EfficientNet-B7](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b7_3rdparty_8xb32-aa_in1k_20220119-bf03951c.pth)\u003Cbr\u002F>[EfficientNet-B8](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b8_3rdparty_8xb32-aa-advprop_in1k_20220119-297ce1b7.pth) |\n| **RepVGG** |[RepVGG-A0](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-A0_3rdparty_4xb64-coslr-120e_in1k_20210909-883ab98c.pth)\u003Cbr\u002F>[RepVGG-A1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-A1_3rdparty_4xb64-coslr-120e_in1k_20210909-24003a24.pth) \u003Cbr\u002F>[RepVGG-A2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-A2_3rdparty_4xb64-coslr-120e_in1k_20210909-97d7695a.pth)\u003Cbr\u002F>[RepVGG-B0](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B0_3rdparty_4xb64-coslr-120e_in1k_20210909-446375f4.pth)\u003Cbr\u002F>[RepVGG-B1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B1_3rdparty_4xb64-coslr-120e_in1k_20210909-750cdf67.pth)\u003Cbr\u002F>[RepVGG-A1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-A1_3rdparty_4xb64-coslr-120e_in1k_20210909-24003a24.pth)\u003Cbr\u002F>[RepVGG-B1g2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B1g2_3rdparty_4xb64-coslr-120e_in1k_20210909-344f6422.pth)\u003Cbr\u002F>[RepVGG-B1g4](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B1g4_3rdparty_4xb64-coslr-120e_in1k_20210909-d4c1a642.pth)\u003Cbr\u002F>[RepVGG-B2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B2_3rdparty_4xb64-coslr-120e_in1k_20210909-bd6b937c.pth)\u003Cbr\u002F>[RepVGG-B2g4](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B2g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-7b7955f0.pth)\u003Cbr\u002F>[RepVGG-B2g4](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B2g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-7b7955f0.pth)\u003Cbr\u002F>[RepVGG-B3](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B3_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-dda968bf.pth)\u003Cbr\u002F>[RepVGG-B3g4](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B3g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-4e54846a.pth)\u003Cbr\u002F>[RepVGG-D2se](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-D2se_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-cf3139b7.pth)| **Res2Net** | [Res2Net-50-14w-8s](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fres2net\u002Fres2net50-w14-s8_3rdparty_8xb32_in1k_20210927-bc967bf1.pth)\u003Cbr\u002F>[Res2Net-50-26w-8s](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fres2net\u002Fres2net50-w26-s8_3rdparty_8xb32_in1k_20210927-f547a94b.pth)\u003Cbr\u002F>[Res2Net-101-26w-4s](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fres2net\u002Fres2net101-w26-s4_3rdparty_8xb32_in1k_20210927-870b6c36.pth)\u003Cbr\u002F> |**ConvNeXt** | [ConvNeXt-Tiny](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvnext\u002Fconvnext-tiny_3rdparty_32xb128_in1k_20220124-18abde00.pth)\u003Cbr\u002F>[ConvNeXt-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvnext\u002Fconvnext-small_3rdparty_32xb128_in1k_20220124-d39b5192.pth)\u003Cbr\u002F>[ConvNeXt-Base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvnext\u002Fconvnext-base_in21k-pre-3rdparty_32xb128_in1k_20220124-eb2d6ada.pth)\u003Cbr\u002F>[ConvNeXt-Large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvnext\u002Fconvnext-large_in21k-pre-3rdparty_64xb64_in1k_20220124-2412403d.pth)\u003Cbr\u002F>[ConvNeXt-XLarge](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvnext\u002Fconvnext-xlarge_in21k-pre-3rdparty_64xb64_in1k_20220124-76b6863d.pth) |\n| **HRNet** |[HRNet-W18](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w18_3rdparty_8xb32_in1k_20220120-0c10b180.pth)\u003Cbr\u002F>[HRNet-W30](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w30_3rdparty_8xb32_in1k_20220120-8aa3832f.pth) \u003Cbr\u002F>[HRNet-W32](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w32_3rdparty_8xb32_in1k_20220120-c394f1ab.pth)\u003Cbr\u002F>[HRNet-W40](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w40_3rdparty_8xb32_in1k_20220120-9a2dbfc5.pth)\u003Cbr\u002F>[HRNet-W44](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w44_3rdparty_8xb32_in1k_20220120-35d07f73.pth)\u003Cbr\u002F>[HRNet-W48](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w48_3rdparty_8xb32_in1k_20220120-e555ef50.pth)\u003Cbr\u002F>[HRNet-W64](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w64_3rdparty_8xb32_in1k_20220120-19126642.pth) | **ConvMixer** | [ConvMixer-768\u002F32](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvmixer\u002Fconvmixer-768-32_3rdparty_10xb64_in1k_20220323-bca1f7b8.pth)\u003Cbr\u002F>[ConvMixer-1024\u002F20](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvmixer\u002Fconvmixer-1024-20_3rdparty_10xb64_in1k_20220323-48f8aeba.pth)\u003Cbr\u002F>[ConvMixer-1536\u002F20](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvmixer\u002Fconvmixer-1536_20_3rdparty_10xb64_in1k_20220323-ea5786f3.pth) |**CSPNet** | [CSPDarkNet50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fcspnet\u002Fcspdarknet50_3rdparty_8xb32_in1k_20220329-bd275287.pth)\u003Cbr\u002F>[CSPResNet50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fcspnet\u002Fcspresnet50_3rdparty_8xb32_in1k_20220329-dd6dddfb.pth)\u003Cbr\u002F>[CSPResNeXt50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fcspnet\u002Fcspresnext50_3rdparty_8xb32_in1k_20220329-2cc84d21.pth) |\n|**Swin Transformer**|[tiny-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fswin_tiny_224_b16x64_300e_imagenet_20210616_090925-66df6be6.pth)\u003Cbr\u002F>[small-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fswin_small_224_b16x64_300e_imagenet_20210615_110219-7f9d988b.pth)\u003Cbr\u002F>[base-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fswin_base_224_b16x64_300e_imagenet_20210616_190742-93230b0d.pth)\u003Cbr\u002F>[large-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fconvert\u002Fswin_large_patch4_window7_224_22kto1k-5f0996db.pth)\u003Cbr\u002F>[base-384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fconvert\u002Fswin_base_patch4_window12_384_22kto1k-d59b0d1d.pth)\u003Cbr\u002F>[large-384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fconvert\u002Fswin_large_patch4_window12_384_22kto1k-0a40944b.pth)|**Vision Transformer**|[vit_base_p16_224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Fpretrain\u002Fvit-base-p16_3rdparty_pt-64xb64_in1k-224_20210928-02284250.pth)\u003Cbr\u002F>[vit_base_p32_224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Fpretrain\u002Fvit-base-p32_3rdparty_pt-64xb64_in1k-224_20210928-eee25dd4.pth)\u003Cbr\u002F>[vit_large_p16_224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Fpretrain\u002Fvit-large-p16_3rdparty_pt-64xb64_in1k-224_20210928-0001f9a1.pth)\u003Cbr\u002F>[vit_base_p16_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Ffinetune\u002Fvit-base-p16_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-98e8652b.pth)\u003Cbr\u002F>[vit_base_p32_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Ffinetune\u002Fvit-base-p32_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-9cea8599.pth)\u003Cbr\u002F>[vit_large_p16_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Ffinetune\u002Fvit-large-p16_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-b20ba619.pth)|**Transformer in Transformer**|[TNT-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftnt\u002Ftnt-small-p16_3rdparty_in1k_20210903-c56ee7df.pth)|\n|**MLP Mixer**|[base_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmlp-mixer\u002Fmixer-base-p16_3rdparty_64xb64_in1k_20211124-1377e3e0.pth)\u003Cbr\u002F>[large_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmlp-mixer\u002Fmixer-large-p16_3rdparty_64xb64_in1k_20211124-5a2519d2.pth)|**Deit**|[DeiT-tiny](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-tiny_pt-4xb256_in1k_20220218-13b382a0.pth)\u003Cbr\u002F>[DeiT-tiny distilled](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-tiny-distilled_3rdparty_pt-4xb256_in1k_20211216-c429839a.pth)\u003Cbr\u002F>[DeiT-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-small_pt-4xb256_in1k_20220218-9425b9bb.pth)\u003Cbr\u002F>[DeiT-small distilled](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-small-distilled_3rdparty_pt-4xb256_in1k_20211216-4de1d725.pth)\u003Cbr\u002F>[DeiT-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-base_pt-16xb64_in1k_20220216-db63c16c.pth)\u003Cbr\u002F>[DeiT-base