gryf/coach
1
0
Fork 0
mirror of https://github.com/gryf/coach.git synced 2026-02-22 01:45:56 +01:00
Code Issues Projects Releases Wiki Activity

Move embedder, middleware, and head parameters to framework agnostic modules. (#45)

Browse Source 
Part of #28
This commit is contained in:
Sina Afrooze
2018-10-29 14:46:40 -07:00
committed by Scott Leishman
parent 16b3e99f37
commit a888226641
60 changed files with 410 additions and 330 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
rl_coach/agents/actor_critic_agent.py
View File

@@ -20,9 +20,9 @@ import numpy as np
import scipy.signal
from rl_coach.agents.policy_optimization_agent import PolicyOptimizationAgent, PolicyGradientRescaler
from rl_coach.architectures.tensorflow_components.heads.policy_head import PolicyHeadParameters
from rl_coach.architectures.tensorflow_components.heads.v_head import VHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import PolicyHeadParameters, VHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import AlgorithmParameters, NetworkParameters, \
AgentParameters
from rl_coach.exploration_policies.categorical import CategoricalParameters
@@ -31,7 +31,6 @@ from rl_coach.logger import screen
from rl_coach.memories.episodic.single_episode_buffer import SingleEpisodeBufferParameters
from rl_coach.spaces import DiscreteActionSpace, BoxActionSpace
from rl_coach.utils import last_sample
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
class ActorCriticAlgorithmParameters(AlgorithmParameters):

6
rl_coach/agents/bc_agent.py
View File

@@ -19,13 +19,13 @@ from typing import Union
import numpy as np
from rl_coach.agents.imitation_agent import ImitationAgent
from rl_coach.architectures.tensorflow_components.heads.policy_head import PolicyHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.head_parameters import PolicyHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.base_parameters import AgentParameters, AlgorithmParameters, NetworkParameters, \
MiddlewareScheme
from rl_coach.exploration_policies.e_greedy import EGreedyParameters
from rl_coach.memories.episodic.episodic_experience_replay import EpisodicExperienceReplayParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.memories.non_episodic.experience_replay import ExperienceReplayParameters

2
rl_coach/agents/categorical_dqn_agent.py
View File

@@ -20,7 +20,7 @@ import numpy as np
from rl_coach.agents.dqn_agent import DQNNetworkParameters, DQNAlgorithmParameters, DQNAgentParameters
from rl_coach.agents.value_optimization_agent import ValueOptimizationAgent
from rl_coach.architectures.tensorflow_components.heads.categorical_q_head import CategoricalQHeadParameters
from rl_coach.architectures.head_parameters import CategoricalQHeadParameters
from rl_coach.base_parameters import AgentParameters
from rl_coach.core_types import StateType
from rl_coach.exploration_policies.e_greedy import EGreedyParameters

6
rl_coach/agents/cil_agent.py
View File

@@ -17,9 +17,9 @@
from typing import Union
from rl_coach.agents.imitation_agent import ImitationAgent
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.architectures.tensorflow_components.heads.cil_head import RegressionHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import RegressionHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import AgentParameters, MiddlewareScheme, NetworkParameters, AlgorithmParameters
from rl_coach.exploration_policies.e_greedy import EGreedyParameters
from rl_coach.memories.non_episodic.balanced_experience_replay import BalancedExperienceReplayParameters

7
rl_coach/agents/clipped_ppo_agent.py
View File

@@ -23,12 +23,11 @@ import numpy as np
from rl_coach.agents.actor_critic_agent import ActorCriticAgent
from rl_coach.agents.policy_optimization_agent import PolicyGradientRescaler
from rl_coach.architectures.tensorflow_components.heads.ppo_head import PPOHeadParameters
from rl_coach.architectures.tensorflow_components.heads.v_head import VHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import PPOHeadParameters, VHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import AlgorithmParameters, NetworkParameters, \
AgentParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import EnvironmentSteps, Batch, EnvResponse, StateType
from rl_coach.exploration_policies.additive_noise import AdditiveNoiseParameters
from rl_coach.exploration_policies.categorical import CategoricalParameters

7
rl_coach/agents/ddpg_agent.py
View File

@@ -22,10 +22,9 @@ import numpy as np
from rl_coach.agents.actor_critic_agent import ActorCriticAgent
from rl_coach.agents.agent import Agent
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.architectures.tensorflow_components.heads.ddpg_actor_head import DDPGActorHeadParameters
from rl_coach.architectures.tensorflow_components.heads.v_head import VHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import DDPGActorHeadParameters, VHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import NetworkParameters, AlgorithmParameters, \
AgentParameters, EmbedderScheme
from rl_coach.core_types import ActionInfo, EnvironmentSteps

7
rl_coach/agents/dfp_agent.py
View File

@@ -21,14 +21,13 @@ from typing import Union
import numpy as np
from rl_coach.agents.agent import Agent
from rl_coach.architectures.head_parameters import MeasurementsPredictionHeadParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.architectures.tensorflow_components.layers import Conv2d, Dense
from rl_coach.architectures.tensorflow_components.heads.measurements_prediction_head import \
MeasurementsPredictionHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.base_parameters import AlgorithmParameters, AgentParameters, NetworkParameters, \
MiddlewareScheme
from rl_coach.core_types import ActionInfo, EnvironmentSteps, RunPhase
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.exploration_policies.e_greedy import EGreedyParameters
from rl_coach.memories.episodic.episodic_experience_replay import EpisodicExperienceReplayParameters
from rl_coach.memories.memory import MemoryGranularity

