def extract(cls, node):
     encoding_map = {0: 'corner', 1: 'center'}
     center_point_box = onnx_attr(node, 'center_point_box', 'i', default=0)
     NonMaxSuppression.update_node_stat(node, {'sort_result_descending': 0,
                                               'output_type': np.int64,
                                               'box_encoding': encoding_map[center_point_box]})
     return cls.enabled
 def extract(cls, node):
     attrs = {
         'sort_result_descending': 1,
         'box_encoding': 'corner',
         'output_type': np.int32
     }
     NonMaxSuppression.update_node_stat(node, attrs)
     return cls.enabled
 def extract(cls, node):
     pad_to_max_output_size = node.pb.attr["pad_to_max_output_size:"].b
     if not pad_to_max_output_size:
         log.warning('The attribute "pad_to_max_output_size" of node {} is equal to False which is not supported.'
                     'Forcing it to be equal to True'.format(node.soft_get('name')))
     attrs = {'sort_result_descending': 1, 'box_encoding': 'corner', 'output_type': np.int32}
     NonMaxSuppression.update_node_stat(node, attrs)
     return cls.enabled
 def extract(cls, node):
     attrs = {'sort_result_descending': 1, 'center_point_box': 0, 'output_type': np.int32}
     NonMaxSuppression.update_node_stat(node, attrs)
     return cls.enabled