Index

A

action (TimeStep attribute) action_space (EnvSpec attribute) (ParallelVecEnvExecutor attribute), [1] (Policy attribute), [1], [2], [3], [4], [5] (VecEnvExecutor attribute), [1] actions (MAMLTrajectoryBatch attribute) (TrajectoryBatch attribute) adapt_policy() (MAML method) (MetaRLAlgorithm method), [1] add_history() (ISSampler method), [1] add_path() (HERReplayBuffer method), [1] (PathBuffer method), [1]

add_transition() (ReplayBuffer method), [1] add_transitions() (ReplayBuffer method), [1] AddGaussianNoise (class in metarl.np.exploration_policies) (class in metarl.np.exploration_policies.add_gaussian_noise) AddOrnsteinUhlenbeckNoise (class in metarl.np.exploration_policies) (class in metarl.np.exploration_policies.add_ornstein_uhlenbeck_noise) agent (DefaultWorker attribute), [1] (TFWorkerWrapper attribute), [1] agent_info (TimeStep attribute) agent_infos (TrajectoryBatch attribute) augmented_observation_space (TaskEmbeddingPolicy attribute), [1]

B

Baseline (class in metarl.np.baselines) (class in metarl.np.baselines.baseline) baselines (MAMLTrajectoryBatch attribute) BaseModel (class in metarl.tf.models) (class in metarl.tf.models.model) BaseSampler (class in metarl.sampler.sampler_deprecated) BatchDataset (class in metarl.np.optimizers) (class in metarl.np.optimizers.minibatch_dataset) BatchSampler (class in metarl.sampler) (class in metarl.sampler.batch_sampler) (class in metarl.tf.samplers) (class in metarl.tf.samplers.batch_sampler) Bernoulli (class in metarl.tf.distributions) (class in metarl.tf.distributions.bernoulli)

BernoulliMLPRegressor (class in metarl.tf.regressors) (class in metarl.tf.regressors.bernoulli_mlp_regressor) build() (BaseModel method), [1] (CategoricalCNNPolicy method), [1] (CategoricalGRUPolicy method), [1] (CategoricalLSTMPolicy method), [1] (CategoricalMLPPolicy method), [1] (GaussianGRUPolicy method), [1] (GaussianLSTMPolicy method), [1] (GaussianMLPEncoder method), [1] (GaussianMLPPolicy method), [1] (GaussianMLPTaskEmbeddingPolicy method), [1] (Model method), [1] (StochasticEncoder method), [1] build_eval() (HessianVectorProduct method)

C

Categorical (class in metarl.tf.distributions) (class in metarl.tf.distributions.categorical) CategoricalCNNModel (class in metarl.tf.models) (class in metarl.tf.models.categorical_cnn_model) CategoricalCNNPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.categorical_cnn_policy) CategoricalGRUModel (class in metarl.tf.models) (class in metarl.tf.models.categorical_gru_model) CategoricalGRUPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.categorical_gru_policy) CategoricalLSTMModel (class in metarl.tf.models) (class in metarl.tf.models.categorical_lstm_model) CategoricalLSTMPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.categorical_lstm_policy) CategoricalMLPModel (class in metarl.tf.models) (class in metarl.tf.models.categorical_mlp_model) CategoricalMLPPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.categorical_mlp_policy) CategoricalMLPRegressor (class in metarl.tf.regressors) (class in metarl.tf.regressors.categorical_mlp_regressor) cdf() (TanhNormal method), [1] CEM (class in metarl.np.algos) (class in metarl.np.algos.cem) center_advs() (in module metarl.tf.misc.tensor_utils) cg() (in module metarl.tf.optimizers.conjugate_gradient_optimizer) child_proc_shutdown() (in module metarl.experiment.experiment_wrapper) clear() (PathBuffer method), [1] clone() (CategoricalCNNPolicy method), [1] (CategoricalGRUPolicy method), [1] (CategoricalLSTMPolicy method), [1] (CategoricalMLPPolicy method), [1] (ContinuousCNNQFunction method), [1] (ContinuousMLPPolicy method), [1] (ContinuousMLPQFunction method), [1] (DiscreteCNNQFunction method), [1] (DiscreteMLPQFunction method), [1] (Encoder method), [1] (GaussianGRUPolicy method), [1] (GaussianLSTMPolicy method), [1] (GaussianMLPEncoder method), [1] (GaussianMLPPolicy method), [1] (GaussianMLPRegressorModel method) (GaussianMLPTaskEmbeddingPolicy method), [1] (QFunction method), [1] close() (in module metarl.sampler.parallel_sampler) (ParallelVecEnvExecutor method), [1] (Plotter method), [1], [2], [3] (StatefulPool method) (VecEnvExecutor method), [1]

CMAES (class in metarl.np.algos) (class in metarl.np.algos.cma_es) cnn() (in module metarl.tf.models.cnn) cnn_with_max_pooling() (in module metarl.tf.models.cnn) CNNMLPMergeModel (class in metarl.tf.models) (class in metarl.tf.models.cnn_mlp_merge_model) CNNModel (class in metarl.tf.models) (class in metarl.tf.models.cnn_model) CNNModelWithMaxPooling (class in metarl.tf.models) (class in metarl.tf.models.cnn_model_max_pooling) collect_rollout() (DefaultWorker method), [1] (TFWorkerWrapper method), [1] (VecWorker method), [1] (Worker method), [1] compile_function() (in module metarl.tf.misc.tensor_utils) compute_advantages() (in module metarl.tf.misc.tensor_utils) (in module metarl.torch) compute_kl_div() (ContextConditionedPolicy method), [1] compute_loss() (GaussianMLPValueFunction method), [1] (ValueFunction method), [1] concat_tensor_dict_list() (in module metarl.misc.tensor_utils) (in module metarl.tf.misc.tensor_utils) concat_tensor_list() (in module metarl.tf.misc.tensor_utils) concatenate() (metarl.TrajectoryBatch class method) ConjugateGradientOptimizer (class in metarl.tf.optimizers) (class in metarl.tf.optimizers.conjugate_gradient_optimizer) (class in metarl.torch.optimizers) (class in metarl.torch.optimizers.conjugate_gradient_optimizer) constraint_val() (ConjugateGradientOptimizer method), [1] (PenaltyLbfgsOptimizer method), [1] ConstructEnvsSampler (class in metarl.experiment.task_sampler) context (ContextConditionedPolicy attribute), [1] ContextConditionedPolicy (class in metarl.torch.policies) (class in metarl.torch.policies.context_conditioned_policy) ContinuousCNNQFunction (class in metarl.tf.q_functions) (class in metarl.tf.q_functions.continuous_cnn_q_function) ContinuousMLPBaseline (class in metarl.tf.baselines) (class in metarl.tf.baselines.continuous_mlp_baseline) ContinuousMLPPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.continuous_mlp_policy) ContinuousMLPQFunction (class in metarl.tf.q_functions) (class in metarl.tf.q_functions.continuous_mlp_q_function) (class in metarl.torch.q_functions) (class in metarl.torch.q_functions.continuous_mlp_q_function) ContinuousMLPRegressor (class in metarl.tf.regressors) (class in metarl.tf.regressors.continuous_mlp_regressor) cross_entropy_sym() (Categorical method), [1]