distilled](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-base-distilled_3rdparty_pt-16xb64_in1k_20211216-42891296.pth)\u003Cbr\u002F>[DeiT-base 384px](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-base_3rdparty_ft-16xb32_in1k-384px_20211124-822d02f2.pth)\u003Cbr\u002F>[DeiT-base distilled 384px](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-base-distilled_3rdparty_ft-16xb32_in1k-384px_20211216-e48d6000.pth)|**Conformer**|[Conformer-tiny-p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconformer\u002Fconformer-tiny-p16_3rdparty_8xb128_in1k_20211206-f6860372.pth)\u003Cbr\u002F>[Conformer-small-p32](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconformer\u002Fconformer-small-p32_8xb128_in1k_20211206-947a0816.pth)\u003Cbr\u002F>[Conformer-small-p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconformer\u002Fconformer-small-p16_3rdparty_8xb128_in1k_20211206-3065dcf5.pth)\u003Cbr\u002F>[Conformer-base-p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconformer\u002Fconformer-base-p16_3rdparty_8xb128_in1k_20211206-bfdf8637.pth)|\n|**T2T-ViT**|[T2T-ViT_t-14](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ft2t-vit\u002Ft2t-vit-t-14_8xb64_in1k_20211220-f7378dd5.pth)\u003Cbr\u002F>[T2T-ViT_t-19](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ft2t-vit\u002Ft2t-vit-t-19_8xb64_in1k_20211214-7f5e3aaf.pth)\u003Cbr\u002F>[T2T-ViT_t-24](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ft2t-vit\u002Ft2t-vit-t-24_8xb64_in1k_20211214-b2a68ae3.pth)|**Twins**|[PCPVT-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-pcpvt-small_3rdparty_8xb128_in1k_20220126-ef23c132.pth)\u003Cbr\u002F>[PCPVT-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-pcpvt-base_3rdparty_8xb128_in1k_20220126-f8c4b0d5.pth)\u003Cbr\u002F>[PCPVT-large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-pcpvt-large_3rdparty_16xb64_in1k_20220126-c1ef8d80.pth)\u003Cbr\u002F>[SVT-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-svt-small_3rdparty_8xb128_in1k_20220126-8fe5205b.pth)\u003Cbr\u002F>[SVT-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-svt-base_3rdparty_8xb128_in1k_20220126-e31cc8e9.pth)\u003Cbr\u002F>[SVT-large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-svt-large_3rdparty_16xb64_in1k_20220126-4817645f.pth)|**PoolFormer**|[PoolFormer-S12](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fpoolformer\u002Fpoolformer-s12_3rdparty_32xb128_in1k_20220414-f8d83051.pth)\u003Cbr\u002F>[PoolFormer-S24](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fpoolformer\u002Fpoolformer-s24_3rdparty_32xb128_in1k_20220414-d7055904.pth)\u003Cbr\u002F>[PoolFormer-S36](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fpoolformer\u002Fpoolformer-s36_3rdparty_32xb128_in1k_20220414-d78ff3e8.pth)\u003Cbr\u002F>[PoolFormer-M36](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fpoolformer\u002Fpoolformer-m36_3rdparty_32xb128_in1k_20220414-c55e0949.pth)\u003Cbr\u002F>[PoolFormer-M48](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fpoolformer\u002Fpoolformer-m48_3rdparty_32xb128_in1k_20220414-9378f3eb.pth)|\n|**DenseNet**|[DenseNet121](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdensenet\u002Fdensenet121_4xb256_in1k_20220426-07450f99.pth)\u003Cbr\u002F>[DenseNet161](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdensenet\u002Fdensenet161_4xb256_in1k_20220426-ee6a80a9.pth)\u003Cbr\u002F>[DenseNet169](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdensenet\u002Fdensenet169_4xb256_in1k_20220426-a2889902.pth)\u003Cbr\u002F>[DenseNet201](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdensenet\u002Fdensenet201_4xb256_in1k_20220426-05cae4ef.pth)|**Visual Attention Network(VAN)**|[VAN-Tiny](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvan\u002Fvan-tiny_8xb128_in1k_20220501-385941af.pth)\u003Cbr\u002F>[VAN-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvan\u002Fvan-small_8xb128_in1k_20220501-17bc91aa.pth)\u003Cbr\u002F>[VAN-Base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvan\u002Fvan-base_8xb128_in1k_20220501-6a4cc31b.pth)\u003Cbr\u002F>[VAN-Large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvan\u002Fvan-large_8xb128_in1k_20220501-f212ba21.pth)|**Wide-ResNet**|[WRN-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fwrn\u002Fwide-resnet50_3rdparty-timm_8xb32_in1k_20220304-83ae4399.pth)\u003Cbr\u002F>[WRN-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fwrn\u002Fwide-resnet101_3rdparty_8xb32_in1k_20220304-8d5f9d61.pth)|\n|**HorNet**|[HorNet-Tiny](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-tiny_3rdparty_in1k_20220915-0e8eedff.pth)\u003Cbr\u002F>[HorNet-Tiny-GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-tiny-gf_3rdparty_in1k_20220915-4c35a66b.pth)\u003Cbr\u002F>[HorNet-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-small_3rdparty_in1k_20220915-5935f60f.pth)\u003Cbr\u002F>[HorNet-Small-GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-small-gf_3rdparty_in1k_20220915-649ca492.pth)\u003Cbr\u002F>[HorNet-Base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-base_3rdparty_in1k_20220915-a06176bb.pth)\u003Cbr\u002F>[HorNet-Base-GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-base-gf_3rdparty_in1k_20220915-82c06fa7.pth)\u003Cbr\u002F>[HorNet-Large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-large_3rdparty_in21k_20220909-9ccef421.pth)\u003Cbr\u002F>[HorNet-Large-GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-large-gf_3rdparty_in21k_20220909-3aea3b61.pth)\u003Cbr\u002F>[HorNet-Large-GF384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-base-gf_3rdparty_in1k_20220915-82c06fa7.pth)|**EfficientFormer**|[efficientformer-l1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientformer\u002Fefficientformer-l1_3rdparty_in1k_20220803-d66e61df.pth)\u003Cbr\u002F>[efficientformer-l3](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientformer\u002Fefficientformer-l3_3rdparty_in1k_20220803-dde1c8c5.pth)\u003Cbr\u002F>[efficientformer-l7](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientformer\u002Fefficientformer-l7_3rdparty_in1k_20220803-41a552bb.pth)|**Swin Transformer v2**|[tiny-256 window 8](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-tiny-w8_3rdparty_in1k-256px_20220803-e318968f.pth)\u003Cbr\u002F>[tiny-256 window 16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-tiny-w16_3rdparty_in1k-256px_20220803-9651cdd7.pth)\u003Cbr\u002F>[small-256 window 8](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-small-w8_3rdparty_in1k-256px_20220803-b01a4332.pth)\u003Cbr\u002F>[small-256 window 16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-small-w16_3rdparty_in1k-256px_20220803-b707d206.pth)\u003Cbr\u002F>[base-256 window 8](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-base-w8_3rdparty_in1k-256px_20220803-8ff28f2b.pth)\u003Cbr\u002F>[base-256 window 16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-base-w16_3rdparty_in1k-256px_20220803-5a1886b7.pth)\u003Cbr\u002F>[large-256 window 16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-large-w16_in21k-pre_3rdparty_in1k-256px_20220803-c40cbed7.pth)\u003Cbr\u002F>[large-384 window 