import os
import sys
from os.path import dirname, abspath
sys.path.append(dirname(dirname(abspath(__file__))))
import gym
from agents.actor_critic_agents.SAC_Discrete import SAC_Discrete
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
from environments.DMP_Env_1D_dynamic import deep_mobile_printing_1d1r
config = Config()
config.seed = 1
config.environment = deep_mobile_printing_1d1r()
config.num_episodes_to_run = 10000
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = True
config.GPU = "cuda:0"
config.overwrite_existing_results_file = True
config.randomise_random_seed = False
config.save_model = False
OUT_FILE_NAME = "SAC_1d" + "sin" + "_seed_" + str(config.seed)
config.save_model_path = "/mnt/NAS/home/WenyuHan/SNAC/SAC/1D/dynamic/" + OUT_FILE_NAME + "/"
config.file_to_save_data_results = "/mnt/NAS/home/WenyuHan/SNAC/SAC/1D/dynamic/" + OUT_FILE_NAME + "/" + "Results_Data.pkl"
config.file_to_save_results_graph = "/mnt/NAS/home/WenyuHan/SNAC/SAC/1D/dynamic/" + OUT_FILE_NAME + "/" + "Results_Graph.png"
if os.path.exists(config.save_model_path) == False:
os.makedirs(config.save_model_path)
import sys
from os.path import dirname, abspath
sys.path.append(dirname(dirname(abspath(__file__))))
import gym
from agents.actor_critic_agents.A2C import A2C
from agents.actor_critic_agents.A3C import A3C
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
config = Config()
config.seed = 1
config.environment = gym.make("gym_boxworld:boxworldRandomSmall-v0")
config.num_episodes_to_run = int(1e3)
config.file_to_save_data_results = "results/data_and_graphs/Boxworld_Results_Data.pkl"
config.file_to_save_results_graph = "results/data_and_graphs/Boxworld_Results_Graph.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = True
config.hyperparameters = {
"Actor_Critic_Agents": {
from agents.DQN_agents.DDQN import DDQN
from agents.hierarchical_agents.DIAYN import DIAYN
from agents.hierarchical_agents.h_DQN import h_DQN
from agents.hierarchical_agents.HIRO import HIRO
from agents.hierarchical_agents.SNN_HRL import SNN_HRL
from agents.policy_gradient_agents.PPO import PPO
from agents.policy_gradient_agents.REINFORCE import REINFORCE
from environments.FaceDiscreete import FaceEnvironementDiscreete
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
config = Config()
config.seed = 1
config.environment = FaceEnvironementDiscreete(
"../weights/blg_small_12_5e-06_5e-05_2_8_small_big_noisy_first_True_512")
config.num_episodes_to_run = 500
config.file_to_save_data_results = "Data_and_Graphs/FaceDiscreete.pkl"
config.file_to_save_results_graph = "Data_and_Graphs/FaceDiscreete.png"
config.show_solution_score = True
config.visualise_individual_results = True
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = True
config.overwrite_existing_results_file = True
config.randomise_random_seed = True
config.save_model = True
actor_critic_agent_hyperparameters = {
from A3C import A3C
from agents.DQN_agents.DQN_HER import DQN_HER
from DDQN import DDQN
from environments.Four_Rooms_Environment import Four_Rooms_Environment
from hierarchical_agents.DIAYN import DIAYN
from hierarchical_agents.HRL import HRL
from hierarchical_agents.SNN_HRL import SNN_HRL
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
from agents.DQN_agents.DQN import DQN
config = Config()
config.seed = 1
config.environment = Four_Rooms_Environment(
15,
15,
stochastic_actions_probability=0.25,
random_start_user_place=True,
random_goal_place=False)
config.num_episodes_to_run = 200
config.file_to_save_data_results = "Data_and_Graphs/Four_Rooms.pkl"
config.file_to_save_results_graph = "Data_and_Graphs/Four_Rooms.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 3
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
from A2C import A2C
from Dueling_DDQN import Dueling_DDQN
from SAC_Discrete import SAC_Discrete
from agents.actor_critic_agents.A3C import A3C
from agents.policy_gradient_agents.PPO import PPO
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
from agents.DQN_agents.DDQN import DDQN
from agents.DQN_agents.DDQN_With_Prioritised_Experience_Replay import DDQN_With_Prioritised_Experience_Replay
from agents.DQN_agents.DQN import DQN
from agents.DQN_agents.DQN_With_Fixed_Q_Targets import DQN_With_Fixed_Q_Targets