D

DDPG (class in metarl.tf.algos.ddpg) (class in metarl.torch.algos.ddpg) default() (LogEncoder method) DEFAULT_N_ENVS (VecWorker attribute), [1] DefaultWorker (class in metarl.sampler) (class in metarl.sampler.default_worker) DeterministicMLPPolicy (class in metarl.torch.policies) (class in metarl.torch.policies.deterministic_mlp_policy) DiagonalGaussian (class in metarl.tf.distributions) (class in metarl.tf.distributions.diagonal_gaussian) dict_np_to_torch() (in module metarl.torch) DifferentiableSGD (class in metarl.torch.optimizers) (class in metarl.torch.optimizers.differentiable_sgd) dim (Bernoulli attribute), [1] (Categorical attribute), [1] (DiagonalGaussian attribute), [1] (Distribution attribute), [1] (RecurrentCategorical attribute), [1] disable() (Plotter static method), [1] discount_cumsum() (in module metarl.misc.tensor_utils) discounted_returns() (in module metarl.tf.misc.tensor_utils) DiscreteCNNQFunction (class in metarl.tf.q_functions) (class in metarl.tf.q_functions.discrete_cnn_q_function) DiscreteMLPQFunction (class in metarl.tf.q_functions) (class in metarl.tf.q_functions.discrete_mlp_q_function) DiscreteQfDerivedPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.discrete_qf_derived_policy) dist_info() (StochasticPolicy method), [1]

dist_info_keys (Bernoulli attribute), [1] (Distribution attribute), [1] dist_info_specs (Categorical attribute), [1] (DiagonalGaussian attribute), [1] (Distribution attribute), [1] (RecurrentCategorical attribute), [1] dist_info_sym() (BernoulliMLPRegressor method), [1] (CategoricalMLPRegressor method), [1] (GaussianCNNRegressor method), [1] distribution (BernoulliMLPRegressor attribute), [1] (CategoricalCNNPolicy attribute), [1] (CategoricalGRUPolicy attribute), [1] (CategoricalLSTMPolicy attribute), [1] (CategoricalMLPPolicy attribute), [1] (CategoricalMLPRegressor attribute), [1] Distribution (class in metarl.tf.distributions) (class in metarl.tf.distributions.distribution) distribution (GaussianCNNRegressor attribute), [1] (GaussianGRUPolicy attribute), [1] (GaussianLSTMPolicy attribute), [1] (GaussianMLPEncoder attribute), [1] (GaussianMLPPolicy attribute), [1] (GaussianMLPRegressor attribute), [1] (GaussianMLPTaskEmbeddingPolicy attribute), [1] (StochasticEncoder attribute), [1] (StochasticModule attribute), [1] (StochasticPolicy attribute), [1] DQN (class in metarl.tf.algos.dqn) dump_json() (in module metarl.experiment.experiment)

E

enable (Plotter attribute), [1] Encoder (class in metarl.np.embeddings) (class in metarl.np.embeddings.encoder) (class in metarl.tf.embeddings) (class in metarl.tf.embeddings.encoder) encoder (TaskEmbeddingPolicy attribute), [1] encoder_distribution (TaskEmbeddingPolicy attribute), [1] entropy() (Bernoulli method), [1] (Categorical method), [1] (DiagonalGaussian method), [1] (Distribution method), [1] (GaussianMLPPolicy method), [1] (RecurrentCategorical method), [1] (TanhGaussianMLPPolicy method), [1] (TanhNormal method), [1] entropy_sym() (Categorical method), [1] (DiagonalGaussian method), [1] (Distribution method), [1] (RecurrentCategorical method), [1] enumerate_support() (TanhNormal method), [1] env (DefaultWorker attribute), [1] (TFWorkerWrapper attribute), [1]

env_info (TimeStep attribute) env_infos (TrajectoryBatch attribute) env_spec (Policy attribute), [1] (TimeStep attribute) (TrajectoryBatch attribute) EnvPoolSampler (class in metarl.experiment.task_sampler) EnvSpec (class in metarl.envs.env_spec) EnvUpdate (class in metarl.sampler.env_update) EpsilonGreedyPolicy (class in metarl.np.exploration_policies) (class in metarl.np.exploration_policies.epsilon_greedy_policy) ERWR (class in metarl.tf.algos.erwr) evaluate() (MetaEvaluator method), [1] ExistingEnvUpdate (class in metarl.sampler.env_update) expand() (TanhNormal method), [1] ExperimentContext (class in metarl.experiment.experiment) ExperimentStats (class in metarl.experiment.local_runner) ExperimentTemplate (class in metarl.experiment.experiment) explained_variance_1d() (in module metarl.misc.tensor_utils) ExplorationPolicy (class in metarl.np.exploration_policies) (class in metarl.np.exploration_policies.exploration_policy)

F

filter_valids() (in module metarl.tf.misc.tensor_utils) (in module metarl.torch) filter_valids_dict() (in module metarl.tf.misc.tensor_utils) FiniteDifferenceHvp (class in metarl.tf.optimizers) (class in metarl.tf.optimizers.conjugate_gradient_optimizer) FirstOrderOptimizer (class in metarl.tf.optimizers) (class in metarl.tf.optimizers.first_order_optimizer) fit() (Baseline method), [1] (BernoulliMLPRegressor method), [1] (CategoricalMLPRegressor method), [1] (ContinuousMLPBaseline method), [1] (ContinuousMLPRegressor method), [1] (GaussianCNNBaseline method), [1] (GaussianCNNRegressor method), [1] (GaussianMLPBaseline method), [1] (GaussianMLPRegressor method), [1] (LinearFeatureBaseline method), [1] (Regressor method), [1] (ZeroBaseline method), [1] FixedPolicy (class in metarl.np.policies) (class in metarl.np.policies.fixed_policy) flat_to_params() (Module method), [1]

flatten_batch() (in module metarl.tf.misc.tensor_utils) (in module metarl.torch) flatten_batch_dict() (in module metarl.tf.misc.tensor_utils) flatten_inputs() (in module metarl.tf.misc.tensor_utils) flatten_tensor_variables() (in module metarl.tf.misc.tensor_utils) flatten_tensors() (in module metarl.misc.tensor_utils) forward() (ContextConditionedPolicy method), [1] (ContinuousMLPQFunction method), [1] (GaussianMLPBaseModule method) (GaussianMLPValueFunction method), [1] (MLPModule method), [1] (MultiHeadedMLPModule method), [1] from_onehot() (in module metarl.tf.distributions.categorical) from_trajectory_list() (metarl.TrajectoryBatch class method) from_worker_factory() (metarl.sampler.local_sampler.LocalSampler class method) (metarl.sampler.LocalSampler class method) (metarl.sampler.MultiprocessingSampler class method) (metarl.sampler.RaySampler class method) (metarl.sampler.Sampler class method) (metarl.sampler.multiprocessing_sampler.MultiprocessingSampler class method) (metarl.sampler.ray_sampler.RaySampler class method) (metarl.sampler.sampler.Sampler class method) full (ReplayBuffer attribute), [1]

