AttributeError: 'Tensor' object has no attribute 'numpy' while loading keras model

Question

I'm using a pretrained yolov4 keras model for object detection in tf2 So I loaded the model and call it like

new_model = tf.keras.models.load_model('./checkpoints/yolov4-416')
pred_bbox = new_model(image_data)
for key, value in pred_bbox.items():

then an error occurred:

AttributeError: 'Tensor' object has no attribute 'items'

So I modified my code

new_model = tf.keras.models.load_model('./checkpoints/yolov4-416')
prediction_list = new_model.predict(image_data)
prediction_dict = {name: pred for name, pred in zip(new_model.output_names, prediction_list)}
for key, value in pred_bbox.items():
    boxes = value[:, :, 0:4]
    pred_conf = value[:, :, 4:] 

boxes, scores, classes, valid_detections = tf.image.combined_non_max_suppression(
    boxes=tf.reshape(boxes, (tf.shape(boxes)[0], -1, 1, 4)),
    scores=tf.reshape( pred_conf,
    (tf.shape(pred_conf)[0], -1, tf.shape(pred_conf)[-1])),
    max_output_size_per_class=50,
    max_total_size=50,
    iou_threshold=FLAGS.iou,
    score_threshold=FLAGS.score
)
pred_bbox = [boxes.numpy(), scores.numpy(), classes.numpy(), valid_detections.numpy()]
image = utils.draw_bbox(original_image, pred_bbox)

It was doing great until box.numpy() is called and another error occurred

AttributeError: 'Tensor' object has no attribute 'numpy'

I wonder how to solve this, and I think it's because it is in graph mode instead of eager execution?

Yes. It's because of graph mode. But are you using tf1 or disabled eager mode in tf2 or using @tf.function? — Kaveh
As @BestDogeStackoverflow said, you may called boxes.numpy() where it may run in graph mode. So please add complete code. — Kaveh

jimmy12487 jimmy12487 · Accepted Answer · 2021-08-09T15:19:33

OK I just kind of solve the problem When I'm saving model, it's somehow saved in graph mode, which is why I can't call boxes.numpy(). So what I do is just load weights and rebuild the model and the problem is just gone. But I still want to know what makes it wrong

save_model code:

STRIDES, ANCHORS, NUM_CLASS, XYSCALE = utils.load_config(FLAGS)
input_layer = tf.keras.layers.Input([FLAGS.input_size, FLAGS.input_size, 3])
feature_maps = YOLO(input_layer, NUM_CLASS, FLAGS.model, FLAGS.tiny)
bbox_tensors = []
prob_tensors = []
for i, fm in enumerate(feature_maps):
  if i == 0:
    output_tensors = decode(fm, FLAGS.input_size // 8, NUM_CLASS, STRIDES, ANCHORS, i, XYSCALE, FLAGS.framework)
  elif i == 1:
    output_tensors = decode(fm, FLAGS.input_size // 16, NUM_CLASS, STRIDES, ANCHORS, i, XYSCALE, FLAGS.framework)
  else:
    output_tensors = decode(fm, FLAGS.input_size // 32, NUM_CLASS, STRIDES, ANCHORS, i, XYSCALE, FLAGS.framework)
  bbox_tensors.append(output_tensors[0])
  prob_tensors.append(output_tensors[1])
pred_bbox = tf.concat(bbox_tensors, axis=1)
pred_prob = tf.concat(prob_tensors, axis=1)
boxes, pred_conf = filter_boxes(pred_bbox, 
  pred_prob,`score_thershold=FLAGS.score_thres, 
  input_shape=tf.constant([FLAGS.input_size, FLAGS.input_size]))
  pred = tf.concat([boxes, pred_conf], axis=-1)
model = tf.keras.Model(input_layer, pred)
utils.load_weights(model, FLAGS.weights, FLAGS.model, FLAGS.tiny)
model.summary()
model.save(FLAGS.output)

AttributeError: 'Tensor' object has no attribute 'numpy' while loading keras model

1 Answers