24](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-large-w24_in21k-pre_3rdparty_in1k-384px_20220803-3b36c165.pth)|\n|**MViTv2**|[MViTv2-Tiny](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmvit\u002Fmvitv2-tiny_3rdparty_in1k_20220722-db7beeef.pth)\u003Cbr\u002F>[MViTv2-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmvit\u002Fmvitv2-small_3rdparty_in1k_20220722-986bd741.pth)\u003Cbr\u002F>[MViTv2-Base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmvit\u002Fmvitv2-base_3rdparty_in1k_20220722-9c4f0a17.pth)\u003Cbr\u002F>[MViTv2-Large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmvit\u002Fmvitv2-large_3rdparty_in1k_20220722-2b57b983.pth)|**MobileVit**|[MobileViT-XXSmall](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilevit\u002Fmobilevit-xxsmall_3rdparty_in1k_20221018-77835605.pth)\u003Cbr>[MobileViT-XSmall](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilevit\u002Fmobilevit-xsmall_3rdparty_in1k_20221018-be39a6e7.pth)\u003Cbr>[MobileViT-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilevit\u002Fmobilevit-small_3rdparty_in1k_20221018-cb4f741c.pth)|**DaViT**|[DaViT-T](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdavit\u002Fdavit-tiny_3rdparty_in1k_20221116-700fdf7d.pth)\u003Cbr>[DaViT-S](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdavit\u002Fdavit-small_3rdparty_in1k_20221116-51a849a6.pth)\u003Cbr>[DaViT-B](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdavit\u002Fdavit-base_3rdparty_in1k_20221116-19e0d956.pth)|\n|**RepLKNet**|[RepLKNet-31B-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Freplknet\u002Freplknet-31B_in21k-pre_3rdparty_in1k_20221118-54ed5c46.pth)\u003Cbr>[RepLKNet-31B-384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Freplknet\u002Freplknet-31B_in21k-pre_3rdparty_in1k-384px_20221118-76c92b24.pth)\u003Cbr>[RepLKNet-31L-384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Freplknet\u002Freplknet-31L_in21k-pre_3rdparty_in1k-384px_20221118-dc3fc07c.pth)\u003Cbr>[RepLKNet-XL](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Freplknet\u002Freplknet-XL_meg73m-pre_3rdparty_in1k-320px_20221118-88259b1d.pth)|**BEiT**|[BEiT-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fbeit\u002Fbeit-base_3rdparty_in1k_20221114-c0a4df23.pth)|**EVA**|[EVA-G-p14-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-g-p14_30m-pre_3rdparty_in21k_20221213-d72285b7.pth)\u003Cbr>[EVA-G-p14-336](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-g-p14_30m-in21k-pre_3rdparty_in1k-336px_20221213-210f9071.pth)\u003Cbr>[EVA-G-p14-560](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-g-p14_30m-in21k-pre_3rdparty_in1k-560px_20221213-fa1c3652.pth)\u003Cbr>[EVA-G-p16-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-g-p16_3rdparty_30m_20221213-7bed23ee.pth)\u003Cbr>[EVA-L-p14-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-l-p14_mim-pre_3rdparty_in21k_20221213-8f194fa2.pth)\u003Cbr>[EVA-L-p14-196](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-l-p14_mim-in21k-pre_3rdparty_in1k-196px_20221213-b730c7e7.pth)\u003Cbr>[EVA-L-p14-336](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-l-p14_mim-in21k-pre_3rdparty_in1k-336px_20221213-f25b7634.pth)\n|**MixMIM**|[mixmim-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmixmim\u002Fmixmim-base_3rdparty_in1k_20221206-e40e2c8c.pth)|**EfficientNetV2**|[EfficientNetV2-b0](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-b0_3rdparty_in1k_20221221-9ef6e736.pth)\u003Cbr>[EfficientNetV2-b1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-b1_3rdparty_in1k_20221221-6955d9ce.pth)\u003Cbr>[EfficientNetV2-b2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-b2_3rdparty_in1k_20221221-74f7d493.pth)\u003Cbr>[EfficientNetV2-b3](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-b3_3rdparty_in1k_20221221-b6f07a36.pth)\u003Cbr>[EfficientNetV2-s](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-s_in21k-pre-3rdparty_in1k_20221220-7a7c8475.pth)\u003Cbr>[EfficientNetV2-m](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-m_in21k-pre-3rdparty_in1k_20221220-a1013a04.pth)\u003Cbr>[EfficientNetV2-l](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-l_in21k-pre-3rdparty_in1k_20221220-63df0efd.pth)\u003Cbr>[EfficientNetV2-xl](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-xl_in21k-pre-3rdparty_in1k_20221220-583ac18b.pth)|**DeiT3**|[deit3_small_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-small-p16_3rdparty_in1k_20221008-0f7c70cf.pth)\u003Cbr\u002F>[deit3_small_p16_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-small-p16_3rdparty_in1k-384px_20221008-a2c1a0c7.pth)\u003Cbr\u002F>[deit3_base_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-base-p16_3rdparty_in1k_20221008-60b8c8bf.pth)\u003Cbr\u002F>[deit3_base_p16_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-base-p16_3rdparty_in1k-384px_20221009-e19e36d4.pth)\u003Cbr\u002F>[deit3_medium_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-medium-p16_3rdparty_in1k_20221008-3b21284d.pth)\u003Cbr\u002F>[deit3_large_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-large-p16_3rdparty_in1k_20221009-03b427ea.pth)\u003Cbr\u002F>[deit3_large_p16_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-large-p16_3rdparty_in1k-384px_20221009-4317ce62.pth)\u003Cbr\u002F>[deit3_huge_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-huge-p14_3rdparty_in1k_20221009-e107bcb7.pth)|\n|**EdgeNeXt**|[edgenext-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fedgenext\u002Fedgenext-base_3rdparty_in1k_20220801-9ade408b.pth)\u003Cbr>[edgenext-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fedgenext\u002Fedgenext-small_3rdparty_in1k_20220801-d00db5f8.pth)\u003Cbr>[edgenext-X-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fedgenext\u002Fedgenext-xsmall_3rdparty_in1k_20220801-974f9fe7.pth)\u003Cbr>[edgenext-XX-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fedgenext\u002Fedgenext-xxsmall_3rdparty_in1k_20220801-7ca8a81d.pth)|**RevVisionTransformer**|[revvit-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frevvit\u002Frevvit-base_3rdparty_in1k_20221213-87a7b0a5.pth)\u003Cbr>[revvit-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frevvit\u002Frevvit-small_3rdparty_in1k_20221213-a3a34f5c.pth)\n\n## 我维护的其他项目\n- [**图片数据不够？我做了一款图像增强软件**](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FImage-Augmentation)\n[![GitHub stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FFafa-DL\u002FImage-Augmentation)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FImage-Augmentation)\n[![GitHub forks](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fforks\u002FFafa-DL\u002FImage-Augmentation)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FImage-Augmentation)\n- [**一键转换与编辑图像标注文件软件，极大提高效率**](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FLabelConvert)\n[![GitHub stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FFafa-DL\u002FLabelConvert)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FLabelConvert)\n[![GitHub forks](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fforks\u002FFafa-DL\u002FLabelConvert)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FLabelConvert)\n\n## 参考\n```\n@repo{2020mmclassification,\n    title={OpenMMLab's Image Classification Toolbox and Benchmark},\n    author={MMClassification Contributors},\n    howpublished = {\\url{https:\u002F\u002Fgithub.com\u002Fopen-mmlab\u002Fmmclassification}},\n    year={2020}\n}\n```\n","用于图像分类的优秀骨干网络\n===========================\n\n\u003Cdiv align=\"center\">\n\n[![BILIBILI](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_readme_c7801df1a6d3.png)](https:\u002F\u002Fspace.bilibili.com\u002F46880349)\n\n![](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FAwesome%20Backbones-v0.6.3-brightgreen)\n![](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPyTorch-%3E%3Dv1.7.1-green)\n![](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPython-%3E%3Dv3.6-yellowgreen)\n[![GitHub forks](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fforks\u002FFafa-DL\u002FAwesome-Backbones)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones)\n[![GitHub stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FFafa-DL\u002FAwesome-Backbones)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones)\n\n\u003C\u002Fdiv>\n\n## 写在前面\n- 若训练效果不佳，首先需要调整学习率和Batch size，这俩超参很大程度上影响收敛。其次，从关闭图像增强手段（尤其小数据集）开始，有的图像增强方法会污染数据，如\n\n![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_readme_54d1539a62e8.jpg) ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_readme_c2cf1354964f.jpg) ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_readme_3e7dd95a32fb.jpg)\n\n  如何去除增强？如[efficientnetv2-b0](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fmodels\u002Fefficientnetv2\u002Fefficientnetv2_b0.py)配置文件中train_pipeline可更改为如下\n```yaml\ntrain_pipeline = [\n    dict(type='LoadImageFromFile'),\n    dict(\n        type='RandomResizedCrop',\n        size=192,\n        efficientnet_style=True,\n        interpolation='bicubic'),\n    dict(type='Normalize', **img_norm_cfg),\n    dict(type='ImageToTensor', keys=['img']),\n    dict(type='ToTensor', keys=['gt_label']),\n    dict(type='Collect', keys=['img', 'gt_label'])\n]\n```\n  若你的数据集提前已经将shape更改为网络要求的尺寸，那么`Resize`操作也可以去除。