G

GaussianCNNBaseline (class in metarl.tf.baselines) (class in metarl.tf.baselines.gaussian_cnn_baseline) GaussianCNNModel (class in metarl.tf.models) (class in metarl.tf.models.gaussian_cnn_model) GaussianCNNRegressor (class in metarl.tf.regressors) (class in metarl.tf.regressors.gaussian_cnn_regressor) GaussianCNNRegressorModel (class in metarl.tf.regressors) (class in metarl.tf.regressors.gaussian_cnn_regressor_model) GaussianGRUModel (class in metarl.tf.models) (class in metarl.tf.models.gaussian_gru_model) GaussianGRUPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.gaussian_gru_policy) GaussianLSTMModel (class in metarl.tf.models) (class in metarl.tf.models.gaussian_lstm_model) GaussianLSTMPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.gaussian_lstm_policy) GaussianMLPBaseline (class in metarl.tf.baselines) (class in metarl.tf.baselines.gaussian_mlp_baseline) GaussianMLPBaseModule (class in metarl.torch.modules.gaussian_mlp_module) GaussianMLPEncoder (class in metarl.tf.embeddings) (class in metarl.tf.embeddings.gaussian_mlp_encoder) GaussianMLPIndependentStdModule (class in metarl.torch.modules) (class in metarl.torch.modules.gaussian_mlp_module) GaussianMLPModel (class in metarl.tf.models) (class in metarl.tf.models.gaussian_mlp_model) GaussianMLPModule (class in metarl.torch.modules) (class in metarl.torch.modules.gaussian_mlp_module) GaussianMLPPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.gaussian_mlp_policy) (class in metarl.torch.policies) (class in metarl.torch.policies.gaussian_mlp_policy) GaussianMLPRegressor (class in metarl.tf.regressors) (class in metarl.tf.regressors.gaussian_mlp_regressor) GaussianMLPRegressorModel (class in metarl.tf.regressors.gaussian_mlp_regressor_model) GaussianMLPTaskEmbeddingPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.gaussian_mlp_task_embedding_policy) GaussianMLPTwoHeadedModule (class in metarl.torch.modules) (class in metarl.torch.modules.gaussian_mlp_module) GaussianMLPValueFunction (class in metarl.torch.value_functions) (class in metarl.torch.value_functions.gaussian_mlp_value_function) get() (LazyDict method) get_action() (AddGaussianNoise method), [1] (AddOrnsteinUhlenbeckNoise method), [1] (CategoricalCNNPolicy method), [1] (CategoricalGRUPolicy method), [1] (CategoricalLSTMPolicy method), [1] (CategoricalMLPPolicy method), [1] (ContextConditionedPolicy method), [1] (ContinuousMLPPolicy method), [1] (DeterministicMLPPolicy method), [1] (DiscreteQfDerivedPolicy method), [1] (EpsilonGreedyPolicy method), [1] (ExplorationPolicy method), [1] (FixedPolicy method), [1] (GaussianGRUPolicy method), [1] (GaussianLSTMPolicy method), [1] (GaussianMLPPolicy method), [1], [2], [3] (GaussianMLPTaskEmbeddingPolicy method), [1] (Policy method), [1], [2], [3], [4], [5] (ScriptedPolicy method), [1] (TanhGaussianMLPPolicy method), [1] (TaskEmbeddingPolicy method), [1] (UniformControlPolicy method) get_action_given_latent() (GaussianMLPTaskEmbeddingPolicy method), [1] (TaskEmbeddingPolicy method), [1] get_action_given_task() (GaussianMLPTaskEmbeddingPolicy method), [1] (TaskEmbeddingPolicy method), [1] get_action_sym() (ContinuousMLPPolicy method), [1] get_actions() (AddGaussianNoise method), [1] (AddOrnsteinUhlenbeckNoise method), [1] (CategoricalCNNPolicy method), [1] (CategoricalGRUPolicy method), [1] (CategoricalLSTMPolicy method), [1] (CategoricalMLPPolicy method), [1] (ContinuousMLPPolicy method), [1] (DeterministicMLPPolicy method), [1] (DiscreteQfDerivedPolicy method), [1] (EpsilonGreedyPolicy method), [1] (ExplorationPolicy method), [1] (FixedPolicy method), [1] (GaussianGRUPolicy method), [1] (GaussianLSTMPolicy method), [1] (GaussianMLPPolicy method), [1], [2], [3] (GaussianMLPTaskEmbeddingPolicy method), [1] (Policy method), [1], [2], [3] (ScriptedPolicy method), [1] (TanhGaussianMLPPolicy method), [1] (TaskEmbeddingPolicy method), [1] (UniformControlPolicy method)

get_actions_given_latents() (GaussianMLPTaskEmbeddingPolicy method), [1] (TaskEmbeddingPolicy method), [1] get_actions_given_tasks() (GaussianMLPTaskEmbeddingPolicy method), [1] (TaskEmbeddingPolicy method), [1] get_env_copy() (LocalRunner method), [1] get_exploration_policy() (MAML method) (MetaRLAlgorithm method), [1] get_global_vars() (Module method), [1] (QFunction method), [1] (TaskEmbeddingPolicy method), [1] get_history_list() (ISSampler method), [1] get_latent() (Encoder method), [1], [2], [3] (GaussianMLPEncoder method), [1] (TaskEmbeddingPolicy method), [1] get_latents() (Encoder method), [1], [2], [3] (GaussianMLPEncoder method), [1] get_metadata() (in module metarl.experiment.experiment) get_minibatch() (OptimizerWrapper method), [1] get_param_shapes() (Module method), [1] get_param_values() (Baseline method), [1] (ContinuousMLPBaseline method), [1] (ExplorationPolicy method), [1] (FixedPolicy method), [1] (GaussianCNNBaseline method), [1] (GaussianMLPBaseline method), [1] (LinearFeatureBaseline method), [1] (Module method), [1] (Policy method), [1] (ScriptedPolicy method), [1] (ZeroBaseline method), [1] get_params() (Module method), [1] get_params_internal() (ContinuousMLPBaseline method), [1] (GaussianCNNBaseline method), [1] (GaussianMLPBaseline method), [1] get_plotters() (Plotter static method), [1], [2], [3] get_qval() (ContinuousCNNQFunction method), [1] (ContinuousMLPQFunction method), [1] get_qval_sym() (ContinuousCNNQFunction method), [1] (ContinuousMLPQFunction method), [1] (DiscreteCNNQFunction method), [1] (DiscreteMLPQFunction method), [1] (QFunction method), [1] get_regularizable_vars() (CategoricalMLPPolicy method), [1] (ContinuousMLPPolicy method), [1] (QFunction method), [1] get_seed() (in module metarl.experiment.deterministic) get_target_ops() (in module metarl.tf.misc.tensor_utils) get_trainable_vars() (Module method), [1] (QFunction method), [1] (TaskEmbeddingPolicy method), [1] global_device() (in module metarl.torch) graph_inputs() (in module metarl.tf.misc.tensor_utils) grow_pool() (EnvPoolSampler method) gru() (in module metarl.tf.models.gru) GRUModel (class in metarl.tf.models) (class in metarl.tf.models.gru_model)

