Refactor ONNXRuntime Python interface

[zhiqwang]

We introduce the yolort.runtime interface to accelerate the application on the Python frontend. For example, you can define the runtime and do inferencing with ONNXRuntime as follows

from yolort.runtime import PredictorORT

detector = PredictorORT("best.onnx")
img_path = "bus.jpg"
scores, class_ids, boxes = detector.run_on_image(img_path)

[codecov]

Codecov Report

Merging #176 (39efaa5) into master (f0746c6) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #176   +/-   ##
=======================================
  Coverage   96.48%   96.48%           
=======================================
  Files          10       10           
  Lines         598      598           
=======================================
  Hits          577      577           
  Misses         21       21           

Flag	Coverage Δ
unittests	96.48% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f0746c6...39efaa5. Read the comment docs.

[itsnine]

@zhiqwang hi, do you want to have that feature or you already started to write? I can do that feature if nobody already started doing it :)

[zhiqwang]

@zhiqwang hi, do you want to have that feature or you already started to write? I can do that feature if nobody already started doing it :)

Hi @itsnine , I haven't started this work yet, so far I've only worked on fixing some docstrings. We're welcome for this features, go ahead and feel free to start this job with a new PR. I'll hold this PR for fixing the docstrings. We can move forward to your PR first.

My current plan is to put the core code for ORT inference into the directory yolort/runtime. And furthermore we could name the python scripts as ort_modeling.py, you can also use others if you have a better suggestion. We plan to add more python inferencing backends in the future in this directory. And then add a CLI tool named run_model.py (just for example) to the directory tools for use cases.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
1 out of 2 committers have signed the CLA.

✅ zhiqwang
❌ pre-commit-ci[bot]
_{You have signed the CLA already but the status is still pending? Let us recheck it.}