Hello, the shape of the npy file I calculated through the compute_FMM_Pool function is (2, H, W). Are the FFM image given in the paper and the opts["weight_dir"] parameter options in train_mdnet_weight.py the information of the first dimension in the npy file?
Hello, the shape of the npy file I calculated through the compute_FMM_Pool function is (2, H, W). Are the FFM image given in the paper and the opts["weight_dir"] parameter options in train_mdnet_weight.py the information of the first dimension in the npy file?