[add] modelExporters

[amkrajewski]

• This PR introduces the modelExporters module with 3 Exporter classes, allowing conversion and saving of the loaded models (including ones modified within pySIPFENN transfer learning runs) into ONNX, PyTorch, and CoreML formats.
• The ONNXExporter also allows for model conversion to FP16 precision, lowering the size by a factor of 2, as well as model simplification with the ONNX optimizer toolset.

[amkrajewski]

TODO:

• Testing the runtime of the 3 exporters
• Usage examples in the docs.

[codecov[bot]]

Codecov Report

Attention: 3 lines in your changes are missing coverage. Please review.

Comparison is base (45202a4) 92.20% compared to head (1cc47f2) 93.22%.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #8 +/- ## ========================================== + Coverage 92.20% 93.22% +1.01% ========================================== Files 12 14 +2 Lines 1181 1402 +221 ========================================== + Hits 1089 1307 +218 - Misses 92 95 +3 ``` | [Files](https://app.codecov.io/gh/PhasesResearchLab/pySIPFENN/pull/8?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=PhasesResearchLab) | Coverage Δ | | |---|---|---| | [pysipfenn/\_\_init\_\_.py](https://app.codecov.io/gh/PhasesResearchLab/pySIPFENN/pull/8?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=PhasesResearchLab#diff-cHlzaXBmZW5uL19faW5pdF9fLnB5) | `100.00% <100.00%> (ø)` | | | [pysipfenn/core/pysipfenn.py](https://app.codecov.io/gh/PhasesResearchLab/pySIPFENN/pull/8?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=PhasesResearchLab#diff-cHlzaXBmZW5uL2NvcmUvcHlzaXBmZW5uLnB5) | `83.58% <100.00%> (+0.51%)` | :arrow_up: | | [pysipfenn/tests/test\_ModelExporters.py](https://app.codecov.io/gh/PhasesResearchLab/pySIPFENN/pull/8?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=PhasesResearchLab#diff-cHlzaXBmZW5uL3Rlc3RzL3Rlc3RfTW9kZWxFeHBvcnRlcnMucHk=) | `100.00% <100.00%> (ø)` | | | [pysipfenn/tests/test\_pysipfenn.py](https://app.codecov.io/gh/PhasesResearchLab/pySIPFENN/pull/8?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=PhasesResearchLab#diff-cHlzaXBmZW5uL3Rlc3RzL3Rlc3RfcHlzaXBmZW5uLnB5) | `95.41% <100.00%> (ø)` | | | [pysipfenn/core/modelExporters.py](https://app.codecov.io/gh/PhasesResearchLab/pySIPFENN/pull/8?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=PhasesResearchLab#diff-cHlzaXBmZW5uL2NvcmUvbW9kZWxFeHBvcnRlcnMucHk=) | `97.85% <97.85%> (ø)` | |

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[amkrajewski]

So if the calculator is empty it is not an error?

Based on the current PR state, it does raise an AssertionError with the message 'No models loaded in calculator. Nothing to export.'

I will add tests to verify these errors occur as intended.

[amkrajewski]

I added the aforementioned tests and documentation page on why-and-how to use Exporters.