Feature/export docs

[DatSplit]

Description

Added a docs folder in the root of the repository with documentation about exporting trained models as suggested by @SkalskiP in #170:

rf-detr /
├── docs /
│   └── export.md
├── README.md
...

With the following subsections:

• ONNX export
• TensorRT conversion
• Deploying on Jetson Orin Nano (8 GB developer kit)

Type of change

Documentation only

How has this change been tested, please provide a testcase or example of how you tested the change?

Ran the Python snippets on my Jetson Orin Nano 8 GB developer kit (Ubuntu: 22.04.4 LTS, TensorRT-10.0.1.6, CUDA 12.0, Python 3.10)

Any specific deployment considerations

No

Docs

This feature only adds docs!

[CLAassistant]

All committers have signed the CLA.

[DatSplit]

Good afternoon @SkalskiP,

As mentioned by you in #170 "the sooner the better. It can be really rough. I can help you out with structuring it properly. I mostly care about the steps to follow."

I believe there is still a bug in the newly created code snippet in (Deploying on Jetson Orin Nano (8 GB developer kit)).
Will work on this tonight or tomorrow (first check whether there is actually a bug), as I got to go now.

Questions:

1. (ONNX export) more details on other possible arguments that can be passed to the export() method?
2. (TensorRT conversion) More details required on how to set-up TensorRT correctly?
3. Other: might it be benifical to add a tensorRT export option to model.export()?
4. Should I explain the run_sync() method in more detail?

Any feedback is welcome!

[DatSplit]

I have read the CLA Document and I sign the CLA.