New FINN setup on dev

[auphelia]

For all FINN-users that work with dev:

We just merged PR-844 that is changing the setup of FINN

• The Docker environment was updated to Ubuntu 22.04
• Python was updated to 3.10
• Brevitas was updated to latest master (10/07/2023)
  • As a result of this change, we have completely switched to the QONNX export. Internally FINN still uses the FINN-ONNX format. Please have a look at the QONNX export tutorial for more information and check out this section in the documentation.
• Other dependencies were also updated, e.g. numpy, onnxruntime. (for more details, see requirements.txt)
• With these updates, we also moved to Vivado/Vitis Version 2022.2 and implemented experimental support for 2023.1 on dev. The xrt version that is installed in the FINN Docker container is 2022.2

Thanks to @fpjentzsch for his work to make this PR possible!

I would also like to take the chance to make you aware of a new tutorial we added in the past weeks!
The tutorial allows you to vary the folding factors for a small network and to investigate the consequences (in terms of resources and performance) for the resulting FINN accelerator. In addition to the interactive Jupyter Notebook, documentation has been added to summarize the constraints for setting the folding factors.

Thanks to @shashwat1198 for his contribution!

Have fun experimenting with the new features!