I'm new to addons, and I'm trying to use the schedule sampling in my model, which now is implemented by tf2.1 functional api.
I have a few questions about it:
Can I use a multi layer rnn with the tfa sampler(e.g. two layer lstm with a sampler)?
Where can I find a full example with schedule sampler?
During inference, should I just use the inference sampler?