\n\n## 更新日志\n\n**`2025.01.17`** \n- 支持转ONNX[#136](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fpull\u002F136) @PurpleSky-NS\n- 类别激活图相关脚本[#114](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fpull\u002F114) @jackyjinjing\n\n**`2024.09.06`** \n- 修复高频反馈的**评估时结果浮动大**的问题\n\n**`2023.12.02`** \n- 新增Issue中多人提及的输出**Train Acc**与**Val loss**\n    - `metrics_outputs.csv`保存每周期`train_loss, train_acc, train_precision, train_recall, train_f1-score, val_loss, val_acc, val_precision, val_recall, val_f1-score`方便各位绘图\n    - 终端由原先仅输出**Val**相关metrics升级为Train与Val都输出\n\n    ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_readme_ba74ba9fe021.jpg)\n\n**`2023.08.05`** \n- 新增**TinyViT**(预训练权重不匹配)、**DeiT3**、**EdgeNeXt**、**RevVisionTransformer**\n\n**`2023.03.07`** \n- 新增**MobileViT**、**DaViT**、**RepLKNet**、**BEiT**、**EVA**、**MixMIM**、**EfficientNetV2**\n\n\n## 测试环境\n\n- Pytorch      1.7.1+\n- Python       3.6+\n\n## 资料\n|数据集|视频教程|人工智能技术探讨群|\n|---|---|---|\n|[`花卉数据集` 提取码：0zat](https:\u002F\u002Fpan.baidu.com\u002Fs\u002F1137y4l-J3AgyCiC_cXqIqw)|[点我跳转](https:\u002F\u002Fwww.bilibili.com\u002Fvideo\u002FBV1SY411P7Nd)|[1群：78174903](https:\u002F\u002Fjq.qq.com\u002F?_wv=1027&k=lY5KVICA)\u003Cbr\u002F>[3群：584723646](https:\u002F\u002Fjq.qq.com\u002F?_wv=1027&k=bakez5Yz)\n\n## 快速开始\n\n- 遵循[环境搭建](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FEnvironment_setting.md)完成配置\n- 下载[MobileNetV3-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilenet_v3\u002Fconvert\u002Fmobilenet_v3_small-8427ecf0.pth)权重至**datas**下\n- **Awesome-Backbones**文件夹下终端输入\n```bash\npython tools\u002Fsingle_test.py datas\u002Fcat-dog.png models\u002Fmobilenet\u002Fmobilenet_v3_small.py --classes-map datas\u002FimageNet1kAnnotation.txt\n```\n\n## 教程\n- [环境搭建](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FEnvironment_setting.md)\n- [数据集准备](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FData_preparing.md)\n- [配置文件解释](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FConfigs_description.md)\n- [训练](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FHow_to_train.md)\n- [模型评估](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FHow_to_eval.md)\n- [计算Flops&Params](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FCalculate_Flops.md)\n- [添加新的模型组件](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FAdd_modules.md)\n- [类别激活图可视化](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FCAM_visualization.md)\n- [学习率策略可视化](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FLr_visualization.md)\n- [数据增强策略可视化](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fblob\u002Fmain\u002Fdatas\u002Fdocs\u002FPipeline_visualization.md)\n\n## 模型\n- [x] [LeNet5](https:\u002F\u002Fieeexplore.ieee.org\u002Fdocument\u002F6795724)\n- [x] [AlexNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124461111)\n- [x] [VGG](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124477080)\n- [x] [DenseNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124630832)\n- [x] [ResNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124477575)\n- [x] [Wide-ResNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124754437)\n- [x] [ResNeXt](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124477919)\n- [x] [SEResNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124478157)\n- [x] [SEResNeXt](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124478347)\n- [x] [RegNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124478426)\n- [x] [MobileNetV2](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124478681)\n- [x] [MobileNetV3](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124478770)\n- [x] [ShuffleNetV1](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124479156)\n- [x] [ShuffleNetV2](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124479336)\n- [x] [EfficientNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124754493)\n- [x] [RepVGG](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124479644)\n- [x] [Res2Net](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124479467)\n- [x] [ConvNeXt](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124481466)\n- [x] [HRNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124481590)\n- [x] [ConvMixer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124481766)\n- [x] [CSPNet](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124481930)\n- [x] [Swin-Transformer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124538198)\n- [x] [Vision-Transformer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124567953)\n- [x] [Transformer-in-Transformer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596023)\n- [x] [MLP-Mixer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596093)\n- [x] [DeiT](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124591888)\n- [x] [Conformer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596343)\n- [x] [T2T-ViT](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596425)\n- [x] [Twins](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596619)\n- [x] [PoolFormer](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124596740)\n- [x] [VAN](https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\u002Farticle\u002Fdetails\u002F124630541)\n- [x] [HorNet](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2207.14284v2.pdf)\n- [x] [EfficientFormer](https:\u002F\u002Farxiv.org\u002Fabs\u002F2206.01191)\n- [x] [Swin Transformer V2](https:\u002F\u002Farxiv.org\u002Fabs\u002F2111.09883.pdf)\n- [x] [MViT V2](http:\u002F\u002Fopenaccess.thecvf.com\u002F\u002Fcontent\u002FCVPR2022\u002Fpapers\u002FLi_MViTv2_Improved_Multiscale_Vision_Transformers_for_Classification_and_Detection_CVPR_2022_paper.pdf)\n- [x] [MobileViT](https:\u002F\u002Farxiv.org\u002Fabs\u002F2110.02178)\n- [x] [DaViT](https:\u002F\u002Farxiv.org\u002Fabs\u002F2204.03645v1)\n- [x] [replknet](https:\u002F\u002Farxiv.org\u002Fabs\u002F2203.06717)\n- [x] [BEiT](https:\u002F\u002Farxiv.org\u002Fabs\u002F2106.08254)\n- [x] [EVA](https:\u002F\u002Farxiv.org\u002Fabs\u002F2211.07636)\n- [x] [MixMIM](https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.13137)\n- [x] [EfficientNetV2](https:\u002F\u002Farxiv.org\u002Fabs\u002F2104.00298)\n## 预训练权重\n\n| 名称 | 权重 | 名称 | 权重 | 名称 | 权重 |\n| :-----: | :-----: | :------: | :------: | :------: | :-----: |\n| **LeNet5** | None | **AlexNet** | None | **VGG** | [VGG-11](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg11_batch256_imagenet_20210208-4271cd6c.pth)\u003Cbr\u002F>[VGG-13](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg13_batch256_imagenet_20210208-4d1d6080.pth)\u003Cbr\u002F>[VGG-16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg16_batch256_imagenet_20210208-db26f1a5.pth)\u003Cbr\u002F>[VGG-19](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg19_batch256_imagenet_20210208-e6920e4a.pth)\u003Cbr\u002F>[VGG-11-BN](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg11_bn_batch256_imagenet_20210207-f244902c.pth)\u003Cbr\u002F>[VGG-13-BN](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg13_bn_batch256_imagenet_20210207-1a8b7864.pth)\u003Cbr\u002F>[VGG-16-BN](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg16_bn_batch256_imagenet_20210208-7e55cd29.pth)\u003Cbr\u002F>[VGG-19-BN](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvgg\u002Fvgg19_bn_batch256_imagenet_20210208-da620c4f.pth)|\n| **ResNet** |[ResNet-18](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnet18_8xb32_in1k_20210831-fbbb1da6.pth)\u003Cbr\u002F>[ResNet-34](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnet34_8xb32_in1k_20210831-f257d4e6.pth)\u003Cbr\u002F>[ResNet-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnet50_8xb32_in1k_20210831-ea4938fc.pth)\u003Cbr\u002F>[ResNet-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnet101_8xb32_in1k_20210831-539c63f8.pth)\u003Cbr\u002F>[ResNet-152](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnet152_8xb32_in1k_20210901-4d7582fa.pth) | **ResNetV1C** | [ResNetV1C-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1c50_8xb32_in1k_20220214-3343eccd.pth)\u003Cbr\u002F>[ResNetV1C-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1c101_8xb32_in1k_20220214-434fe45f.pth)\u003Cbr\u002F>[ResNetV1C-152](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1c152_8xb32_in1k_20220214-c013291f.pth) |**ResNetV1D** | [ResNetV1D-