open-aff:注意特征融合的代码和训练有素的模型

上传者: 42180863 | 上传时间: 2023-04-03 16:44:16 | 文件大小: 386.7MB | 文件类型: ZIP
注意特征融合 用于“注意特征融合”的MXNet / Gluon代码 到目前为止,此仓库中有什么: ImageNet的代码,训练有素的模型和训练日志 PS: 如果您是我们提交的论文的审稿人,请注意,当前实现的准确性比本文中的准确性要高一些,因为它是一个带有很多技巧的新实现。 如果您是我的学位论文评估专家,发现论文与这个repo的数字有些出入,那是因为在论文提交后我又将代码重新实现一遍遍,添加了AutoAugment,Labelinging这些技巧,因此目前这个repo中的分类准确率会比论文中的数字高一些,还请见谅。 更改日志: 2020-10-08:通过一整套技巧重新实现图像分类代码 2020-09-29:上载所提交论文的图像分类代码和训练有素的模型 去做: 在ImageNet上运行AFF-ResNeXt-50和AFF-ResNet-50 在新的训练模型上更新Grad-CAM结果 重新实

文件下载

资源详情

[{"title":"( 36 个子文件 386.7MB ) open-aff:注意特征融合的代码和训练有素的模型","children":[{"title":"open-aff-master","children":[{"title":"train_imagenet.py <span style='color:#111;'> 23.74KB </span>","children":null,"spread":false},{"title":"model","children":[{"title":"fusion.py <span style='color:#111;'> 4.26KB </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 123B </span>","children":null,"spread":false},{"title":"classification.py <span style='color:#111;'> 18.96KB </span>","children":null,"spread":false},{"title":"block.py <span style='color:#111;'> 8.56KB </span>","children":null,"spread":false}],"spread":true},{"title":"aff_pytorch","children":[{"title":"README_CN.md <span style='color:#111;'> 1.42KB </span>","children":null,"spread":false},{"title":"aff_net","children":[{"title":"fusion.py <span style='color:#111;'> 5.03KB </span>","children":null,"spread":false},{"title":"aff_resnet.py <span style='color:#111;'> 14.12KB </span>","children":null,"spread":false}],"spread":true},{"title":"README.md <span style='color:#111;'> 1.45KB </span>","children":null,"spread":false}],"spread":true},{"title":"params","children":[{"title":"cifar100","children":[{"title":"AFF-ResNet-32","children":[{"title":"superclasses","children":[{"title":"train_cifar100_cifar100-ASKCFuse-resnet-32-c-4-s-1.log <span style='color:#111;'> 58.63KB </span>","children":null,"spread":false},{"title":"From_Scratch_Log_train_cifar100_cifar100-ASKCFuse-resnet-32-c-4-s-1.log <span style='color:#111;'> 103.77KB </span>","children":null,"spread":false},{"title":"0.9078-cifar100-cifar100-ASKCFuse-resnet-32-c-4-s-1-best.params <span style='color:#111;'> 30.24MB </span>","children":null,"spread":false}],"spread":true},{"title":"train_cifar100-ASKCFuse-resnet-32-c-4-s-1.log <span style='color:#111;'> 59.13KB </span>","children":null,"spread":false},{"title":"0.8357-cifar100-ASKCFuse-resnet-32-c-4-s-1-best.params <span style='color:#111;'> 30.24MB </span>","children":null,"spread":false}],"spread":true},{"title":"AFF-ResNeXt-38-32x4d","children":[{"title":"train_cifar100-ASKCFuse-resneXt-38-c-4-s-1.log <span style='color:#111;'> 58.62KB </span>","children":null,"spread":false},{"title":"From_Scratch_Log_train_cifar100_cifar100-ASKCFuse-resneXt-38-c-4-s-1.log <span style='color:#111;'> 50.32KB </span>","children":null,"spread":false},{"title":"0.9123-cifar100-ASKCFuse-resneXt-38-c-4-s-1-best.params <span style='color:#111;'> 29.92MB </span>","children":null,"spread":false}],"spread":true},{"title":"AFF-ResNeXt-47-32x4d","children":[{"title":"0.8538-cifar100-ASKCFuse-resneXt-47-32x4-c-4-s-1-best.params <span style='color:#111;'> 36.99MB </span>","children":null,"spread":false},{"title":"train_cifar100-ASKCFuse-resneXt-47-32x4-c-4-s-1.log <span style='color:#111;'> 235.85KB </span>","children":null,"spread":false},{"title":"superclass","children":[{"title":"train_cifar100-ASKCFuse-resneXt-47-32x4-c-4-s-1.log <span style='color:#111;'> 58.64KB </span>","children":null,"spread":false},{"title":"0.9184-cifar100-ASKCFuse-resneXt-47-32x4-c-4-s-1-best.params <span style='color:#111;'> 36.99MB </span>","children":null,"spread":false},{"title":"From_Scratch_Log_train_cifar100-ASKCFuse-resneXt-47-c-4-s-1.log <span style='color:#111;'> 58.56KB </span>","children":null,"spread":false}],"spread":true}],"spread":true}],"spread":true},{"title":"imagenet","children":[{"title":"AFF-ResNet-50","children":[{"title":"aff_resnet50.log <span style='color:#111;'> 975.48KB </span>","children":null,"spread":false},{"title":"0.2028-imagenet-ImageNet-ASKCFuse-resnet--50-s-3-235-best.params <span style='color:#111;'> 58.57MB </span>","children":null,"spread":false}],"spread":true},{"title":"AFF-ResNeXt-50","children":[{"title":"0.2028-imagenet-ImageNet-ASKCFuse-resnext--50-s-3-231-best.params <span style='color:#111;'> 57.86MB </span>","children":null,"spread":false},{"title":"aff_resnext50.log <span style='color:#111;'> 975.58KB </span>","children":null,"spread":false}],"spread":true},{"title":"iAFF-ResNeXt-50","children":[{"title":"iaff_resnext50.log <span style='color:#111;'> 981.76KB </span>","children":null,"spread":false},{"title":"0.1978-imagenet-ImageNet-ResGlobLocaforGlobLocaCha-resnext--50-s-3-238-best.params <span style='color:#111;'> 67.33MB </span>","children":null,"spread":false}],"spread":true},{"title":"iAFF-ResNet-50","children":[{"title":"0.2022-imagenet-ImageNet-ResGlobLocaforGlobLocaCha-resnet--50-s-3-238-best.params <span style='color:#111;'> 68.04MB </span>","children":null,"spread":false},{"title":"iaff_resnet50.log <span style='color:#111;'> 975.50KB </span>","children":null,"spread":false}],"spread":true}],"spread":true}],"spread":true},{"title":"train_cifar.py <span style='color:#111;'> 16.28KB </span>","children":null,"spread":false},{"title":".gitignore <span style='color:#111;'> 12B </span>","children":null,"spread":false},{"title":"cmd_scripts.txt <span style='color:#111;'> 1.83KB </span>","children":null,"spread":false},{"title":"README.md <span style='color:#111;'> 5.97KB </span>","children":null,"spread":false},{"title":"utils","children":[{"title":"flops.py <span style='color:#111;'> 6.33KB </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 27B </span>","children":null,"spread":false}],"spread":true}],"spread":true}],"spread":true}]

评论信息

免责申明

【只为小站】的资源来自网友分享,仅供学习研究,请务必在下载后24小时内给予删除,不得用于其他任何用途,否则后果自负。基于互联网的特殊性,【只为小站】 无法对用户传输的作品、信息、内容的权属或合法性、合规性、真实性、科学性、完整权、有效性等进行实质审查;无论 【只为小站】 经营者是否已进行审查,用户均应自行承担因其传输的作品、信息、内容而可能或已经产生的侵权或权属纠纷等法律责任。
本站所有资源不代表本站的观点或立场,基于网友分享,根据中国法律《信息网络传播权保护条例》第二十二条之规定,若资源存在侵权或相关问题请联系本站客服人员,zhiweidada#qq.com,请把#换成@,本站将给予最大的支持与配合,做到及时反馈和处理。关于更多版权及免责申明参见 版权及免责申明