Owl_rllib.PolicyPolicies: how a network's output becomes a distribution and an action.
A Gaussian policy stores the unclipped sample, so log-probabilities stay those of the Gaussian; only the action sent to the environment is clipped to the action bounds.
type ('tensor, 'array, 'action) t = {sample : rng:Random.State.t -> 'tensor -> 'array;one sample per row of the output
*)log_prob : 'tensor -> 'tensor -> 'tensor;log_prob output samples, shape [batch; 1]
entropy : 'tensor -> 'tensor;actions : 'array -> 'action array;environment actions from samples
*)}module S : sig ... endmodule D : sig ... end