config = Config()
config.seed = 1
config.environment = gym.make("CartPole-v0")
config.num_episodes_to_run = 450
config.file_to_save_data_results = "data_and_graphs/Cart_Pole_Results_Data.pkl"
config.file_to_save_results_graph = "data_and_graphs/Cart_Pole_Results_Graph.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
config.hyperparameters = {
sys.path.append(dirname(dirname(abspath(__file__))))
import gym
from agents.actor_critic_agents.SAC_Discrete import SAC_Discrete
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
from environments.DMP_Env_2D_dynamic import deep_mobile_printing_2d1r
config = Config()
config.seed = 1
# 0: sin
# 1: gaussian
# 2: step
config.environment = deep_mobile_printing_2d1r(plan_choose=0)
config.num_episodes_to_run = 5000
config.file_to_save_data_results = "results/data_and_graphs/Cart_Pole_Results_Data.pkl"
config.file_to_save_results_graph = "results/data_and_graphs/Cart_Pole_Results_Graph.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = True
config.overwrite_existing_results_file = True
config.randomise_random_seed = True
config.save_model = False
config.hyperparameters = {
from environments.j2n6s300.DDPG_HER_env_Gazebo import j2n6s300_Environment
from agents.actor_critic_agents.DDPG_HER import DDPG_HER
from utilities.data_structures.Config import Config
from agents.Trainer import Trainer
from datetime import datetime
import os
now = datetime.now() # current date and time
date_str = now.strftime("%Y-%m-%d_%H-%M-%S")
os.mkdir('Data_and_Graphs/results_' + date_str)
path = 'Data_and_Graphs/results_' + date_str + '/'
config = Config()
config.seed = 1
config.environment = j2n6s300_Environment(proxyID='Env1')
config.num_episodes_to_run = 1
config.file_to_save_config = path + "config.json"
config.file_to_save_data_results = path + "jaco_DDPG-HER.pkl"
config.file_to_save_results_graph = path + "jaco_DDPG-HER.png"
config.show_solution_score = False
config.visualise_results_while_training = True
config.visualise_individual_results = True
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = True
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.load_model = False
config.load_model_path = "Models/model.pt"
config.save_model = True
from agents.hierarchical_agents.SNN_HRL import SNN_HRL
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
from agents.DQN_agents.DQN import DQN
from agents.hierarchical_agents.h_DQN import h_DQN
from environments.Long_Corridor_Environment import Long_Corridor_Environment
config = Config()
config.seed = 1
config.env_parameters = {"stochasticity_of_action_right": 0.5}
config.environment = Long_Corridor_Environment(
stochasticity_of_action_right=config.
env_parameters["stochasticity_of_action_right"])
config.num_episodes_to_run = 10000
config.file_to_save_data_results = "Data_and_Graphs/Long_Corridor_Results_Data.pkl"
config.file_to_save_results_graph = "Data_and_Graphs/Long_Corridor_Results_Graph.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 3
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
config.load_model = False
config.hyperparameters = {
"h_DQN": {
"CONTROLLER": {
"batch_size":
"clip_rewards":False
}
}
num_vehicles = 30
num_episode = 10
trials = 100
action_type = ["random","greedy"]
task_num = 32
task_file = "../sac/tasks.txt"
# config.environment = VEC_Environment(num_vehicles=50, task_num=task_num)
# config.environment.generate_change_tasks("../change_tasks.txt", 8)
# with open(count_file,'w+') as f:
# f.write("")
config.environment = VEC_Environment(num_vehicles=num_vehicles, task_num=task_num)
config.environment.load_offloading_tasks(task_file, 3)
for iter in [5]:
count_file = "../../../learningrate_{}.txt".format(iter)
with open(count_file,'w+') as f:
f.write("")
config.environment.count_file = count_file
for learning_rate in [0.00002, 0.00008,0.0002,0.0008,0.002,0.008]:
print("num_vehicles=",num_vehicles)
config.hyperparameters["Actor_Critic_Agents"]["Actor"]["learning_rate"]=learning_rate
config.hyperparameters["Actor_Critic_Agents"]["Critic"]["learning_rate"]=learning_rate
# for i in action_type:
# print(i)
# with open("../finish_count.txt",'a') as f:
# f.write(i+'\n')
# results = []
import os
import sys
