PixelCNN example from the documentation does not work

How to repro?

TF version: 2.19.0

1. Take an example from https://www.tensorflow.org/probability/api_docs/python/tfp/distributions/PixelCNN#examples
2. Create Google Colab
3. Past the example into the Google Colab
4. Run it

The following exception occurs:

    ---------------------------------------------------------------------------
    ValueError                                Traceback (most recent call last)
    [/tmp/ipython-input-624558695.py](https://localhost:8080/#) in <cell line: 0>()
         33 
         34 # Define the log likelihood for the loss fn
    ---> 35 log_prob = dist.log_prob(image_input)
         36 
         37 # Define the model
    
    6 frames
    <decorator-gen-128> in shape_v2(input, out_type, name)
    
    [/usr/local/lib/python3.12/dist-packages/keras/src/backend/common/keras_tensor.py](https://localhost:8080/#) in __tf_tensor__(self, dtype, name)
        154 
        155     def __tf_tensor__(self, dtype=None, name=None):
    --> 156         raise ValueError(
        157             "A KerasTensor cannot be used as input to a TensorFlow function. "
        158             "A KerasTensor is a symbolic placeholder for a shape and dtype, "
    
    ValueError: A KerasTensor cannot be used as input to a TensorFlow function. A KerasTensor is a symbolic placeholder for a shape and dtype, used when constructing Keras Functional models or Keras Functions. You can only use it as input to a Keras layer or a Keras operation (from the namespaces `keras.layers` and `keras.ops`). You are likely doing something like:
    
    ```
    x = Input(...)
    ...
    tf_fn(x)  # Invalid.
    ```
    
    What you should do instead is wrap `tf_fn` in a layer:
    
    ```
    class MyLayer(Layer):
        def call(self, x):
            return tf_fn(x)
    
    x = MyLayer()(x)
    ```

Start with the PixelCNN example at the TensorFlow Probability API documentation URL and reproduce it in Google Colab with TensorFlow 2.19. Inspect the failure at dist.log_prob(image_input) and trace the reported KerasTensor exception. Done means the documented example runs successfully without this exception.