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1d50_b32x8_imagenet_20210531-db14775a.pth)\u003Cbr\u002F>[ResNetV1D-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1d101_b32x8_imagenet_20210531-6e13bcd3.pth)\u003Cbr\u002F>[ResNetV1D-152](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnet\u002Fresnetv1d152_b32x8_imagenet_20210531-278cf22a.pth) |\n| **ResNeXt** | [ResNeXt-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnext\u002Fresnext50_32x4d_b32x8_imagenet_20210429-56066e27.pth)\u003Cbr\u002F>[ResNeXt-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnext\u002Fresnext101_32x4d_b32x8_imagenet_20210506-e0fa3dd5.pth)\u003Cbr\u002F>[ResNeXt-152](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fresnext\u002Fresnext152_32x4d_b32x8_imagenet_20210524-927787be.pth) | **SEResNet** | [SEResNet-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fse-resnet\u002Fse-resnet50_batch256_imagenet_20200804-ae206104.pth)\u003Cbr\u002F>[SEResNet-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fse-resnet\u002Fse-resnet101_batch256_imagenet_20200804-ba5b51d4.pth)| **SEResNeXt**| None|\n| **RegNet** |[RegNetX-400MF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-400mf_8xb128_in1k_20211213-89bfc226.pth)\u003Cbr\u002F>[RegNetX-800MF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-800mf_8xb128_in1k_20211213-222b0f11.pth)\u003Cbr\u002F>[RegNetX-1.6GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-1.6gf_8xb128_in1k_20211213-d1b89758.pth)\u003Cbr\u002F>[RegNetX-3.2GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-3.2gf_8xb64_in1k_20211213-1fdd82ae.pth)\u003Cbr\u002F>[RegNetX-4.0GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-4.0gf_8xb64_in1k_20211213-efed675c.pth)\u003Cbr\u002F>[RegNetX-6.4GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-6.4gf_8xb64_in1k_20211215-5c6089da.pth)\u003Cbr\u002F>[RegNetX-8.0GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-8.0gf_8xb64_in1k_20211213-9a9fcc76.pth)\u003Cbr\u002F>[RegNetX-12GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fregnet\u002Fregnetx-12gf_8xb64_in1k_20211213-5df8c2f8.pth) | **MobileNetV2** | [MobileNetV2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilenet_v2\u002Fmobilenet_v2_batch256_imagenet_20200708-3b2dc3af.pth) |**MobileNetV3** | [MobileNetV3-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilenet_v3\u002Fconvert\u002Fmobilenet_v3_small-8427ecf0.pth)\u003Cbr\u002F>[MobileNetV3-Large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilenet_v3\u002Fconvert\u002Fmobilenet_v3_large-3ea3c186.pth) |\n| **ShuffleNetV1** |[ShuffleNetV1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fshufflenet_v1\u002Fshufflenet_v1_batch1024_imagenet_20200804-5d6cec73.pth) | **ShuffleNetV2** | [ShuffleNetV2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fshufflenet_v2\u002Fshufflenet_v2_batch1024_imagenet_20200812-5bf4721e.pth) |**EfficientNet** | [EfficientNet-B0](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b0_3rdparty_8xb32_in1k_20220119-a7e2a0b1.pth)\u003Cbr\u002F>[EfficientNet-B1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b1_3rdparty_8xb32_in1k_20220119-002556d9.pth)\u003Cbr\u002F>[EfficientNet-B2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b2_3rdparty_8xb32_in1k_20220119-ea374a30.pth)\u003Cbr\u002F>[EfficientNet-B3](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b3_3rdparty_8xb32_in1k_20220119-4b4d7487.pth)\u003Cbr\u002F>[EfficientNet-B4](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b4_3rdparty_8xb32_in1k_20220119-81fd4077.pth)\u003Cbr\u002F>[EfficientNet-B5](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b5_3rdparty_8xb32_in1k_20220119-e9814430.pth)\u003Cbr\u002F>[EfficientNet-B6](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b6_3rdparty_8xb32-aa_in1k_20220119-45b03310.pth)\u003Cbr\u002F>[EfficientNet-B7](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b7_3rdparty_8xb32-aa_in1k_20220119-bf03951c.pth)\u003Cbr\u002F>[EfficientNet-B8](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnet\u002Fefficientnet-b8_3rdparty_8xb32-aa-advprop_in1k_20220119-297ce1b7.pth) |\n| **RepVGG** |[RepVGG-A0](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-A0_3rdparty_4xb64-coslr-120e_in1k_20210909-883ab98c.pth)\u003Cbr\u002F>[RepVGG-A1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-A1_3rdparty_4xb64-coslr-120e_in1k_20210909-24003a24.pth) \u003Cbr\u002F>[RepVGG-A2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-A2_3rdparty_4xb64-coslr-120e_in1k_20210909-97d7695a.pth)\u003Cbr\u002F>[RepVGG-B0](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B0_3rdparty_4xb64-coslr-120e_in1k_20210909-446375f4.pth)\u003Cbr\u002F>[RepVGG-B1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B1_3rdparty_4xb64-coslr-120e_in1k_20210909-750cdf67.pth)\u003Cbr\u002F>[RepVGG-A1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-A1_3rdparty_4xb64-coslr-120e_in1k_20210909-24003a24.pth)\u003Cbr\u002F>[RepVGG-B1g2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B1g2_3rdparty_4xb64-coslr-120e_in1k_20210909-344f6422.pth)\u003Cbr\u002F>[RepVGG-B1g4](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B1g4_3rdparty_4xb64-coslr-120e_in1k_20210909-d4c1a642.pth)\u003Cbr\u002F>[RepVGG-B2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B2_3rdparty_4xb64-coslr-120e_in1k_20210909-bd6b937c.pth)\u003Cbr\u002F>[RepVGG-B2g4](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B2g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-7b7955f0.pth)\u003Cbr\u002F>[RepVGG-B2g4](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B2g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-7b7955f0.pth)\u003Cbr\u002F>[RepVGG-B3](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B3_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-dda968bf.pth)\u003Cbr\u002F>[RepVGG-B3g4](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-B3g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-4e54846a.pth)\u003Cbr\u002F>[RepVGG-D2se](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frepvgg\u002Frepvgg-D2se_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-cf3139b7.pth)| **Res2Net** | [Res2Net-50-14w-8s](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fres2net\u002Fres2net50-w14-s8_3rdparty_8xb32_in1k_20210927-bc967bf1.pth)\u003Cbr\u002F>[Res2Net-50-26w-8s](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fres2net\u002Fres2net50-w26-s8_3rdparty_8xb32_in1k_20210927-f547a94b.pth)\u003Cbr\u002F>[Res2Net-101-26w-4s](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fres2net\u002Fres2net101-w26-s4_3rdparty_8xb32_in1k_20210927-870b6c36.pth)\u003Cbr\u002F> |**ConvNeXt** | [ConvNeXt-Tiny](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvnext\u002Fconvnext-tiny_3rdparty_32xb128_in1k_20220124-18abde00.pth)\u003Cbr\u002F>[ConvNeXt-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvnext\u002Fconvnext-small_3rdparty_32xb128_in1k_20220124-d39b5192.pth)\u003Cbr\u002F>[ConvNeXt-Base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvnext\u002Fconvnext-base_in21k-pre-3rdparty_32xb128_in1k_20220124-eb2d6ada.pth)\u003Cbr\u002F>[ConvNeXt-Large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvnext\u002Fconvnext-large_in21k-pre-3rdparty_64xb64_in1k_20220124-2412403d.pth)\u003Cbr\u002F>[ConvNeXt-XLarge](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvnext\u002Fconvnext-xlarge_in21k-pre-3rdparty_64xb64_in1k_20220124-76b6863d.pth) |\n| **HRNet** |[HRNet-W18](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w18_3rdparty_8xb32_in1k_20220120-0c10b180.pth)\u003Cbr\u002F>[HRNet-W30](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w30_3rdparty_8xb32_in1k_20220120-8aa3832f.pth) \u003Cbr\u002F>[HRNet-W32](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w32_3rdparty_8xb32_in1k_20220120-c394f1ab.pth)\u003Cbr\u002F>[HRNet-W40](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w40_3rdparty_8xb32_in1k_20220120-9a2dbfc5.pth)\u003Cbr\u002F>[HRNet-W44](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w44_3rdparty_8xb32_in1k_20220120-35d07f73.pth)\u003Cbr\u002F>[HRNet-W48](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w48_3rdparty_8xb32_in1k_20220120-e555ef50.pth)\u003Cbr\u002F>[HRNet-W64](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhrnet\u002Fhrnet-w64_3rdparty_8xb32_in1k_20220120-19126642.pth) | **ConvMixer** | [ConvMixer-768\u002F32](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvmixer\u002Fconvmixer-768-32_3rdparty_10xb64_in1k_20220323-bca1f7b8.pth)\u003Cbr\u002F>[ConvMixer-1024\u002F20](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvmixer\u002Fconvmixer-1024-20_3rdparty_10xb64_in1k_20220323-48f8aeba.pth)\u003Cbr\u002F>[ConvMixer-1536\u002F20](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconvmixer\u002Fconvmixer-1536_20_3rdparty_10xb64_in1k_20220323-ea5786f3.pth) |**CSPNet** | [CSPDarkNet50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fcspnet\u002Fcspdarknet50_3rdparty_8xb32_in1k_20220329-bd275287.pth)\u003Cbr\u002F>[CSPResNet50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fcspnet\u002Fcspresnet50_3rdparty_8xb32_in1k_20220329-dd6dddfb.pth)\u003Cbr\u002F>[CSPResNeXt50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fcspnet\u002Fcspresnext50_3rdparty_8xb32_in1k_20220329-2cc84d21.pth) |\n|**Swin