owl-rllib
Owl_rllib.On_policy
Rollout collection and updates shared by the actor-critic methods (Owl_rllib.A2c, Owl_rllib.Ppo) and by language-model RL.
Owl_rllib.A2c
Owl_rllib.Ppo
Actor and critic are separate networks with separate optimisers.
owl-rllib.bridges
owl-rllib.llm
module Make (Graph : Owl_neural_graph_sig.Sig) : sig ... end