Fixed exception on linear_quantize_activations for ImageType output

[Zerui18]

Issue

Calling cto.coreml.experimental.linear_quantize_activations on a model with ct.ImageType output(s) fails at the model output because check_intermediate_output expects output_value to be a np.ndarray but the model's final output is PIL.Image.

Solution

This patch adds a simple attempt to convert the output_value to np.ndarray before the computations.

[TobyRoseman]

Thanks for the fix. Can you add a unit test that fails without this fix, but passes with this fix?

[Zerui18]

Hi @TobyRoseman I've just discovered that the root cause of this issue is the function ModelDebugger.step which applies check_intermediate_output explicitly to all original model outputs as seen here:

coremltools/coremltools/optimize/coreml/experimental/_model_debugger.py

Lines 369 to 372 in 68b4f5a

	for (output_name, output_value) in model_outputs.items():
	operation = self.block_info.operations.get(output_name, None)
	if not step_fn(output_value, output_name, operation, activation_stats_dict):
	return

Hence, I believe rather than changing check_intermediate_output which is correctly expecting only np.ndarray, it's better to adjust ModelDebugger.step so that it doesn't call check_intermediate_output on the final model outputs. More concretely I'm thinking that we can add a new argument intermediate_outputs_only=False which can be set to True to skip the above block of code, whilst retaining the original functionality of step by default.

What do you think?

[TobyRoseman]

Thanks @Zerui18 for looking into this. I'm not very familiar with that part of the code. I think @yixingli-apple or @junpeiz should be able to help here.

[Zerui18]

Hi, is there any updates on this?