H

HERReplayBuffer (class in metarl.replay_buffer) (class in metarl.replay_buffer.her_replay_buffer)

HessianVectorProduct (class in metarl.tf.optimizers.conjugate_gradient_optimizer) history (ISSampler attribute), [1]

I

icdf() (TanhNormal method), [1] identity_function() (in module metarl.sampler.worker_factory) infer_posterior() (ContextConditionedPolicy method), [1] init_opt() (DDPG method) (DQN method) (NOP method), [1] (NPO method) (OffPolicyRLAlgorithm method), [1] (REPS method) init_plot() (Plotter method), [1] init_worker() (Plotter method), [1] initialize() (in module metarl.sampler.parallel_sampler) (StatefulPool method) initialize_tf_vars() (LocalTFRunner method), [1] InOutSpec (class in metarl) input (DiscreteCNNQFunction attribute), [1] (DiscreteMLPQFunction attribute), [1] (GaussianMLPEncoder attribute), [1] (Model attribute), [1] (Network attribute) (Sequential attribute), [1]

input_dim (Encoder attribute), [1] (GaussianMLPEncoder attribute), [1] (MLPEncoder attribute), [1] input_space (InOutSpec attribute) inputs (ContinuousCNNQFunction attribute), [1] (ContinuousMLPQFunction attribute), [1] (Model attribute), [1] (Network attribute) (Sequential attribute), [1] ISSampler (class in metarl.sampler) (class in metarl.sampler.is_sampler) iterate() (BatchDataset method), [1]

K

kl() (Bernoulli method), [1] (Categorical method), [1] (DiagonalGaussian method), [1] (Distribution method), [1] (RecurrentCategorical method), [1]

kl_sym() (Bernoulli method), [1] (Categorical method), [1] (DiagonalGaussian method), [1] (Distribution method), [1] (RecurrentCategorical method), [1]

L

last_observations (TrajectoryBatch attribute) latent_mean (GaussianMLPEncoder attribute), [1] latent_space (TaskEmbeddingPolicy attribute), [1] latent_std_param (GaussianMLPEncoder attribute), [1] LazyDict (class in metarl.tf.optimizers.utils) LbfgsOptimizer (class in metarl.tf.optimizers) (class in metarl.tf.optimizers.lbfgs_optimizer) lengths (TrajectoryBatch attribute) likelihood_ratio_sym() (Bernoulli method), [1] (Categorical method), [1] (DiagonalGaussian method), [1] (Distribution method), [1] (RecurrentCategorical method), [1] LinearFeatureBaseline (class in metarl.np.baselines) (class in metarl.np.baselines.linear_feature_baseline) LinearMultiFeatureBaseline (class in metarl.np.baselines) (class in metarl.np.baselines.linear_multi_feature_baseline) load() (Snapshotter method), [1] LocalRunner (class in metarl.experiment) (class in metarl.experiment.local_runner) LocalSampler (class in metarl.sampler) (class in metarl.sampler.local_sampler) LocalTFRunner (class in metarl.experiment) (class in metarl.experiment.local_tf_runner) log_diagnostics() (Baseline method), [1] (LocalRunner method), [1] (NPO method) (OffPolicyRLAlgorithm method), [1] (Policy method), [1], [2], [3] (QFunction method), [1] (REPS method)

log_likelihood() (Bernoulli method), [1] (Categorical method), [1] (DiagonalGaussian method), [1] (Distribution method), [1] (GaussianMLPPolicy method), [1] (RecurrentCategorical method), [1] (TanhGaussianMLPPolicy method), [1] log_likelihood_sym() (Bernoulli method), [1] (BernoulliMLPRegressor method), [1] (Categorical method), [1] (CategoricalMLPRegressor method), [1] (DiagonalGaussian method), [1] (Distribution method), [1] (GaussianCNNRegressor method), [1] (RecurrentCategorical method), [1] (StochasticRegressor method), [1] log_multitask_performance() (in module metarl) log_parameters() (in module metarl.experiment.experiment_wrapper) log_performance() (in module metarl) (MAML method) log_prob() (TanhNormal method), [1] LogEncoder (class in metarl.experiment.experiment) loss() (ConjugateGradientOptimizer method), [1] (FirstOrderOptimizer method), [1] (LbfgsOptimizer method), [1] (PenaltyLbfgsOptimizer method), [1] lstm() (in module metarl.tf.models.lstm) LSTMModel (class in metarl.tf.models) (class in metarl.tf.models.lstm_model)

