Added support for QONNX Resize node ingestion and tested with tiny UNet model

[nghielme]

Description

All the necessary parts are added in order to properly ingest QONNX models containing Resize node

Type of change

Extended the capability of QONNX ingestion of hls4ml

• New feature (non-breaking change which adds functionality)

Tests

I added two tests based on tiny UNet model in test_qonnx.py and one based on a simpler model taken from tiny UNet model. They have been both added in pytest directory.

Checklist

• I have read the guidelines for contributing.
• I have commented my code, particularly in hard-to-understand areas.
• I have made corresponding changes to the documentation.
• My changes generate no new warnings.
• I have installed and run pre-commit on the files I edited or added.
• I have added tests that prove my fix is effective or that my feature works.

[jmitrevs]

Isn't this identical?

[nghielme]

Here the output of the tests if you run the code with the original line:

    def match(self, node):
        if isinstance(node, (Pooling1D, Pooling2D, GlobalPooling1D, GlobalPooling2D)) and node.get_attr('pool_op') == 'Max':
>           return isinstance(node.get_input_variable().type.precision, XnorPrecisionType) and not isinstance(
                node.get_output_variable().type.precision, XnorPrecisionType
            )
E           AttributeError: 'FixedPrecisionType' object has no attribute 'precision'

hls4ml/backends/fpga/passes/xnor_pooling.py:14: AttributeError

Solved if you update it as proposed.

[nghielme]

I think I addressed all the comments. I am noticing the CI/CD pipeline fails in 2 tests, 1 and 2, but apparently they are unrelated to this contribution.

[jmitrevs]

Note, this now includes part (but not all) of #1130.

[nghielme]

I reviewed all the comments, let me know if there's anything else to address. I see the pipeline failed but again it was due to the large time taken by some tests