Is it possible to set up a joint probability distribution with learnable parameters using with keras? Does anybody can show me how to define the output layers?
Hi @ch3f , Yes it is possible with the help of Tf probability it allows you to define probabilistic models with learnable parameters, and you can integrate this with keras by defining output layers that represent distributions, such as normal, Bernoulli, or more complex distributions .
kindly refer to this documentation for better understanding .
Thank You