M

make_launcher_archive() (in module metarl.experiment.experiment) make_optimizer() (in module metarl) make_sampler() (LocalRunner method), [1] (LocalTFRunner method), [1] MAML (class in metarl.torch.algos.maml) MAMLPPO (class in metarl.torch.algos.maml_ppo) MAMLTrajectoryBatch (class in metarl.torch.algos.maml) MAMLTRPO (class in metarl.torch.algos.maml_trpo) MAMLVPG (class in metarl.torch.algos.maml_vpg) mean (TanhNormal attribute), [1] MetaEvaluator (class in metarl.experiment) (class in metarl.experiment.meta_evaluator) metarl (module) metarl.envs.env_spec (module) metarl.experiment (module) metarl.experiment.deterministic (module) metarl.experiment.experiment (module) metarl.experiment.experiment_wrapper (module) metarl.experiment.local_runner (module) metarl.experiment.local_tf_runner (module) metarl.experiment.meta_evaluator (module) metarl.experiment.snapshotter (module) metarl.experiment.task_sampler (module) metarl.misc (module) metarl.misc.tensor_utils (module) metarl.np (module) metarl.np.algos (module) metarl.np.algos.cem (module) metarl.np.algos.cma_es (module) metarl.np.algos.meta_rl_algorithm (module) metarl.np.algos.nop (module) metarl.np.algos.off_policy_rl_algorithm (module) metarl.np.algos.rl_algorithm (module) metarl.np.baselines (module) metarl.np.baselines.baseline (module) metarl.np.baselines.linear_feature_baseline (module) metarl.np.baselines.linear_multi_feature_baseline (module) metarl.np.baselines.zero_baseline (module) metarl.np.embeddings (module) metarl.np.embeddings.encoder (module) metarl.np.exploration_policies (module) metarl.np.exploration_policies.add_gaussian_noise (module) metarl.np.exploration_policies.add_ornstein_uhlenbeck_noise (module) metarl.np.exploration_policies.epsilon_greedy_policy (module) metarl.np.exploration_policies.exploration_policy (module) metarl.np.optimizers (module) metarl.np.optimizers.minibatch_dataset (module) metarl.np.policies (module) metarl.np.policies.fixed_policy (module) metarl.np.policies.policy (module) metarl.np.policies.scripted_policy (module) metarl.np.q_functions (module) metarl.np.q_functions.q_function (module) metarl.plotter (module) metarl.plotter.plotter (module) metarl.replay_buffer (module) metarl.replay_buffer.her_replay_buffer (module) metarl.replay_buffer.path_buffer (module) metarl.replay_buffer.replay_buffer (module) metarl.sampler (module) metarl.sampler.batch_sampler (module) metarl.sampler.default_worker (module) metarl.sampler.env_update (module) metarl.sampler.is_sampler (module) metarl.sampler.local_sampler (module) metarl.sampler.multiprocessing_sampler (module) metarl.sampler.off_policy_vectorized_sampler (module) metarl.sampler.on_policy_vectorized_sampler (module) metarl.sampler.parallel_sampler (module) metarl.sampler.parallel_vec_env_executor (module) metarl.sampler.ray_sampler (module) metarl.sampler.sampler (module) metarl.sampler.sampler_deprecated (module) metarl.sampler.stateful_pool (module) metarl.sampler.utils (module) metarl.sampler.vec_env_executor (module) metarl.sampler.vec_worker (module) metarl.sampler.worker (module) metarl.sampler.worker_factory (module) metarl.tf (module) metarl.tf.algos.ddpg (module) metarl.tf.algos.dqn (module) metarl.tf.algos.erwr (module) metarl.tf.algos.npo (module) metarl.tf.algos.ppo (module) metarl.tf.algos.reps (module) metarl.tf.algos.vpg (module) metarl.tf.baselines (module) metarl.tf.baselines.continuous_mlp_baseline (module) metarl.tf.baselines.gaussian_cnn_baseline (module) metarl.tf.baselines.gaussian_mlp_baseline (module) metarl.tf.distributions (module) metarl.tf.distributions.bernoulli (module) metarl.tf.distributions.categorical (module) metarl.tf.distributions.diagonal_gaussian (module) metarl.tf.distributions.distribution (module) metarl.tf.distributions.recurrent_categorical (module) metarl.tf.distributions.recurrent_diagonal_gaussian (module) metarl.tf.embeddings (module) metarl.tf.embeddings.encoder (module) metarl.tf.embeddings.gaussian_mlp_encoder (module) metarl.tf.misc (module) metarl.tf.misc.tensor_utils (module) metarl.tf.models (module) metarl.tf.models.categorical_cnn_model (module) metarl.tf.models.categorical_gru_model (module) metarl.tf.models.categorical_lstm_model (module) metarl.tf.models.categorical_mlp_model (module) metarl.tf.models.cnn (module) metarl.tf.models.cnn_mlp_merge_model (module) metarl.tf.models.cnn_model (module)

metarl.tf.models.cnn_model_max_pooling (module) metarl.tf.models.gaussian_cnn_model (module) metarl.tf.models.gaussian_gru_model (module) metarl.tf.models.gaussian_lstm_model (module) metarl.tf.models.gaussian_mlp_model (module) metarl.tf.models.gru (module) metarl.tf.models.gru_model (module) metarl.tf.models.lstm (module) metarl.tf.models.lstm_model (module) metarl.tf.models.mlp (module) metarl.tf.models.mlp_dueling_model (module) metarl.tf.models.mlp_merge_model (module) metarl.tf.models.mlp_model (module) metarl.tf.models.model (module) metarl.tf.models.module (module) metarl.tf.models.normalized_input_mlp_model (module) metarl.tf.models.parameter (module) metarl.tf.models.sequential (module) metarl.tf.optimizers (module) metarl.tf.optimizers.conjugate_gradient_optimizer (module) metarl.tf.optimizers.first_order_optimizer (module) metarl.tf.optimizers.lbfgs_optimizer (module) metarl.tf.optimizers.penalty_lbfgs_optimizer (module) metarl.tf.optimizers.utils (module) metarl.tf.plotter (module) metarl.tf.plotter.plotter (module) metarl.tf.policies (module) metarl.tf.policies.categorical_cnn_policy (module) metarl.tf.policies.categorical_gru_policy (module) metarl.tf.policies.categorical_lstm_policy (module) metarl.tf.policies.categorical_mlp_policy (module) metarl.tf.policies.continuous_mlp_policy (module) metarl.tf.policies.discrete_qf_derived_policy (module) metarl.tf.policies.gaussian_gru_policy (module) metarl.tf.policies.gaussian_lstm_policy (module) metarl.tf.policies.gaussian_mlp_policy (module) metarl.tf.policies.gaussian_mlp_task_embedding_policy (module) metarl.tf.policies.policy (module) metarl.tf.policies.task_embedding_policy (module) metarl.tf.policies.uniform_control_policy (module) metarl.tf.q_functions (module) metarl.tf.q_functions.continuous_cnn_q_function (module) metarl.tf.q_functions.continuous_mlp_q_function (module) metarl.tf.q_functions.discrete_cnn_q_function (module) metarl.tf.q_functions.discrete_mlp_q_function (module) metarl.tf.q_functions.q_function (module) metarl.tf.regressors (module) metarl.tf.regressors.bernoulli_mlp_regressor (module) metarl.tf.regressors.categorical_mlp_regressor (module) metarl.tf.regressors.continuous_mlp_regressor (module) metarl.tf.regressors.gaussian_cnn_regressor (module) metarl.tf.regressors.gaussian_cnn_regressor_model (module) metarl.tf.regressors.gaussian_mlp_regressor (module) metarl.tf.regressors.gaussian_mlp_regressor_model (module) metarl.tf.regressors.regressor (module) metarl.tf.samplers (module) metarl.tf.samplers.batch_sampler (module) metarl.tf.samplers.worker (module) metarl.torch (module) metarl.torch.algos.ddpg (module) metarl.torch.algos.maml (module) metarl.torch.algos.maml_ppo (module) metarl.torch.algos.maml_trpo (module) metarl.torch.algos.maml_vpg (module) metarl.torch.algos.ppo (module) metarl.torch.algos.trpo (module) metarl.torch.algos.vpg (module) metarl.torch.distributions (module) metarl.torch.distributions.tanh_normal (module) metarl.torch.embeddings (module) metarl.torch.embeddings.mlp_encoder (module) metarl.torch.modules (module) metarl.torch.modules.gaussian_mlp_module (module) metarl.torch.modules.mlp_module (module) metarl.torch.modules.multi_headed_mlp_module (module) metarl.torch.optimizers (module) metarl.torch.optimizers.conjugate_gradient_optimizer (module) metarl.torch.optimizers.differentiable_sgd (module) metarl.torch.optimizers.optimizer_wrapper (module) metarl.torch.policies (module) metarl.torch.policies.context_conditioned_policy (module) metarl.torch.policies.deterministic_mlp_policy (module) metarl.torch.policies.gaussian_mlp_policy (module) metarl.torch.policies.policy (module) metarl.torch.policies.tanh_gaussian_mlp_policy (module) metarl.torch.q_functions (module) metarl.torch.q_functions.continuous_mlp_q_function (module) metarl.torch.value_functions (module) metarl.torch.value_functions.gaussian_mlp_value_function (module) metarl.torch.value_functions.value_function (module) MetaRLAlgorithm (class in metarl.np.algos) (class in metarl.np.algos.meta_rl_algorithm) mlp() (in module metarl.tf.models.mlp) MLPDuelingModel (class in metarl.tf.models) (class in metarl.tf.models.mlp_dueling_model) MLPEncoder (class in metarl.torch.embeddings) (class in metarl.torch.embeddings.mlp_encoder) MLPMergeModel (class in metarl.tf.models) (class in metarl.tf.models.mlp_merge_model) MLPModel (class in metarl.tf.models) (class in metarl.tf.models.mlp_model) MLPModule (class in metarl.torch.modules) (class in metarl.torch.modules.mlp_module) Model (class in metarl.tf.models) (class in metarl.tf.models.model) Module (class in metarl.tf.models) (class in metarl.tf.models.module) MultiHeadedMLPModule (class in metarl.torch.modules) (class in metarl.torch.modules.multi_headed_mlp_module) MultiprocessingSampler (class in metarl.sampler) (class in metarl.sampler.multiprocessing_sampler)

