Cannot directly use Prebuilt Models in other models

[zhengqing-cn]

Hi,

I want to use the prebuilt conv ODE blocks in a the network. However, there always shows below error:

Traceback (most recent call last): File "C:\Program Files\JetBrains\PyCharm 2021.1.1\plugins\python\helpers\pydev_pydevd_bundle\pydevd_exec2.py", line 3, in Exec exec(exp, global_vars, local_vars) File "", line 1, in File "C:\Users\ZhengQing\anaconda3\envs\tensorflow-gpu_115\lib\site-packages\tensorflow_core\python\keras\engine\base_layer.py", line 869, in call outputs = self._symbolic_call(inputs) File "C:\Users\ZhengQing\anaconda3\envs\tensorflow-gpu_115\lib\site-packages\tensorflow_core\python\keras\engine\base_layer.py", line 2158, in _symbolic_call output_shapes = self.compute_output_shape(input_shapes) File "C:\Users\ZhengQing\anaconda3\envs\tensorflow-gpu_115\lib\site-packages\tensorflow_core\python\keras\engine\network.py", line 699, in compute_output_shape return super(Network, self).compute_output_shape(input_shape) File "C:\Users\ZhengQing\anaconda3\envs\tensorflow-gpu_115\lib\site-packages\tensorflow_core\python\keras\engine\base_layer.py", line 646, in compute_output_shape raise NotImplementedError NotImplementedError

And Here is my code:

from tensorflow.keras.models import Model from tensorflow.keras.layers import Conv2D, Dense, Flatten, Input, MaxPooling2D, Layer, Lambda, Conv2DTranspose, BatchNormalization, AveragePooling2D, UpSampling2D, concatenate

from tfdiffeq.models import Conv2dODENet

def test_model(input_layer): x = Conv2D(8, (3, 3), activation="relu", kernel_initializer='he_normal', padding="same")(input_layer) x = Conv2D(8, (3, 3), activation="relu", kernel_initializer='he_normal', padding="same")(x) output_layer = Conv2dODENet(num_filters = 2, augment_dim=0, time_dependent=False)(x)

model = Model(inputs=input_layer, outputs=output_layer)

model.compile(...)

return model

input_layer = Input(shape=(160, 208, 4)) model = test_model(input_layer)

Could you kindly help me to check whether my implementation is correct? Many thanks in advance!

Best, ZQ