NOTE: One thing to remember is that if your network is not a sequential, it will need to be saved in a separate file and available in the python path

[shexuan]

As noted in documentation, why non-sequential network have to saved in a seperate file? When I deploy a pytorch model over pyspark to inference, I also met the problem —— sequential network can work and non-sequential network failed. How dose this happend and how to deal with this?

Any suggestions would be greatly appreciated!