RuntimeError: The layer has never been called and thus has no defined output shape.

[PRBtc]

Hi I trying to train train.py with backend model as Full Yolo or ResNet50. But I encounter the following error. CCould you please let me know what could be the error here. tHanks

Traceback (most recent call last):
File "train.py", line 101, in
main(args)
File "train.py", line 70, in main
anchors = config['model']['anchors'])
File "/Users/priyanka/Desktop/Yolo/YOLO_V2_Resnet/keras-yolo2/frontend.py", line 60, in init
print(self.feature_extractor.get_output_shape())
File "/Users/priyanka/Desktop/Yolo/YOLO_V2_Resnet/keras-yolo2/backend.py", line 33, in get_output_shape
return self.feature_extractor.get_output_shape_at(-1)[1:3]
File "//anaconda3/lib/python3.7/site-packages/tensorflow/python/keras/engine/base_layer.py", line 2040, in get_output_shape_at
'output shape')
File "//anaconda3/lib/python3.7/site-packages/tensorflow/python/keras/engine/base_layer.py", line 2631, in _get_node_attribute_at_index
'and thus has no defined ' + attr_name + '.')
RuntimeError: The layer has never been called and thus has no defined output shape.

[jia0511]

+1

[EduardoHT]

with open(config_path) as config_buffer:
config = json.load(config_buffer)

###############################

Make the model

###############################

yolo = YOLO(backend = config['model']['backend'],
input_size = config['model']['input_size'],
labels = config['model']['labels'],
max_box_per_image = config['model']['max_box_per_image'],
anchors = config['model']['anchors'])

###############################

Load trained weights

###############################

yolo.load_weights(weights_path)

Output:::::::::::::::::::::::::::::::::

RuntimeError Traceback (most recent call last)
~\AppData\Local\Temp/ipykernel_6012/3719954046.py in
6 ###############################
7
----> 8 yolo = YOLO(backend = config['model']['backend'],
9 input_size = config['model']['input_size'],
10 labels = config['model']['labels'],

~\Eduardo\ModelosProntosSingapura\keras-yolo2-master\frontend.py in init(self, backend, input_size, labels, max_box_per_image, anchors)
56 raise Exception('Architecture not supported! Only support Full Yolo, Tiny Yolo, MobileNet, SqueezeNet, VGG16, ResNet50, and Inception3 at the moment!')
57
---> 58 print(self.feature_extractor.get_output_shape())
59 self.grid_h, self.grid_w = self.feature_extractor.get_output_shape()
60 features = self.feature_extractor.extract(input_image)

~\Eduardo\ModelosProntosSingapura\keras-yolo2-master\backend.py in get_output_shape(self)
29
30 def get_output_shape(self):
---> 31 return self.feature_extractor.get_output_shape_at(-1)[1:3]
32
33 def extract(self, input_image):

~\Anaconda3\envs\TesteComYolo\lib\site-packages\keras\engine\base_layer.py in get_output_shape_at(self, node_index)
2075 RuntimeError: If called in Eager mode.
2076 """
-> 2077 return self._get_node_attribute_at_index(node_index, 'output_shapes',
2078 'output shape')
2079

~\Anaconda3\envs\TesteComYolo\lib\site-packages\keras\engine\base_layer.py in _get_node_attribute_at_index(self, node_index, attr, attr_name)
2681 """
2682 if not self._inbound_nodes:
-> 2683 raise RuntimeError(f'The layer {self.name} has never been called '
2684 'and thus has no defined {attr_name}.')
2685 if not len(self._inbound_nodes) > node_index:

RuntimeError: The layer model has never been called and thus has no defined {attr_name}.