ImageNet and pytorch pretrained model id mapping

Showing the relationship between ImageNet ID and labels and pytorch pre-trained model output ID and labels.

Requirments

See file requirment.txt.

Show the mapping

Open jupyter notebook imagenet_and_pytorch_pretrained_model_id_mapping.ipynb and you will see.

In "data" filefolder

Filefolder sample_1000 contains sample images. Image name format: (ImageNet_ID)_(WNID).JPEG. I only upload 100 sample images to GitHub. If you want full version of sample_1000 (1000 sample images, one image each class), please download them from https://drive.google.com/drive/folders/1XF_n21S1OFyLO61ruHwJd_tj3SlwNuAd?usp=sharing.

File imagenet_id_wnid_class.mat gives ImageNet IDs, WNIDs, and class names.

File modeloutput_id_wnid_class.json gives pytorch pre-trained model output IDs, WNIDs, and class names.

ImageNet ID和pytorch预训练模型输出ID的对应关系

需要的python包

见文件requirment.txt

展示

打开jupyter notebook imagenet_and_pytorch_pretrained_model_id_mapping.ipynb

在"data"文件夹中

文件夹sample_1000中有示例图片，其命名格式为：(ImageNet_ID)_(WNID).JPEG；我仅上传了100张示例图片；如需要完整1000张图，请点击如下链接下载 https://drive.google.com/drive/folders/1XF_n21S1OFyLO61ruHwJd_tj3SlwNuAd?usp=sharing

文件imagenet_id_wnid_class.mat: ImageNet ID，WNID和类别名称

文件modeloutput_id_wnid_class.json: pytorch pre-trained model output ID，WNID和类别名称