tensorflow: Best methods to debug and identify dynamic-sized tensors to work around: TensorFlow Lite Error: Attempting to use a delegate that only supports static-sized tensors with a graph that has dynamic-sized tensors?

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow):
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): macOS 11.2.3
• Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device: iPhone 12 pro
• TensorFlow installed from (source or binary): binary
• TensorFlow version (use command below): v1.12.1-53831-ga8b6d5ff93a 2.5.0-rc0
• Python version: 3.8

Describe the current behavior Converted Google’s Repnet Model to a tflite model to run on ios. I appear to be using too much ram to be able to run with the cpu, so I am trying to run on gpu. Except TensorFlow Lite Error: Attempting to use a delegate that only supports static-sized tensors with a graph that has dynamic-sized tensors. is being thrown at runtime on ios.

Describe the expected behavior I understand dynamic-sized tensors cannot be used with tflite (edit: Tflite GPU), but how can I debug to identity the use of dynamic-sized tensors?

Standalone code to reproduce the issue https://colab.research.google.com/github/google-research/google-research/blob/master/repnet/repnet_colab.ipynb

# converting to tflite with
model = get_repnet_model(PATH_TO_CKPT)

tf.keras.models.save_model( model,
    "repnet_savedmodel"
)

# Convert the model using TFLiteConverter
converter = tf.lite.TFLiteConverter.from_saved_model("repnet_savedmodel")
converter.optimizations = [tf.lite.Optimize.OPTIMIZE_FOR_SIZE] # .OPTIMIZE_FOR_SIZE]
converter.target_spec.supported_ops = [
    tf.lite.OpsSet.TFLITE_BUILTINS, # enable TensorFlow Lite ops.
    tf.lite.OpsSet.SELECT_TF_OPS # enable TensorFlow ops.
]
tflite_model = converter.convert()
with open("repnet.tflite", 'wb') as f:
  f.write(tflite_model)