N

n_tasks (ConstructEnvsSampler attribute) (EnvPoolSampler attribute) (SetTaskSampler attribute) (TaskSampler attribute), [1], [2], [3] n_transitions_stored (PathBuffer attribute), [1] (ReplayBuffer attribute), [1] name (BaseModel attribute), [1] (Model attribute), [1] (Module attribute), [1] (Policy attribute), [1] Network (class in metarl.tf.models.model) network_input_spec() (CNNMLPMergeModel method), [1] (GRUModel method), [1] (GaussianGRUModel method), [1] (GaussianLSTMModel method), [1] (LSTMModel method), [1] (MLPMergeModel method), [1] (Model method), [1] network_output_spec() (CategoricalGRUModel method), [1] (CategoricalLSTMModel method), [1] (GRUModel method), [1] (GaussianCNNModel method), [1] (GaussianCNNRegressorModel method), [1] (GaussianGRUModel method), [1] (GaussianLSTMModel method), [1] (GaussianMLPModel method), [1] (GaussianMLPRegressorModel method) (LSTMModel method), [1] (Model method), [1] (NormalizedInputMLPModel method), [1]

networks (ContextConditionedPolicy attribute), [1] (Model attribute), [1] new_tensor() (in module metarl.tf.misc.tensor_utils) new_tensor_like() (in module metarl.tf.misc.tensor_utils) NewEnvUpdate (class in metarl.sampler.env_update) NOP (class in metarl.np.algos) (class in metarl.np.algos.nop) normalize_pixel_batch() (in module metarl.misc.tensor_utils) NormalizedInputMLPModel (class in metarl.tf.models) (class in metarl.tf.models.normalized_input_mlp_model) NotAFileError NotSetupError NPO (class in metarl.tf.algos.npo) num_envs (ParallelVecEnvExecutor attribute), [1] (VecEnvExecutor attribute), [1] number_batches (BatchDataset attribute), [1]

O

observation (TimeStep attribute) observation_space (EnvSpec attribute) (ParallelVecEnvExecutor attribute), [1] (Policy attribute), [1], [2], [3], [4], [5] (VecEnvExecutor attribute), [1] observations (MAMLTrajectoryBatch attribute) (TrajectoryBatch attribute) obtain_exact_trajectories() (LocalSampler method), [1] (MultiprocessingSampler method), [1] (RaySampler method), [1] obtain_samples() (BatchSampler method), [1], [2], [3] (ISSampler method), [1] (LocalRunner method), [1] (LocalSampler method), [1] (MultiprocessingSampler method), [1] (OffPolicyVectorizedSampler method), [1] (OnPolicyVectorizedSampler method), [1] (RaySampler method), [1] (Sampler method), [1], [2] OffPolicyRLAlgorithm (class in metarl.np.algos) (class in metarl.np.algos.off_policy_rl_algorithm) OffPolicyVectorizedSampler (class in metarl.sampler) (class in metarl.sampler.off_policy_vectorized_sampler)

OnPolicyVectorizedSampler (class in metarl.sampler) (class in metarl.sampler.on_policy_vectorized_sampler) optimize() (ConjugateGradientOptimizer method), [1] (FirstOrderOptimizer method), [1] (LbfgsOptimizer method), [1] (PenaltyLbfgsOptimizer method), [1] optimize_policy() (DDPG method), [1] (DQN method) (NOP method), [1] (NPO method) (REPS method) OptimizerWrapper (class in metarl.torch.optimizers) (class in metarl.torch.optimizers.optimizer_wrapper) output (Model attribute), [1] (Network attribute) (Sequential attribute), [1] output_dim (Encoder attribute), [1] (GaussianMLPEncoder attribute), [1] (MLPEncoder attribute), [1] (MLPModule attribute), [1] output_space (InOutSpec attribute) outputs (Model attribute), [1] (Network attribute) (Sequential attribute), [1]

P

pad_tensor() (in module metarl.misc.tensor_utils) (in module metarl.tf.misc.tensor_utils) pad_tensor_dict() (in module metarl.misc.tensor_utils) (in module metarl.tf.misc.tensor_utils) pad_tensor_n() (in module metarl.misc.tensor_utils) (in module metarl.tf.misc.tensor_utils) pad_to_last() (in module metarl.torch) ParallelVecEnvExecutor (class in metarl.sampler) (class in metarl.sampler.parallel_vec_env_executor) parameter() (in module metarl.tf.models.parameter) parameters (BaseModel attribute), [1] (Model attribute), [1] PathBuffer (class in metarl.replay_buffer) (class in metarl.replay_buffer.path_buffer) paths (MAMLTrajectoryBatch attribute) paths_to_tensors() (in module metarl.np) (in module metarl.tf) PearlmutterHvp (class in metarl.tf.optimizers.conjugate_gradient_optimizer) PenaltyLbfgsOptimizer (class in metarl.tf.optimizers) (class in metarl.tf.optimizers.penalty_lbfgs_optimizer) Plotter (class in metarl.plotter) (class in metarl.plotter.plotter) (class in metarl.tf.plotter) (class in metarl.tf.plotter.plotter) Policy (class in metarl.np.policies) (class in metarl.np.policies.policy) (class in metarl.tf.policies) (class in metarl.tf.policies.policy) (class in metarl.torch.policies) (class in metarl.torch.policies.policy)

policy (MAML attribute) populate_task() (in module metarl.sampler.parallel_sampler) positive_advs() (in module metarl.tf.misc.tensor_utils) PPO (class in metarl.tf.algos.ppo) (class in metarl.torch.algos.ppo) predict() (Baseline method), [1] (BernoulliMLPRegressor method), [1] (CategoricalMLPRegressor method), [1] (ContinuousMLPBaseline method), [1] (ContinuousMLPRegressor method), [1] (GaussianCNNBaseline method), [1] (GaussianCNNRegressor method), [1] (GaussianMLPBaseline method), [1] (GaussianMLPRegressor method), [1] (LinearFeatureBaseline method), [1] (Regressor method), [1] (ZeroBaseline method), [1] predict_log_likelihood() (BernoulliMLPRegressor method), [1] (CategoricalMLPRegressor method), [1] predict_n() (ZeroBaseline method), [1] predict_sym() (ContinuousMLPRegressor method), [1] prepare_worker_messages() (WorkerFactory method), [1] process_samples() (OffPolicyRLAlgorithm method), [1] (VPG method) product_of_gaussians() (in module metarl.torch)

