(Image from http://188.138.127.15:81/Datasets/Market-1501-v15.09.15.zip)

Shape : (batch, 3, height, width)

Automatically downloads the onnx and prototxt files on the first run. It is necessary to be connected to the Internet while downloading.

For the sample image,

$ python3 abd_net.py

If you want to specify the input image, put the image path after the --input option.
You can use --savepath option to change the name of the output file to save.

$ python3 abd_net.py --input IMAGE_PATH --savepath SAVE_IMAGE_PATH

If you want to specify the directory of gallery image, put the directory path after the --gallery_dir option.

$ python3 abd_net.py --gallery_dir gallery

Now, files in this gallery directory are very restricted.
Many more files can be found in the bounding_box_test directory of Market-1501-v15.09.15.zip or DukeMTMC-VideoReID.zip.

Once the program run, a intermediate result file containing the features of the gallery image will be created.
By adding the intermediate result file name after the --data option, you can use the intermediate result of the previous inference.

$ python3 abd_net.py --data result_resnet50.npy

By adding the model name after the --model option, you can specify the model.
The model name is selected from 'market1501', 'duke', 'msmt17'.

$ python3 abd_net.py --model market1501

• Attentive but Diverse Person Re-Identification

Pytorch

ONNX opset=11

abd_net_market1501.onnx.prototxt
abd_net_duke.prototxt
abd_net_msmt17.prototxt