Transformer**|[tiny-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fswin_tiny_224_b16x64_300e_imagenet_20210616_090925-66df6be6.pth)\u003Cbr\u002F>[small-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fswin_small_224_b16x64_300e_imagenet_20210615_110219-7f9d988b.pth)\u003Cbr\u002F>[base-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fswin_base_224_b16x64_300e_imagenet_20210616_190742-93230b0d.pth)\u003Cbr\u002F>[large-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fconvert\u002Fswin_large_patch4_window7_224_22kto1k-5f0996db.pth)\u003Cbr\u002F>[base-384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fconvert\u002Fswin_base_patch4_window12_384_22kto1k-d59b0d1d.pth)\u003Cbr\u002F>[large-384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-transformer\u002Fconvert\u002Fswin_large_patch4_window12_384_22kto1k-0a40944b.pth)|**Vision Transformer**|[vit_base_p16_224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Fpretrain\u002Fvit-base-p16_3rdparty_pt-64xb64_in1k-224_20210928-02284250.pth)\u003Cbr\u002F>[vit_base_p32_224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Fpretrain\u002Fvit-base-p32_3rdparty_pt-64xb64_in1k-224_20210928-eee25dd4.pth)\u003Cbr\u002F>[vit_large_p16_224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Fpretrain\u002Fvit-large-p16_3rdparty_pt-64xb64_in1k-224_20210928-0001f9a1.pth)\u003Cbr\u002F>[vit_base_p16_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Ffinetune\u002Fvit-base-p16_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-98e8652b.pth)\u003Cbr\u002F>[vit_base_p32_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Ffinetune\u002Fvit-base-p32_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-9cea8599.pth)\u003Cbr\u002F>[vit_large_p16_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvit\u002Ffinetune\u002Fvit-large-p16_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-b20ba619.pth)|**Transformer in Transformer**|[TNT-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftnt\u002Ftnt-small-p16_3rdparty_in1k_20210903-c56ee7df.pth)|\n|**MLP Mixer**|[base_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmlp-mixer\u002Fmixer-base-p16_3rdparty_64xb64_in1k_20211124-1377e3e0.pth)\u003Cbr\u002F>[large_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmlp-mixer\u002Fmixer-large-p16_3rdparty_64xb64_in1k_20211124-5a2519d2.pth)|**Deit**|[DeiT-tiny](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-tiny_pt-4xb256_in1k_20220218-13b382a0.pth)\u003Cbr\u002F>[DeiT-tiny distilled](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-tiny-distilled_3rdparty_pt-4xb256_in1k_20211216-c429839a.pth)\u003Cbr\u002F>[DeiT-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-small_pt-4xb256_in1k_20220218-9425b9bb.pth)\u003Cbr\u002F>[DeiT-small distilled](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-small-distilled_3rdparty_pt-4xb256_in1k_20211216-4de1d725.pth)\u003Cbr\u002F>[DeiT-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-base_pt-16xb64_in1k_20220216-db63c16c.pth)\u003Cbr\u002F>[DeiT-base distilled](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-base-distilled_3rdparty_pt-16xb64_in1k_20211216-42891296.pth)\u003Cbr\u002F>[DeiT-base 384px](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-base_3rdparty_ft-16xb32_in1k-384px_20211124-822d02f2.pth)\u003Cbr\u002F>[DeiT-base distilled 384px](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit\u002Fdeit-base-distilled_3rdparty_ft-16xb32_in1k-384px_20211216-e48d6000.pth)|**Conformer**|[Conformer-tiny-p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconformer\u002Fconformer-tiny-p16_3rdparty_8xb128_in1k_20211206-f6860372.pth)\u003Cbr\u002F>[Conformer-small-p32](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconformer\u002Fconformer-small-p32_8xb128_in1k_20211206-947a0816.pth)\u003Cbr\u002F>[Conformer-small-p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconformer\u002Fconformer-small-p16_3rdparty_8xb128_in1k_20211206-3065dcf5.pth)\u003Cbr\u002F>[Conformer-base-p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fconformer\u002Fconformer-base-p16_3rdparty_8xb128_in1k_20211206-bfdf8637.pth)|\n|**T2T-ViT**|[T2T-ViT_t-14](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ft2t-vit\u002Ft2t-vit-t-14_8xb64_in1k_20211220-f7378dd5.pth)\u003Cbr\u002F>[T2T-ViT_t-19](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ft2t-vit\u002Ft2t-vit-t-19_8xb64_in1k_20211214-7f5e3aaf.pth)\u003Cbr\u002F>[T2T-ViT_t-24](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ft2t-vit\u002Ft2t-vit-t-24_8xb64_in1k_20211214-b2a68ae3.pth)|**Twins**|[PCPVT-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-pcpvt-small_3rdparty_8xb128_in1k_20220126-ef23c132.pth)\u003Cbr\u002F>[PCPVT-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-pcpvt-base_3rdparty_8xb128_in1k_20220126-f8c4b0d5.pth)\u003Cbr\u002F>[PCPVT-large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-pcpvt-large_3rdparty_16xb64_in1k_20220126-c1ef8d80.pth)\u003Cbr\u002F>[SVT-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-svt-small_3rdparty_8xb128_in1k_20220126-8fe5205b.pth)\u003Cbr\u002F>[SVT-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-svt-base_3rdparty_8xb128_in1k_20220126-e31cc8e9.pth)\u003Cbr\u002F>[SVT-large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Ftwins\u002Ftwins-svt-large_3rdparty_16xb64_in1k_20220126-4817645f.pth)|**PoolFormer**|[PoolFormer-S12](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fpoolformer\u002Fpoolformer-s12_3rdparty_32xb128_in1k_20220414-f8d83051.pth)\u003Cbr\u002F>[PoolFormer-S24](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fpoolformer\u002Fpoolformer-s24_3rdparty_32xb128_in1k_20220414-d7055904.pth)\u003Cbr\u002F>[PoolFormer-S36](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fpoolformer\u002Fpoolformer-s36_3rdparty_32xb128_in1k_20220414-d78ff3e8.pth)\u003Cbr\u002F>[PoolFormer-M36](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fpoolformer\u002Fpoolformer-m36_3rdparty_32xb128_in1k_20220414-c55e0949.pth)\u003Cbr\u002F>[PoolFormer-M48](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fpoolformer\u002Fpoolformer-m48_3rdparty_32xb128_in1k_20220414-9378f3eb.pth)|\n|**DenseNet**|[DenseNet121](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdensenet\u002Fdensenet121_4xb256_in1k_20220426-07450f99.pth)\u003Cbr\u002F>[DenseNet161](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdensenet\u002Fdensenet161_4xb256_in1k_20220426-ee6a80a9.pth)\u003Cbr\u002F>[DenseNet169](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdensenet\u002Fdensenet169_4xb256_in1k_20220426-a2889902.pth)\u003Cbr\u002F>[DenseNet201](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdensenet\u002Fdensenet201_4xb256_in1k_20220426-05cae4ef.pth)|**Visual Attention