Q

q_vals (DiscreteCNNQFunction attribute), [1] (DiscreteMLPQFunction attribute), [1] QFunction (class in metarl.np.q_functions) (class in metarl.np.q_functions.q_function) (class in metarl.tf.q_functions) (class in metarl.tf.q_functions.q_function)

R

RaySampler (class in metarl.sampler) (class in metarl.sampler.ray_sampler) recurrent (BernoulliMLPRegressor attribute), [1] (CategoricalMLPRegressor attribute), [1] (ContinuousMLPRegressor attribute), [1] (GaussianCNNRegressor attribute), [1] recurrent_parameter() (in module metarl.tf.models.parameter) RecurrentCategorical (class in metarl.tf.distributions) (class in metarl.tf.distributions.recurrent_categorical) RecurrentDiagonalGaussian (in module metarl.tf.distributions) Regressor (class in metarl.tf.regressors) (class in metarl.tf.regressors.regressor) ReplayBuffer (class in metarl.replay_buffer) (class in metarl.replay_buffer.replay_buffer) REPS (class in metarl.tf.algos.reps) reset() (AddGaussianNoise method), [1] (AddOrnsteinUhlenbeckNoise method), [1] (CategoricalGRUPolicy method), [1] (CategoricalLSTMPolicy method), [1] (DeterministicMLPPolicy method), [1] (Encoder method), [1] (ExplorationPolicy method), [1] (FixedPolicy method), [1] (GaussianGRUPolicy method), [1] (GaussianLSTMPolicy method), [1] (GaussianMLPPolicy method), [1] (MLPEncoder method), [1] (Module method), [1] (ParallelVecEnvExecutor method), [1] (Policy method), [1] (ScriptedPolicy method), [1] (TanhGaussianMLPPolicy method), [1] (VecEnvExecutor method), [1]

reset_belief() (ContextConditionedPolicy method), [1] restore() (LocalRunner method), [1] resume() (LocalRunner method), [1] reward (TimeStep attribute) rewards (MAMLTrajectoryBatch attribute) (TrajectoryBatch attribute) RLAlgorithm (class in metarl.np.algos) (class in metarl.np.algos.rl_algorithm) rollout() (DefaultWorker method), [1] (SamplerWorker method) (TFWorkerWrapper method), [1] (Worker method), [1] (in module metarl.sampler.utils) rrse() (in module metarl.misc.tensor_utils) rsample() (TanhNormal method), [1] rsample_with_pre_tanh_value() (TanhNormal method), [1] run_collect() (StatefulPool method) run_each() (StatefulPool method) run_experiment() (in module metarl.experiment) (in module metarl.experiment.experiment) (in module metarl.experiment.experiment_wrapper) run_imap_unordered() (StatefulPool method) run_map() (StatefulPool method) run_worker() (in module metarl.sampler.multiprocessing_sampler)

S

sample() (Bernoulli method), [1] (Categorical method), [1] (ConstructEnvsSampler method) (DiagonalGaussian method), [1] (EnvPoolSampler method) (ReplayBuffer method), [1] (SetTaskSampler method) (TanhNormal method), [1] (TaskSampler method), [1] sample_from_belief() (ContextConditionedPolicy method), [1] sample_path() (PathBuffer method), [1] sample_paths() (in module metarl.sampler.parallel_sampler) sample_predict() (BernoulliMLPRegressor method), [1] sample_sym() (Categorical method), [1] (DiagonalGaussian method), [1] sample_transitions() (PathBuffer method), [1] Sampler (class in metarl.sampler) (class in metarl.sampler.sampler) (class in metarl.sampler.sampler_deprecated) SamplerWorker (class in metarl.sampler.ray_sampler) save() (LocalRunner method), [1] save_itr_params() (Snapshotter method), [1] ScriptedPolicy (class in metarl.np.policies) (class in metarl.np.policies.scripted_policy) Sequential (class in metarl.tf.models) (class in metarl.tf.models.sequential) set() (LazyDict method) set_gpu_mode() (in module metarl.torch) set_param_values() (Baseline method), [1] (ContinuousMLPBaseline method), [1] (ExplorationPolicy method), [1] (FixedPolicy method), [1] (GaussianCNNBaseline method), [1] (GaussianMLPBaseline method), [1] (LinearFeatureBaseline method), [1] (Module method), [1] (Policy method), [1] (ScriptedPolicy method), [1] (ZeroBaseline method), [1] set_seed() (in module metarl.experiment.deterministic) (in module metarl.sampler.parallel_sampler) SetTaskSampler (class in metarl.experiment.task_sampler) SetTaskUpdate (class in metarl.sampler.env_update) setup() (LocalRunner method), [1] (LocalTFRunner method), [1] SetupArgs (class in metarl.experiment.local_runner) SharedGlobal (class in metarl.sampler.stateful_pool) shutdown() (DefaultWorker method), [1] (SamplerWorker method) (TFWorkerWrapper method), [1] (VecWorker method), [1] (Worker method), [1] shutdown_worker() (BatchSampler method), [1], [2], [3] (LocalSampler method), [1] (MultiprocessingSampler method), [1] (OffPolicyVectorizedSampler method), [1] (OnPolicyVectorizedSampler method), [1] (RaySampler method), [1] (Sampler method), [1], [2] slice_nested_dict() (in module metarl.misc.tensor_utils) sliced_fun() (in module metarl.tf.optimizers.utils) sliding_window() (in module metarl.misc.tensor_utils)