sys.path.append(os.path.dirname(os.path.abspath(os.path.dirname(__file__))))
import gym
from agents.actor_critic_agents.DDPG import DDPG
from agents.actor_critic_agents.DDPG_HER_Che import DDPG_HER_Che
from utilities.data_structures.Config import Config
from agents.Trainer import Trainer
config = Config()
config.seed = 1
config.environment = gym.make("FetchReach-v1")
#config.environment = gym.make("FetchPush-v1")
#config.environment = gym.make("FetchPickAndPlace-v1")
#config.environment = gym.make("FetchSlide-v1")
config.num_episodes_to_run = 2000
#config.num_episodes_to_run = 2000
config.file_to_save_data_results = None
config.file_to_save_results_graph = None
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
#config.runs_per_agent = 3
config.runs_per_agent = 25
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
from agents.actor_critic_agents.A2C import A2C
from agents.DQN_agents.Dueling_DDQN import Dueling_DDQN
from agents.actor_critic_agents.SAC_Discrete import SAC_Discrete
from agents.actor_critic_agents.A3C import A3C
from agents.policy_gradient_agents.PPO import PPO
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
from agents.DQN_agents.DDQN import DDQN
from agents.DQN_agents.DDQN_With_Prioritised_Experience_Replay import DDQN_With_Prioritised_Experience_Replay
from agents.DQN_agents.DQN import DQN
from agents.DQN_agents.DQN_With_Fixed_Q_Targets import DQN_With_Fixed_Q_Targets
config = Config()
config.seed = 1
config.environment = CartPoleEnv() #gym.make("CartPole-v0")
config.num_episodes_to_run = 450
config.file_to_save_data_results = "results/data_and_graphs/Cart_Pole_Results_Data.pkl"
config.file_to_save_results_graph = "results/data_and_graphs/Cart_Pole_Results_Graph.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
config.hyperparameters = {
"DQN_Agents": {
str_to_obj = {
'SAC': SAC,
'DDQN': DDQN,
'SAC_Discrete': SAC_Discrete,
'DIAYN': DIAYN,
'DBH': DBH
}
if args.rts:
config.rts()
AGENTS = [DDQN, SAC_Discrete, DIAYN, DBH]
else:
AGENTS = [str_to_obj[i] for i in args.algorithms]
config.environment_name = args.environment
config.environment = gym.make(config.environment_name)
config.eval = args.evaluate
config.seed = args.seed
config.num_episodes_to_run = args.num_episodes
config.runs_per_agent = args.n_trials
config.use_GPU = args.use_GPU
config.save_results = args.save_results
config.run_prefix = args.run_prefix
config.train_existing_model = args.tem
config.save_directory = 'results/{}'.format(config.run_prefix)
if not os.path.exists(config.save_directory):
os.makedirs(config.save_directory)
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
linear_hidden_units = [128, 128, 32]
"""
from agents.DQN_agents.DQN_HER import DQN_HER
from environments.j2n6s300.HER_env_tf import j2n6s300_Environment
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
from datetime import datetime
now = datetime.now() # current date and time
num_episodes_to_run = 500
eps_decay_rate_denom = round(num_episodes_to_run/6)
config = Config()
config.seed = 1
config.environment = j2n6s300_Environment()
config.num_episodes_to_run = num_episodes_to_run
config.file_to_save_data_results = "Data_and_Graphs/{}jaco.pkl".format(now.strftime("%Y-%m-%d_%H-%M-%S_"))
config.file_to_save_results_graph = "Data_and_Graphs/{}jaco.png".format(now.strftime("%Y-%m-%d_%H-%M-%S_"))
config.show_solution_score = False
config.visualise_results_while_training = True
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = True
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.load_model = False
config.load_model_path = "Models/.pt"
config.save_model = False
import gym
from environments.Atari_Environment import make_atari_game
from agents.DQN_agents.DDQN import DDQN
from agents.hierarchical_agents.HRL.HRL import HRL
from agents.hierarchical_agents.HRL.Model_HRL import Model_HRL
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
config = Config()
config.seed = 1
config.environment = make_atari_game("SpaceInvaders-v0")
config.env_parameters = {}
config.num_episodes_to_run = 500
config.file_to_save_data_results = "data_and_graphs/hrl_experiments/Space_Invaders_Data.pkl"
config.file_to_save_results_graph = "data_and_graphs/hrl_experiments/Space_Invaders.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 10
