search

MILVLG / bottom-up-attention.pytorch

A PyTorch reimplementation of bottom-up-attention models
Apache License 2.0
294 stars 76 forks source link

where to get the box feature? the bbox_file is empty. how to gererate the box feature from image? #115

Open ekma000 opened 8 months ago

ekma000 commented 8 months ago

` elif cfg.MODEL.BUA.EXTRACTOR.MODE == 3: # extract roi features by bbox npy = False bbox_path = os.path.join(args.bbox_dir, im_file.split('.')[0]) if os.path.exists(bbox_path + '.npy'): npy = True bbox_file = bbox_path + '.npy' elif os.path.exists(bbox_path + '.npz'): bbox_file = bbox_path + '.npz' else: continue

        try:
            if npy:
                bbox = torch.from_numpy(np.load(bbox_file, allow_pickle=True).tolist()['bbox']) * dataset_dict['im_scale']
            else:
                bbox = torch.from_numpy(np.load(bbox_file)['bbox']) * dataset_dict['im_scale']
        except Exception as e:
            print(e)
            continue`

where to get the box feature? the bbox_file is empty. how to gererate the box feature from image?