Hello! It could be great write private neural networks granted by PySyft using the modern best practices implemented in FastAI. I tried PySyft 0.2 alongside with FastAI (v1) and they are not perfectly compatible each other. At present, developers are working on PySyft 0.3, so I ask them if they can support FastAI 2 on the new version of PySyft. They have given their ok. Do you think also is possible make compatible the new PySyft 0.3 with FastAI v2?
If you’re interested in it then go for it