gryf/coach
1
0
Fork 0
mirror of https://github.com/gryf/coach.git synced 2026-04-10 15:13:40 +02:00
Code Issues Projects Releases Wiki Activity

Integrate coach.py params with distributed Coach. (#42)

Browse Source 
* Integrate coach.py params with distributed Coach.
* Minor improvements
- Use enums instead of constants.
- Reduce code duplication.
- Ask experiment name with timeout.
This commit is contained in:
Balaji Subramaniam
2018-11-05 09:33:30 -08:00
committed by GitHub
parent 95b4fc6888
commit 7e7006305a
13 changed files with 263 additions and 285 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
rl_coach/presets/CartPole_DQN.py
View File

@@ -1,5 +1,5 @@
from rl_coach.agents.dqn_agent import DQNAgentParameters
from rl_coach.base_parameters import VisualizationParameters, PresetValidationParameters
from rl_coach.base_parameters import VisualizationParameters, PresetValidationParameters, DistributedCoachSynchronizationType
from rl_coach.core_types import TrainingSteps, EnvironmentEpisodes, EnvironmentSteps
from rl_coach.environments.gym_environment import GymVectorEnvironment
from rl_coach.graph_managers.basic_rl_graph_manager import BasicRLGraphManager

5
rl_coach/presets/CartPole_PPO.py
View File

@@ -1,6 +1,6 @@
from rl_coach.agents.clipped_ppo_agent import ClippedPPOAgentParameters
from rl_coach.architectures.tensorflow_components.layers import Dense
from rl_coach.base_parameters import VisualizationParameters, PresetValidationParameters
from rl_coach.base_parameters import VisualizationParameters, PresetValidationParameters, DistributedCoachSynchronizationType
from rl_coach.core_types import TrainingSteps, EnvironmentEpisodes, EnvironmentSteps, RunPhase
from rl_coach.environments.gym_environment import GymVectorEnvironment, mujoco_v2
from rl_coach.exploration_policies.additive_noise import AdditiveNoiseParameters
@@ -44,6 +44,9 @@ agent_params.algorithm.optimization_epochs = 10
agent_params.algorithm.estimate_state_value_using_gae = True
agent_params.algorithm.num_steps_between_copying_online_weights_to_target = EnvironmentSteps(2048)
# Distributed Coach synchronization type.
agent_params.algorithm.distributed_coach_synchronization_type = DistributedCoachSynchronizationType.SYNC
agent_params.exploration = EGreedyParameters()
agent_params.exploration.epsilon_schedule = LinearSchedule(1.0, 0.01, 10000)
agent_params.pre_network_filter.add_observation_filter('observation', 'normalize_observation',