Error when Running on Higher-Resolution Images

[aryker]

I am trying to get this working with images outside the ones in the example folder, and I am running into issues whenever I try to use images that are of a resolution other than 640x480. For example, when trying an image of resolution 5184 x 3456 I get the following stack trace:

File "E:\Data\miniconda3\lib\site-packages\keras\engine\training.py", line 1845, in predict_function  *
        return step_function(self, iterator)
    File "E:\Data\miniconda3\lib\site-packages\keras\engine\training.py", line 1834, in step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    File "E:\Data\miniconda3\lib\site-packages\keras\engine\training.py", line 1823, in run_step  **
        outputs = model.predict_step(data)
    File "E:\Data\miniconda3\lib\site-packages\keras\engine\training.py", line 1791, in predict_step
        return self(x, training=False)
    File "E:\Data\miniconda3\lib\site-packages\keras\utils\traceback_utils.py", line 67, in error_handler
        raise e.with_traceback(filtered_tb) from None
    File "E:\Data\miniconda3\lib\site-packages\keras\engine\input_spec.py", line 264, in assert_input_compatibility
        raise ValueError(f'Input {input_index} of layer "{layer_name}" is '

    ValueError: Input 0 of layer "model_1" is incompatible with the layer: expected shape=(None, None, None, 3), found shape=(None, 3456, 5184, 4)

As another example, when I try an image of resolution 1920 x 1080 I see the following:

File "E:\Data\miniconda3\lib\site-packages\keras\engine\training.py", line 1845, in predict_function  *
        return step_function(self, iterator)
    File "E:\Data\miniconda3\lib\site-packages\keras\engine\training.py", line 1834, in step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    File "E:\Data\miniconda3\lib\site-packages\keras\engine\training.py", line 1823, in run_step  **
        outputs = model.predict_step(data)
    File "E:\Data\miniconda3\lib\site-packages\keras\engine\training.py", line 1791, in predict_step
        return self(x, training=False)
    File "E:\Data\miniconda3\lib\site-packages\keras\utils\traceback_utils.py", line 67, in error_handler
        raise e.with_traceback(filtered_tb) from None
    File "E:\Data\miniconda3\lib\site-packages\keras\backend.py", line 3313, in concatenate
        return tf.concat([to_dense(x) for x in tensors], axis)

    ValueError: Exception encountered when calling layer "up1_concat" (type Concatenate).
    
    Dimension 1 in both shapes must be equal, but are 66 and 67. Shapes are [?,66,120] and [?,67,120]. for '{{node model_1/up1_concat/concat}} = ConcatV2[N=2, T=DT_FLOAT, Tidx=DT_INT32](model_1/up1_upsampling2d/resize/ResizeBilinear, model_1/pool3_pool/AvgPool, model_1/up1_concat/concat/axis)' with input shapes: [?,66,120,1664], [?,67,120,256], [] and with computed input tensors: input[2] = <3>.
    
    Call arguments received by layer "up1_concat" (type Concatenate):
      • inputs=['tf.Tensor(shape=(None, 66, 120, 1664), dtype=float32)', 'tf.Tensor(shape=(None, 67, 120, 256), dtype=float32)']