Python ObservationActionEncoder示例

示例#1

def default_model():
    return models.ActorTwinCriticWithTargets(
        actor=models.Actor(encoder=models.ObservationEncoder(),
                           torso=models.MLP((256, 256), torch.nn.ReLU),
                           head=models.DeterministicPolicyHead()),
        critic=models.Critic(encoder=models.ObservationActionEncoder(),
                             torso=models.MLP((256, 256), torch.nn.ReLU),
                             head=models.ValueHead()),
        observation_normalizer=normalizers.MeanStd())

示例#2

文件： d4pg.py 项目： lorenzosteccanella/tonic

def default_model():
    return models.ActorCriticWithTargets(
        actor=models.Actor(encoder=models.ObservationEncoder(),
                           torso=models.MLP((256, 256), torch.nn.ReLU),
                           head=models.DeterministicPolicyHead()),
        critic=models.Critic(
            encoder=models.ObservationActionEncoder(),
            torso=models.MLP((256, 256), torch.nn.ReLU),
            # These values are for the control suite with 0.99 discount.
            head=models.DistributionalValueHead(-150., 150., 51)),
        observation_normalizer=normalizers.MeanStd())

示例#3

def default_model():
    return models.ActorTwinCriticWithTargets(
        actor=models.Actor(
            encoder=models.ObservationEncoder(),
            torso=models.MLP((256, 256), torch.nn.ReLU),
            head=models.GaussianPolicyHead(
                loc_activation=torch.nn.Identity,
                distribution=models.SquashedMultivariateNormalDiag)),
        critic=models.Critic(encoder=models.ObservationActionEncoder(),
                             torso=models.MLP((256, 256), torch.nn.ReLU),
                             head=models.ValueHead()),
        observation_normalizer=normalizers.MeanStd())