Network(VAN)**|[VAN-Tiny](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvan\u002Fvan-tiny_8xb128_in1k_20220501-385941af.pth)\u003Cbr\u002F>[VAN-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvan\u002Fvan-small_8xb128_in1k_20220501-17bc91aa.pth)\u003Cbr\u002F>[VAN-Base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvan\u002Fvan-base_8xb128_in1k_20220501-6a4cc31b.pth)\u003Cbr\u002F>[VAN-Large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fvan\u002Fvan-large_8xb128_in1k_20220501-f212ba21.pth)|**Wide-ResNet**|[WRN-50](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fwrn\u002Fwide-resnet50_3rdparty-timm_8xb32_in1k_20220304-83ae4399.pth)\u003Cbr\u002F>[WRN-101](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fwrn\u002Fwide-resnet101_3rdparty_8xb32_in1k_20220304-8d5f9d61.pth)|\n|**HorNet**|[HorNet-Tiny](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-tiny_3rdparty_in1k_20220915-0e8eedff.pth)\u003Cbr\u002F>[HorNet-Tiny-GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-tiny-gf_3rdparty_in1k_20220915-4c35a66b.pth)\u003Cbr\u002F>[HorNet-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-small_3rdparty_in1k_20220915-5935f60f.pth)\u003Cbr\u002F>[HorNet-Small-GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-small-gf_3rdparty_in1k_20220915-649ca492.pth)\u003Cbr\u002F>[HorNet-Base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-base_3rdparty_in1k_20220915-a06176bb.pth)\u003Cbr\u002F>[HorNet-Base-GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-base-gf_3rdparty_in1k_20220915-82c06fa7.pth)\u003Cbr\u002F>[HorNet-Large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-large_3rdparty_in21k_20220909-9ccef421.pth)\u003Cbr\u002F>[HorNet-Large-GF](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-large-gf_3rdparty_in21k_20220909-3aea3b61.pth)\u003Cbr\u002F>[HorNet-Large-GF384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fhornet\u002Fhornet-base-gf_3rdparty_in1k_20220915-82c06fa7.pth)|**EfficientFormer**|[efficientformer-l1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientformer\u002Fefficientformer-l1_3rdparty_in1k_20220803-d66e61df.pth)\u003Cbr\u002F>[efficientformer-l3](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientformer\u002Fefficientformer-l3_3rdparty_in1k_20220803-dde1c8c5.pth)\u003Cbr\u002F>[efficientformer-l7](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientformer\u002Fefficientformer-l7_3rdparty_in1k_20220803-41a552bb.pth)|**Swin Transformer v2**|[tiny-256 window 8](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-tiny-w8_3rdparty_in1k-256px_20220803-e318968f.pth)\u003Cbr\u002F>[tiny-256 window 16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-tiny-w16_3rdparty_in1k-256px_20220803-9651cdd7.pth)\u003Cbr\u002F>[small-256 window 8](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-small-w8_3rdparty_in1k-256px_20220803-b01a4332.pth)\u003Cbr\u002F>[small-256 window 16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-small-w16_3rdparty_in1k-256px_20220803-b707d206.pth)\u003Cbr\u002F>[base-256 window 8](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-base-w8_3rdparty_in1k-256px_20220803-8ff28f2b.pth)\u003Cbr\u002F>[base-256 window 16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-base-w16_3rdparty_in1k-256px_20220803-5a1886b7.pth)\u003Cbr\u002F>[large-256 window 16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-large-w16_in21k-pre_3rdparty_in1k-256px_20220803-c40cbed7.pth)\u003Cbr\u002F>[large-384 window 24](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fswin-v2\u002Fswinv2-large-w24_in21k-pre_3rdparty_in1k-384px_20220803-3b36c165.pth)|\n|**MViTv2**|[MViTv2-Tiny](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmvit\u002Fmvitv2-tiny_3rdparty_in1k_20220722-db7beeef.pth)\u003Cbr\u002F>[MViTv2-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmvit\u002Fmvitv2-small_3rdparty_in1k_20220722-986bd741.pth)\u003Cbr\u002F>[MViTv2-Base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmvit\u002Fmvitv2-base_3rdparty_in1k_20220722-9c4f0a17.pth)\u003Cbr\u002F>[MViTv2-Large](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmvit\u002Fmvitv2-large_3rdparty_in1k_20220722-2b57b983.pth)|**MobileVit**|[MobileViT-XXSmall](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilevit\u002Fmobilevit-xxsmall_3rdparty_in1k_20221018-77835605.pth)\u003Cbr>[MobileViT-XSmall](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilevit\u002Fmobilevit-xsmall_3rdparty_in1k_20221018-be39a6e7.pth)\u003Cbr>[MobileViT-Small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilevit\u002Fmobilevit-small_3rdparty_in1k_20221018-cb4f741c.pth)|**DaViT**|[DaViT-T](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdavit\u002Fdavit-tiny_3rdparty_in1k_20221116-700fdf7d.pth)\u003Cbr>[DaViT-S](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdavit\u002Fdavit-small_3rdparty_in1k_20221116-51a849a6.pth)\u003Cbr>[DaViT-B](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdavit\u002Fdavit-base_3rdparty_in1k_20221116-19e0d956.pth)|\n|**RepLKNet**|[RepLKNet-31B-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Freplknet\u002Freplknet-31B_in21k-pre_3rdparty_in1k_20221118-54ed5c46.pth)\u003Cbr>[RepLKNet-31B-384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Freplknet\u002Freplknet-31B_in21k-pre_3rdparty_in1k-384px_20221118-76c92b24.pth)\u003Cbr>[RepLKNet-31L-384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Freplknet\u002Freplknet-31L_in21k-pre_3rdparty_in1k-384px_20221118-dc3fc07c.pth)\u003Cbr>[RepLKNet-XL](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Freplknet\u002Freplknet-XL_meg73m-pre_3rdparty_in1k-320px_20221118-88259b1d.pth)|**BEiT**|[BEiT-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fbeit\u002Fbeit-base_3rdparty_in1k_20221114-c0a4df23.pth)|**EVA**|[EVA-G-p14-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-g-p14_30m-pre_3rdparty_in21k_20221213-d72285b7.pth)\u003Cbr>[EVA-G-p14-336](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-g-p14_30m-in21k-pre_3rdparty_in1k-336px_20221213-210f9071.pth)\u003Cbr>[EVA-G-p14-560](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-g-p14_30m-in21k-pre_3rdparty_in1k-560px_20221213-fa1c3652.pth)\u003Cbr>[EVA-G-p16-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-g-p16_3rdparty_30m_20221213-7bed23ee.pth)\u003Cbr>[EVA-L-p14-224](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-l-p14_mim-pre_3rdparty_in21k_20221213-8f194fa2.pth)\u003Cbr>[EVA-L-p14-196](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-l-p14_mim-in21k-pre_3rdparty_in1k-196px_20221213-b730c7e7.pth)\u003Cbr>[EVA-L-p14-336](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Feva\u002Feva-l-p14_mim-in21k-pre_3rdparty_in1k-336px_20221213-f25b7634.pth)\n|**MixMIM**|[mixmim-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmixmim\u002Fmixmim-base_3rdparty_in1k_20221206-e40e2c8c.pth)|**EfficientNetV2**|[EfficientNetV2-b0](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-b0_3rdparty_in1k_20221221-9ef6e736.pth)\u003Cbr>[EfficientNetV2-b1](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-b1_3rdparty_in1k_20221221-6955d9ce.pth)\u003Cbr>[EfficientNetV2-b2](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-b2_3rdparty_in1k_20221221-74f7d493.pth)\u003Cbr>[EfficientNetV2-b3](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-b3_3rdparty_in1k_20221221-b6f07a36.pth)\u003Cbr>[EfficientNetV2-s](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-s_in21k-pre-3rdparty_in1k_20221220-7a7c8475.pth)\u003Cbr>[EfficientNetV2-m](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-m_in21k-pre-3rdparty_in1k_20221220-a1013a04.pth)\u003Cbr>[EfficientNetV2-l](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-l_in21k-pre-3rdparty_in1k_20221220-63df0efd.pth)\u003Cbr>[EfficientNetV2-xl](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fefficientnetv2\u002Fefficientnetv2-xl_in21k-pre-3rdparty_in1k_20221220-583ac18b.pth)|**DeiT3**|[deit3_small_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-small-p16_3rdparty_in1k_20221008-0f7c70cf.pth)\u003Cbr\u002F>[deit3_small_p16_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-small-p16_3rdparty_in1k-384px_20221008-a2c1a0c7.pth)\u003Cbr\u002F>[deit3_base_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-base-p16_3rdparty_in1k_20221008-60b8c8bf.pth)\u003Cbr\u002F>[deit3_base_p16_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-base-p16_3rdparty_in1k-384px_20221009-e19e36d4.pth)\u003Cbr\u002F>[deit3_medium_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-medium-p16_3rdparty_in1k_20221008-3b21284d.pth)\u003Cbr\u002F>[deit3_large_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-large-p16_3rdparty_in1k_20221009-03b427ea.pth)\u003Cbr\u002F>[deit3_large_p16_384](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-large-p16_3rdparty_in1k-384px_20221009-4317ce62.pth)\u003Cbr\u002F>[deit3_huge_p16](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fdeit3\u002Fdeit3-huge-p14_3rdparty_in1k_20221009-e107bcb7.pth)|\n|**EdgeNeXt**|[edgenext-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fedgenext\u002Fedgenext-base_3rdparty_in1k_20220801-9ade408b.pth)\u003Cbr>[edgenext-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fedgenext\u002Fedgenext-small_3rdparty_in1k_20220801-d00db5f8.pth)\u003Cbr>[edgenext-X-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fedgenext\u002Fedgenext-xsmall_3rdparty_in1k_20220801-974f9fe7.pth)\u003Cbr>[edgenext-XX-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fedgenext\u002Fedgenext-xxsmall_3rdparty_in1k_20220801-7ca8a81d.pth)|**RevVisionTransformer**|[revvit-small](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frevvit\u002Frevvit-base_3rdparty_in1k_20221213-87a7b0a5.pth)\u003Cbr>[revvit-base](https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Frevvit\u002Frevvit-small_3rdparty_in1k_20221213-a3a34f5c.pth)\n\n## 我维护的其他项目\n- [**图片数据不够？