gryf/coach
1
0
Fork 0
mirror of https://github.com/gryf/coach.git synced 2026-02-16 05:55:46 +01:00
Code Issues Projects Releases Wiki Activity

parameter noise exploration - using Noisy Nets

Browse Source
This commit is contained in:
Gal Leibovich
2018-08-27 18:19:01 +03:00
parent 658b437079
commit 1aa2ab0590
49 changed files with 536 additions and 433 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
rl_coach/presets/BitFlip_DQN.py
View File

@@ -1,8 +1,10 @@
from rl_coach.agents.dqn_agent import DQNAgentParameters
from rl_coach.architectures.tensorflow_components.architecture import Dense
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme, InputEmbedderParameters, \
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme, \
PresetValidationParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import TrainingSteps, EnvironmentEpisodes, EnvironmentSteps
from rl_coach.environments.gym_environment import Mujoco
from rl_coach.graph_managers.basic_rl_graph_manager import BasicRLGraphManager
from rl_coach.graph_managers.graph_manager import ScheduleParameters

3
rl_coach/presets/BitFlip_DQN_HER.py
View File

@@ -1,8 +1,9 @@
from rl_coach.agents.dqn_agent import DQNAgentParameters
from rl_coach.architectures.tensorflow_components.architecture import Dense
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme, InputEmbedderParameters, \
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme, \
PresetValidationParameters
from rl_coach.core_types import TrainingSteps, EnvironmentEpisodes, EnvironmentSteps
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.environments.gym_environment import Mujoco
from rl_coach.graph_managers.basic_rl_graph_manager import BasicRLGraphManager
from rl_coach.graph_managers.graph_manager import ScheduleParameters

2
rl_coach/presets/ControlSuite_DDPG.py
View File

@@ -1,6 +1,6 @@
from rl_coach.agents.ddpg_agent import DDPGAgentParameters
from rl_coach.architectures.tensorflow_components.architecture import Dense
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme, PresetValidationParameters
from rl_coach.core_types import TrainingSteps, EnvironmentEpisodes, EnvironmentSteps, RunPhase
from rl_coach.environments.control_suite_environment import ControlSuiteEnvironmentParameters, control_suite_envs
from rl_coach.environments.environment import MaxDumpMethod, SelectedPhaseOnlyDumpMethod, SingleLevelSelection

2
rl_coach/presets/Doom_Health_DFP.py
View File

@@ -1,7 +1,7 @@
from rl_coach.agents.dfp_agent import DFPAgentParameters
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme, MiddlewareScheme, \
PresetValidationParameters
from rl_coach.core_types import EnvironmentSteps, RunPhase
from rl_coach.core_types import EnvironmentSteps, RunPhase, EnvironmentEpisodes
from rl_coach.environments.doom_environment import DoomEnvironmentParameters
from rl_coach.environments.environment import SelectedPhaseOnlyDumpMethod, MaxDumpMethod
from rl_coach.graph_managers.basic_rl_graph_manager import BasicRLGraphManager

7
rl_coach/presets/Doom_Health_Supreme_DFP.py
View File

@@ -1,7 +1,6 @@
from rl_coach.agents.dfp_agent import DFPAgentParameters
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme, MiddlewareScheme, \
PresetValidationParameters
from rl_coach.core_types import EnvironmentSteps, RunPhase
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme, MiddlewareScheme
from rl_coach.core_types import EnvironmentSteps, RunPhase, EnvironmentEpisodes
from rl_coach.environments.doom_environment import DoomEnvironmentParameters
from rl_coach.environments.environment import SelectedPhaseOnlyDumpMethod, MaxDumpMethod
from rl_coach.graph_managers.basic_rl_graph_manager import BasicRLGraphManager
@@ -62,4 +61,4 @@ vis_params.dump_mp4 = False
graph_manager = BasicRLGraphManager(agent_params=agent_params, env_params=env_params,
schedule_params=schedule_params, vis_params=vis_params,
preset_validation_params=preset_validation_params)
)