6
rl_coach/agents/dqn_agent.py
View File

@@ -19,11 +19,11 @@ from typing import Union
import numpy as np
from rl_coach.agents.value_optimization_agent import ValueOptimizationAgent
from rl_coach.architectures.tensorflow_components.heads.q_head import QHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import QHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import AlgorithmParameters, NetworkParameters, AgentParameters, \
MiddlewareScheme
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import EnvironmentSteps
from rl_coach.exploration_policies.e_greedy import EGreedyParameters
from rl_coach.memories.non_episodic.experience_replay import ExperienceReplayParameters

6
rl_coach/agents/human_agent.py
View File

@@ -23,11 +23,11 @@ from pandas import to_pickle
from rl_coach.agents.agent import Agent
from rl_coach.agents.bc_agent import BCNetworkParameters
from rl_coach.architectures.tensorflow_components.heads.policy_head import PolicyHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import PolicyHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import AlgorithmParameters, NetworkParameters, EmbedderScheme, \
AgentParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import ActionInfo
from rl_coach.exploration_policies.e_greedy import EGreedyParameters
from rl_coach.logger import screen

6
rl_coach/agents/n_step_q_agent.py
View File

@@ -20,10 +20,10 @@ import numpy as np
from rl_coach.agents.policy_optimization_agent import PolicyOptimizationAgent
from rl_coach.agents.value_optimization_agent import ValueOptimizationAgent
from rl_coach.architectures.tensorflow_components.heads.q_head import QHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import QHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import AlgorithmParameters, AgentParameters, NetworkParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import EnvironmentSteps
from rl_coach.exploration_policies.e_greedy import EGreedyParameters

6
rl_coach/agents/naf_agent.py
View File

@@ -19,11 +19,11 @@ from typing import Union
import numpy as np
from rl_coach.agents.value_optimization_agent import ValueOptimizationAgent
from rl_coach.architectures.tensorflow_components.heads.naf_head import NAFHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import NAFHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import AlgorithmParameters, AgentParameters, \
NetworkParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import ActionInfo, EnvironmentSteps
from rl_coach.exploration_policies.ou_process import OUProcessParameters

6
rl_coach/agents/nec_agent.py
View File

@@ -21,10 +21,10 @@ from typing import Union
import numpy as np
from rl_coach.agents.value_optimization_agent import ValueOptimizationAgent
from rl_coach.architectures.tensorflow_components.heads.dnd_q_head import DNDQHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import DNDQHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import AlgorithmParameters, NetworkParameters, AgentParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import RunPhase, EnvironmentSteps, Episode, StateType
from rl_coach.exploration_policies.e_greedy import EGreedyParameters

6
rl_coach/agents/policy_gradients_agent.py
View File

@@ -19,11 +19,11 @@ from typing import Union
import numpy as np
from rl_coach.agents.policy_optimization_agent import PolicyOptimizationAgent, PolicyGradientRescaler
from rl_coach.architectures.tensorflow_components.heads.policy_head import PolicyHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import PolicyHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import NetworkParameters, AlgorithmParameters, \
AgentParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.exploration_policies.additive_noise import AdditiveNoiseParameters
from rl_coach.exploration_policies.categorical import CategoricalParameters

7
rl_coach/agents/ppo_agent.py
View File

@@ -22,12 +22,11 @@ import numpy as np
from rl_coach.agents.actor_critic_agent import ActorCriticAgent
from rl_coach.agents.policy_optimization_agent import PolicyGradientRescaler
from rl_coach.architectures.tensorflow_components.heads.ppo_head import PPOHeadParameters
from rl_coach.architectures.tensorflow_components.heads.v_head import VHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.embedder_parameters import InputEmbedderParameters
from rl_coach.architectures.head_parameters import PPOHeadParameters, VHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import AlgorithmParameters, NetworkParameters, \
AgentParameters, DistributedTaskParameters
from rl_coach.architectures.tensorflow_components.embedders.embedder import InputEmbedderParameters
from rl_coach.core_types import EnvironmentSteps, Batch
from rl_coach.exploration_policies.additive_noise import AdditiveNoiseParameters

3
rl_coach/agents/qr_dqn_agent.py
View File

@@ -20,8 +20,7 @@ import numpy as np
from rl_coach.agents.dqn_agent import DQNAgentParameters, DQNNetworkParameters, DQNAlgorithmParameters
from rl_coach.agents.value_optimization_agent import ValueOptimizationAgent
from rl_coach.architectures.tensorflow_components.heads.quantile_regression_q_head import \
QuantileRegressionQHeadParameters
from rl_coach.architectures.head_parameters import QuantileRegressionQHeadParameters
from rl_coach.core_types import StateType
from rl_coach.schedules import LinearSchedule

4
rl_coach/agents/rainbow_dqn_agent.py
View File

@@ -21,8 +21,8 @@ import numpy as np
from rl_coach.agents.categorical_dqn_agent import CategoricalDQNAlgorithmParameters, \
CategoricalDQNAgent, CategoricalDQNAgentParameters
from rl_coach.agents.dqn_agent import DQNNetworkParameters
from rl_coach.architectures.tensorflow_components.heads.rainbow_q_head import RainbowQHeadParameters
from rl_coach.architectures.tensorflow_components.middlewares.fc_middleware import FCMiddlewareParameters
from rl_coach.architectures.head_parameters import RainbowQHeadParameters
from rl_coach.architectures.middleware_parameters import FCMiddlewareParameters
from rl_coach.base_parameters import MiddlewareScheme
from rl_coach.exploration_policies.parameter_noise import ParameterNoiseParameters
from rl_coach.memories.non_episodic.prioritized_experience_replay import PrioritizedExperienceReplayParameters, \