How to use tensorflow serving function post conversion to ONNX

[chaitanya-basava]

Hi, wanted to check how we can use the serving function that can be saved using tf.saved_model.save function

sample code of what I am referring to:

tf.saved_model.save(
        model, export_dir="some path", signatures={'predict': serving_function}
)

Would the serving_function saved as a signature be accessible from post onnx conversion? it would be helpful if I can pointed to a document around this if any.

In case it won't be accessible then any alternative suggestions that I can pursue.

Thanks in advance.