snapshot_dir (SnapshotConfig attribute), [1] (Snapshotter attribute), [1] snapshot_gap (SnapshotConfig attribute), [1] (Snapshotter attribute), [1] snapshot_mode (SnapshotConfig attribute), [1] (Snapshotter attribute), [1] SnapshotConfig (class in metarl.experiment) (class in metarl.experiment.snapshotter) Snapshotter (class in metarl.experiment) (class in metarl.experiment.snapshotter) spec (Encoder attribute), [1] (GaussianMLPEncoder attribute), [1] (MLPEncoder attribute), [1] split() (TrajectoryBatch method) split_augmented_observation() (TaskEmbeddingPolicy method), [1] split_tensor_dict_list() (in module metarl.misc.tensor_utils) (in module metarl.tf.misc.tensor_utils) stack_and_pad_tensor_dict_list() (in module metarl.misc.tensor_utils) stack_tensor_dict_list() (in module metarl.misc.tensor_utils) (in module metarl.tf.misc.tensor_utils) start() (Plotter method), [1] start_rollout() (DefaultWorker method), [1] (TFWorkerWrapper method), [1] (VecWorker method), [1] (Worker method), [1] start_worker() (BatchSampler method), [1], [2], [3] (OffPolicyVectorizedSampler method), [1] (OnPolicyVectorizedSampler method), [1] (RaySampler method), [1] (Sampler method), [1], [2] state (ConjugateGradientOptimizer attribute), [1] state_info_keys (Module attribute), [1] (Policy attribute), [1] state_info_specs (CategoricalGRUPolicy attribute), [1] (CategoricalLSTMPolicy attribute), [1] (GaussianGRUPolicy attribute), [1] (GaussianLSTMPolicy attribute), [1] (Module attribute), [1] StatefulPool (class in metarl.sampler.stateful_pool) step() (ConjugateGradientOptimizer method), [1] (DifferentiableSGD method), [1] (OptimizerWrapper method), [1] (ParallelVecEnvExecutor method), [1] (VecEnvExecutor method), [1] step_epochs() (LocalRunner method), [1] step_rollout() (DefaultWorker method), [1] (TFWorkerWrapper method), [1] (VecWorker method), [1] (Worker method), [1] StochasticEncoder (class in metarl.np.embeddings) (class in metarl.np.embeddings.encoder) (class in metarl.tf.embeddings) (class in metarl.tf.embeddings.encoder) StochasticModule (class in metarl.tf.models) (class in metarl.tf.models.module) StochasticPolicy (class in metarl.np.policies) (class in metarl.np.policies.policy) (class in metarl.tf.policies) (class in metarl.tf.policies.policy) StochasticRegressor (class in metarl.tf.regressors) (class in metarl.tf.regressors.regressor) store_episode() (ReplayBuffer method), [1]

T

TanhGaussianMLPPolicy (class in metarl.torch.policies) (class in metarl.torch.policies.tanh_gaussian_mlp_policy) TanhNormal (class in metarl.torch.distributions) (class in metarl.torch.distributions.tanh_normal) task_space (TaskEmbeddingPolicy attribute), [1] TaskEmbeddingPolicy (class in metarl.tf.policies) (class in metarl.tf.policies.task_embedding_policy) TaskSampler (class in metarl.experiment) (class in metarl.experiment.task_sampler) terminals (TimeStep attribute) (TrajectoryBatch attribute) terminate() (Module method), [1] (Policy method), [1] terminate_task() (in module metarl.sampler.parallel_sampler) TFWorkerClassWrapper (class in metarl.tf.samplers) (class in metarl.tf.samplers.worker) TFWorkerWrapper (class in metarl.tf.samplers) (class in metarl.tf.samplers.worker) TimeStep (class in metarl) to() (GaussianMLPBaseModule method) to_local_command() (in module metarl.experiment) (in module metarl.experiment.experiment) to_trajectory_list() (TrajectoryBatch method) torch_to_np() (in module metarl.torch)

total_env_steps (LocalRunner attribute), [1] train() (CEM method), [1] (CMAES method), [1] (LocalRunner method), [1] (MAML method) (NOP method), [1] (NPO method) (OffPolicyRLAlgorithm method), [1] (REPS method) (RLAlgorithm method), [1] (VPG method) train_once() (CEM method), [1] (CMAES method), [1] (DDPG method), [1] (DQN method) (MAML method) (NPO method) (OffPolicyRLAlgorithm method), [1] (REPS method) (VPG method) TrainArgs (class in metarl.experiment.local_runner) TrajectoryBatch (class in metarl) TRPO (class in metarl.torch.algos.trpo) truncate_paths() (in module metarl.sampler.utils) truncate_tensor_dict() (in module metarl.misc.tensor_utils)

U

unflatten_tensors() (in module metarl.misc.tensor_utils) UniformControlPolicy (class in metarl.tf.policies.uniform_control_policy) update() (BatchDataset method), [1] (SamplerWorker method) update_agent() (DefaultWorker method), [1] (TFWorkerWrapper method), [1] (VecWorker method), [1] (Worker method), [1] update_context() (ContextConditionedPolicy method), [1] update_env() (DefaultWorker method), [1] (TFWorkerWrapper method), [1] (VecWorker method), [1] (Worker method), [1]

update_hvp() (FiniteDifferenceHvp method), [1] (HessianVectorProduct method) (PearlmutterHvp method) update_module_params() (in module metarl.torch) update_opt() (ConjugateGradientOptimizer method), [1] (FirstOrderOptimizer method), [1] (LbfgsOptimizer method), [1] (PenaltyLbfgsOptimizer method), [1] update_plot() (Plotter method), [1], [2], [3] update_target() (DDPG method)

V

valids (MAMLTrajectoryBatch attribute) ValueFunction (class in metarl.torch.value_functions) (class in metarl.torch.value_functions.value_function) variance (TanhNormal attribute), [1] VecEnvExecutor (class in metarl.sampler) (class in metarl.sampler.vec_env_executor) vectorized (BernoulliMLPRegressor attribute), [1] (CategoricalCNNPolicy attribute), [1] (CategoricalGRUPolicy attribute), [1] (CategoricalLSTMPolicy attribute), [1] (CategoricalMLPPolicy attribute), [1] (CategoricalMLPRegressor attribute), [1] (ContinuousMLPPolicy attribute), [1] (ContinuousMLPRegressor attribute), [1] (DiscreteQfDerivedPolicy attribute), [1] (GaussianCNNRegressor attribute), [1] (GaussianGRUPolicy attribute), [1] (GaussianLSTMPolicy attribute), [1] (GaussianMLPEncoder attribute), [1] (GaussianMLPPolicy attribute), [1], [2], [3] (GaussianMLPRegressor attribute), [1] (Module attribute), [1] (Policy attribute), [1], [2], [3] (TanhGaussianMLPPolicy attribute), [1] (UniformControlPolicy attribute)

VecWorker (class in metarl.sampler) (class in metarl.sampler.vec_worker) VPG (class in metarl.tf.algos.vpg) (class in metarl.torch.algos.vpg)

W

Worker (class in metarl.sampler) (class in metarl.sampler.worker) worker_init() (DefaultWorker method), [1] (TFWorkerWrapper method), [1] worker_init_envs() (in module metarl.sampler.parallel_vec_env_executor) worker_init_tf() (in module metarl.tf.samplers.batch_sampler)

worker_init_tf_vars() (in module metarl.tf.samplers.batch_sampler) worker_run_reset() (in module metarl.sampler.parallel_vec_env_executor) worker_run_step() (in module metarl.sampler.parallel_vec_env_executor) WorkerFactory (class in metarl.sampler) (class in metarl.sampler.worker_factory) wrap_experiment() (in module metarl) (in module metarl.experiment.experiment)

Z

zero_grad() (DifferentiableSGD method), [1] (OptimizerWrapper method), [1]

ZeroBaseline (class in metarl.np.baselines) (class in metarl.np.baselines.zero_baseline)