我做了一款图像增强软件**](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FImage-Augmentation)\n[![GitHub 星标](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FFafa-DL\u002FImage-Augmentation)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FImage-Augmentation)\n[![GitHub 分支](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fforks\u002FFafa-DL\u002FImage-Augmentation)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FImage-Augmentation)\n- [**一键转换与编辑图像标注文件软件，极大提高效率**](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FLabelConvert)\n[![GitHub 星标](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FFafa-DL\u002FLabelConvert)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FLabelConvert)\n[![GitHub 分支](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fforks\u002FFafa-DL\u002FLabelConvert)](https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FLabelConvert)\n\n## 参考\n```\n@repo{2020mmclassification,\n    title={OpenMMLab 的图像分类工具箱及基准测试},\n    author={MMClassification 贡献者},\n    howpublished = {\\url{https:\u002F\u002Fgithub.com\u002Fopen-mmlab\u002Fmmclassification}},\n    year={2020}\n}\n```","# Awesome-Backbones 快速上手指南\n\nAwesome-Backbones 是一个专注于图像分类任务的开源主干网络（Backbone）集合，集成了从经典的 LeNet、ResNet 到最新的 Swin Transformer、ConvNeXt 等多种主流模型。本指南将帮助你快速搭建环境并运行第一个测试示例。\n\n## 1. 环境准备\n\n在开始之前，请确保你的开发环境满足以下最低要求：\n\n*   **操作系统**: Linux \u002F Windows \u002F macOS\n*   **Python**: >= 3.6\n*   **PyTorch**: >= 1.7.1\n*   **其他依赖**: 建议安装 `mmcv` 及相关图像处理库（如 `opencv-python`, `pillow`）\n\n> **提示**：国内用户推荐使用清华源或阿里源加速 Python 包的安装。\n\n## 2. 安装步骤\n\n### 第一步：克隆项目\n将代码仓库克隆到本地：\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones.git\ncd Awesome-Backbones\n```\n\n### 第二步：安装依赖\n根据项目根目录下的 `requirements.txt` 安装所需库（如果存在），或直接安装核心依赖：\n```bash\npip install torch>=1.7.1 torchvision -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\npip install -r requirements.txt -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n```\n*(注：若项目中无 requirements.txt，请参照 `datas\u002Fdocs\u002FEnvironment_setting.md` 文档进行详细配置)*\n\n### 第三步：下载预训练权重\n为了快速验证，我们需要下载一个轻量级模型的权重。以 **MobileNetV3-Small** 为例，将其下载到项目的 `datas` 目录下：\n\n```bash\n# 创建目录（如果不存在）\nmkdir -p datas\n\n# 下载权重 (使用 wget 或手动下载后放入)\nwget https:\u002F\u002Fdownload.openmmlab.com\u002Fmmclassification\u002Fv0\u002Fmobilenet_v3\u002Fconvert\u002Fmobilenet_v3_small-8427ecf0.pth -P datas\u002F\n```\n*国内用户若下载缓慢，可复制链接在浏览器中使用迅雷等工具下载，然后手动移动文件至 `datas\u002F` 文件夹。*\n\n### 第四步：准备测试图片\n确保你有一张用于测试的图片（例如 `cat-dog.png`），并将其放置在 `datas` 目录下，或者修改后续命令中的图片路径。\n\n## 3. 基本使用\n\n完成上述准备后，即可通过命令行运行单张图片的推理测试。\n\n在项目根目录 (`Awesome-Backbones`) 下执行以下命令：\n\n```bash\npython tools\u002Fsingle_test.py datas\u002Fcat-dog.png models\u002Fmobilenet\u002Fmobilenet_v3_small.py --classes-map datas\u002FimageNet1kAnnotation.txt\n```\n\n**命令参数说明：**\n*   `datas\u002Fcat-dog.png`: 待测试的图片路径。\n*   `models\u002Fmobilenet\u002Fmobilenet_v3_small.py`: 模型配置文件路径。\n*   `--classes-map`: 类别映射文件，用于将输出索引转换为具体的类别名称（如 ImageNet 标签）。\n\n**预期输出：**\n终端将显示模型对图片的分类结果，包括预测类别、置信度分数等信息。\n\n---\n*更多高级功能（如训练自定义数据集、可视化热力图、计算 Flops 等）请参考项目 `datas\u002Fdocs` 目录下的详细教程文档。*","某初创团队正在开发一款基于移动端的花卉识别应用，需要在有限的算力下快速验证多种主流骨干网络（Backbone）的精度与速度平衡。\n\n### 没有 Awesome-Backbones 时\n- **模型复现成本高**：工程师需手动从不同论文仓库扒取代码，花费数天时间统一数据加载、训练循环和评估接口，极易因版本不兼容导致环境崩溃。\n- **调参盲目且低效**：面对小数据集训练效果差的问题，缺乏系统性的排查指引，往往在错误的图像增强策略上浪费大量时间，不知如何关闭干扰项。\n- **性能对比困难**：难以在同一框架下公平对比 MobileNet、EfficientNet 或 ViT 等不同架构的 FLOPs 与准确率，每次切换模型都需重写配置文件。\n- **可视化功能缺失**：想要分析模型“关注”花朵的哪个部位以优化误判，需额外编写复杂的类别激活图（CAM）脚本，开发周期被拉长。\n\n### 使用 Awesome-Backbones 后\n- **开箱即用的集成**：直接调用内置的 MobileViT、EfficientNetV2 等 30+ 种预置模型，统一了 PyTorch 接口，将模型验证周期从数天缩短至几小时。\n- **精准的调试指引**：利用官方提供的“训练失败排查指南”，快速定位并关闭了不适合小样本的增强操作，显著提升了收敛速度和最终精度。\n- **一键式基准测试**：通过修改简单配置文件即可自动输出 Train\u002FVal 的 Loss、Accuracy 及 F1-score 报表，并自动计算参数量，轻松选出最适合移动端的模型。\n- **内置可解释性工具**：直接运行自带脚本即可生成类别激活图，直观展示模型决策依据，帮助团队迅速迭代优化特征提取能力。\n\nAwesome-Backbones 通过标准化流程与丰富的预置组件，让开发者从繁琐的代码整合中解放出来，专注于核心算法的选型与优化。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FFafa-DL_Awesome-Backbones_6858ff0e.png","Fafa-DL","Fafa Bro","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FFafa-DL_fa48721e.jpg","B站：\r\nhttps:\u002F\u002Fspace.bilibili.com\u002F46880349\r\n\r\nCSDN：https:\u002F\u002Fblog.csdn.net\u002Fzzh516451964zzh\r\n\r\n公众号 :  啥都会一点的研究生",null,"https:\u002F\u002Fgithub.com\u002FFafa-DL",[79],{"name":80,"color":81,"percentage":82},"Python","#3572A5",100,1929,276,"2026-04-02T15:51:16","未说明","未说明 (基于 PyTorch，通常建议配备 NVIDIA GPU 以加速训练，具体显存需求取决于模型大小)",{"notes":89,"python":90,"dependencies":91},"README 中仅明确列出了 PyTorch (>=1.7.1) 和 Python (>=3.6) 的版本要求。未提及具体的操作系统、GPU 型号、显存大小或内存需求。项目支持多种图像分类骨干网络（如 ResNet, ViT, MobileNet 等），不同模型对硬件资源的需求差异较大。快速开始部分提到需要下载预训练权重文件。",">=3.6",[92],"PyTorch>=1.7.1",[14,35,15],[95,96,97,98,99,100,101,102],"pytorch","image-classification","transformer","cnn","pytorch-classification","deep-learning","resnet","swin-transformer","2026-03-27T02:49:30.150509","2026-04-08T03:57:07.338048",[106,111,116,121,126,131,136],{"id":107,"question_zh":108,"answer_zh":109,"source_url":110},23711,"为什么模型评估结果每次运行都不一样，或者与训练时保存的精度不一致？","这通常是由随机种子数设置不同导致的。如果多次执行评估脚本结果浮动，请检查是否固定了随机种子。此外，已修复评估时结果不稳定的问题，建议拉取最新代码重试。如果验证集评估精度比训练保存的最佳点略低（如低 2% 左右），在数据集未变动的情况下属于正常波动范围。","https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fissues\u002F31",{"id":112,"question_zh":113,"answer_zh":114,"source_url":115},23712,"调用 evaluation 脚本时提示\"The model and loaded state dict do not match exactly\"且准确率差异大，如何解决？","出现该报错通常是因为没有使用正确的权重文件或配置文件未对应修改。请仔细检查配置文件中的权重路径，确保加载的是与自己训练模型结构完全匹配的权重文件。如果是用自己训练的权重进行评估，理论上不应出现 mismatch 问题，请确认配置文件中调用的权重路径是否正确。","https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fissues\u002F104",{"id":117,"question_zh":118,"answer_zh":119,"source_url":120},23713,"使用该项目训练 Swin Transformer 等模型时，精度比官方源码低很多，可能是什么原因？","精度较低可能与数据增强策略有关。建议尝试选择性关闭一些图像增强操作，经测试某些增强操作在特定数据集上反而会污染数据导致精度下降。此外，学习率（lr）的设置也非常关键，可以尝试调整为 1e-4 等其他数值进行调试。","https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fissues\u002F57",{"id":122,"question_zh":123,"answer_zh":124,"source_url":125},23714,"训练过程中因内存不足（OOM）导致中断，且 log 文件夹下无对应的.pth 文件，如何处理？","这是因为显存（memory）不够导致的。解决方法是降低 batch_size，同时记得同步调整学习率（通常 batch_size 减小，学习率也应相应减小）以保持训练稳定性。","https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fissues\u002F26",{"id":127,"question_zh":128,"answer_zh":129,"source_url":130},23715,"导入模块时报错\"ModuleNotFoundError: No module named 'utils.train_utils'\"怎么办？","该错误通常不是模块缺失，而是标注文件（如 annotations.txt）格式问题。请检查标注文件中是否存在多余的空格（space），特别是在行首或索引位置，删除多余空格后通常即可解决。","https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fissues\u002F21",{"id":132,"question_zh":133,"answer_zh":134,"source_url":135},23716,"如何扩展模型结构（如增加新的 Head 或多模态输入）而不破坏原始代码？","如果在组件中新增层或更改结构，教程建议新建文件以保护原始版本，并将新模块添加至__init__.py 进行注册。如果完全定义了一个新的 Head，也应新建文件并注册。若需支持多输出头（Multi-head）或多模态任务，通常需要修改 Loss 计算部分。为了降低修改成本，可以参考 MMCV 等框架的设计思路，通过配置文件灵活调整，但深度定制仍需阅读源码并进行相应改动。","https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fissues\u002F11",{"id":137,"question_zh":138,"answer_zh":139,"source_url":140},23717,"如何将训练过程中的 Loss 和精度信息保存下来以便绘图分析？","目前项目已支持保存 train 与 val 相关的 metrics 和 loss 信息。通常这些信息会被记录在日志文件或特定的输出文件中，方便后续提取数据进行画图分析。如果需要更详细的 CSV 格式记录或控制台输出同步到日志文件，可以查看最新代码更新或提交功能请求。","https:\u002F\u002Fgithub.com\u002FFafa-DL\u002FAwesome-Backbones\u002Fissues\u002F2",[]]