6
rl_coach/presets/Fetch_DDPG_HER_baselines.py
View File

@@ -1,9 +1,9 @@
from rl_coach.agents.ddpg_agent import DDPGAgentParameters
from rl_coach.architectures.tensorflow_components.architecture import Dense
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme, InputEmbedderParameters, \
PresetValidationParameters
from rl_coach.core_types import TrainingSteps, EnvironmentEpisodes, EnvironmentSteps, RunPhase
from rl_coach.base_parameters import VisualizationParameters, EmbedderScheme, PresetValidationParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import EnvironmentEpisodes, EnvironmentSteps, TrainingSteps, RunPhase
from rl_coach.environments.environment import SelectedPhaseOnlyDumpMethod, MaxDumpMethod, SingleLevelSelection
from rl_coach.environments.gym_environment import Mujoco, MujocoInputFilter, fetch_v1
from rl_coach.exploration_policies.e_greedy import EGreedyParameters

4
rl_coach/presets/Mujoco_A3C_LSTM.py
View File

@@ -1,8 +1,8 @@
from rl_coach.agents.actor_critic_agent import ActorCriticAgentParameters
from rl_coach.architectures.tensorflow_components.architecture import Dense
from rl_coach.architectures.tensorflow_components.middlewares.lstm_middleware import LSTMMiddlewareParameters
from rl_coach.base_parameters import VisualizationParameters, InputEmbedderParameters, MiddlewareScheme, \
PresetValidationParameters
from rl_coach.base_parameters import VisualizationParameters, MiddlewareScheme, PresetValidationParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import TrainingSteps, EnvironmentEpisodes, EnvironmentSteps, RunPhase
from rl_coach.environments.environment import MaxDumpMethod, SelectedPhaseOnlyDumpMethod, SingleLevelSelection
from rl_coach.environments.gym_environment import Mujoco, mujoco_v2, MujocoInputFilter

5
rl_coach/presets/Pendulum_HAC.py
View File

@@ -2,8 +2,9 @@ import numpy as np
from rl_coach.agents.hac_ddpg_agent import HACDDPGAgentParameters
from rl_coach.architectures.tensorflow_components.architecture import Dense
from rl_coach.base_parameters import VisualizationParameters, EmbeddingMergerType, EmbedderScheme, \
InputEmbedderParameters
from rl_coach.base_parameters import VisualizationParameters, EmbeddingMergerType, EmbedderScheme
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import EnvironmentEpisodes, EnvironmentSteps, RunPhase, TrainingSteps
from rl_coach.environments.environment import SelectedPhaseOnlyDumpMethod
from rl_coach.environments.gym_environment import Mujoco

3
rl_coach/presets/Starcraft_CollectMinerals_A3C.py
View File

@@ -1,8 +1,9 @@
from rl_coach.agents.actor_critic_agent import ActorCriticAgentParameters
from rl_coach.agents.policy_optimization_agent import PolicyGradientRescaler
from rl_coach.base_parameters import VisualizationParameters, InputEmbedderParameters
from rl_coach.base_parameters import VisualizationParameters
from rl_coach.core_types import RunPhase
from rl_coach.core_types import TrainingSteps, EnvironmentEpisodes, EnvironmentSteps
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.environments.environment import MaxDumpMethod, SelectedPhaseOnlyDumpMethod
from rl_coach.environments.starcraft2_environment import StarCraft2EnvironmentParameters
from rl_coach.exploration_policies.additive_noise import AdditiveNoiseParameters

3
rl_coach/presets/Starcraft_CollectMinerals_Dueling_DDQN.py
View File

@@ -2,8 +2,9 @@ from collections import OrderedDict
from rl_coach.agents.ddqn_agent import DDQNAgentParameters
from rl_coach.architectures.tensorflow_components.heads.dueling_q_head import DuelingQHeadParameters
from rl_coach.base_parameters import VisualizationParameters, InputEmbedderParameters
from rl_coach.base_parameters import VisualizationParameters
from rl_coach.core_types import RunPhase
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import TrainingSteps, EnvironmentEpisodes, EnvironmentSteps
from rl_coach.environments.environment import MaxDumpMethod, SelectedPhaseOnlyDumpMethod
from rl_coach.environments.starcraft2_environment import StarCraft2EnvironmentParameters