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
# Loss is not drawing a random sample! otherwise wouldnt jump around that much!!
linear_hidden_units = [32, 32]
learning_rate = 0.005 # 0.001 taxi
buffer_size = 1000000
batch_size = 256
import gym
import sys
sys.path.append('./')
import agents
from agents.policy_gradient_agents.PPO import PPO
from agents.actor_critic_agents.DDPG import DDPG
from agents.actor_critic_agents.SAC import SAC
from agents.actor_critic_agents.TD3 import TD3
from agents.Trainer import Trainer
from agents.hierarchical_agents.DIAYN import DIAYN
from utilities.data_structures.Config import Config
config = Config()
config.seed = 1
config.environment = gym.make("HalfCheetah-v2")
config.num_episodes_to_run = 1000
config.file_to_save_data_results = "data_and_graphs/Hopper_Results_Data.pkl"
config.file_to_save_results_graph = "data_and_graphs/Hopper_Results_Graph.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 3
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
actor_critic_agent_hyperparameters = {
"Actor": {
"learning_rate": 0.0003,
from utilities.data_structures.Config import Config
from agents.DQN_agents.DQN import DQN
config = Config()
config.seed = 1
height = 15
width = 15
random_goal_place = False
num_possible_states = (height * width)**(1 + 1 * random_goal_place)
embedding_dimensions = [[num_possible_states, 20]]
print("Num possible states ", num_possible_states)
config.environment = Four_Rooms_Environment(
height,
width,
stochastic_actions_probability=0.0,
random_start_user_place=True,
random_goal_place=random_goal_place)
config.num_episodes_to_run = 1000
config.file_to_save_data_results = "Data_and_Graphs/Four_Rooms.pkl"
config.file_to_save_results_graph = "Data_and_Graphs/Four_Rooms.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 3
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
from utilities.data_structures.Config import Config from agents.Trainer import Trainer from agents.DQN_agents.DQN import DQN from agents.DQN_agents.DDQN import DDQN from agents.DQN_agents.Dueling_DDQN import Dueling_DDQN from agents.DQN_agents.DDQN_With_Prioritised_Experience_Replay import DDQN_With_Prioritised_Experience_Replay from agents.DQN_agents.DRQN import DRQN from models.FCNN import FCNN from gym.core import Wrapper config = Config() config.environment = Wrapper(SimpleISC(mode="DISCRETE")) config.num_episodes_to_run = 50 config.file_to_save_data_results = "results/data_and_graphs/isc/IllinoisSolarCar_Results_Data.pkl" config.file_to_save_results_graph = "results/data_and_graphs/isc/IllinoisSolarCar_Results_Graph.png" config.show_solution_score = True config.visualise_individual_results = True config.visualise_overall_agent_results = True config.standard_deviation_results = 1.0 config.runs_per_agent = 1 config.use_GPU = False config.overwrite_existing_results_file = True config.randomise_random_seed = False config.save_model = False config.seed = 0 config.debug_mode = True
sys.path.append(
os.path.join(config.Constants['SUMO_PATH'], os.sep, 'tools'))
sumoBinary = config.Constants["SUMO_GUI_PATH"]
sumoCmd = [
sumoBinary, '-S', '-d', config.Constants['Simulation_Delay'], "-c",
config.Constants["SUMO_CONFIG"], "--no-warnings", "true"
]
traci.start(sumoCmd)
init_traci()
config.network = RoadNetworkModel(config.Constants["ROOT"],
config.Constants["Network_XML"])
config.utils = Utils(config)
config.environment = AR_SUMO_Environment(config)
config.network_state_size = config.environment.utils.get_network_state_size()
if config.routing_mode in ["Q_routing_1_hop", "Q_routing_2_hop"]:
assert (config.network_state_size == num_GAT_features_per_layer[0])
gat = GAT(
config=config,
num_of_layers=config.hyperparameters["GAT"]['num_of_layers'],
num_heads_per_layer=config.hyperparameters["GAT"]['num_heads_per_layer'],
num_features_per_layer=config.hyperparameters["GAT"]
['num_features_per_layer'],
add_skip_connection=config.hyperparameters["GAT"]['add_skip_connection'],
bias=config.hyperparameters["GAT"]['bias'],
dropout=config.hyperparameters["GAT"]['dropout'],
log_attention_weights=not config.training_mode).to(config.device)
import gym
import random
import numpy as np
import torch
from hierarchical_agents.HIRO import HIRO
from utilities.data_structures.Config import Config
random.seed(1)
np.random.seed(1)
torch.manual_seed(1)
config = Config()
config.seed = 1
config.environment = gym.make("Pendulum-v0")
config.num_episodes_to_run = 1500
config.file_to_save_data_results = None
config.file_to_save_results_graph = None
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
config.hyperparameters = {
"LOWER_LEVEL": {
import gym
from agents.hierarchical_agents.HRL.HRL import HRL
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
config = Config()
config.environment = gym.make("Taxi-v2")
config.seed = 1
config.env_parameters = {}
config.num_episodes_to_run = 2000
config.file_to_save_data_results = None
config.file_to_save_results_graph = None
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 3
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
linear_hidden_units = [32, 32]
learning_rate = 0.01
buffer_size = 100000
batch_size = 256
batch_norm = False
embedding_dimensionality = 10
gradient_clipping_norm = 5
update_every_n_steps = 1
import gym
from agents.policy_gradient_agents.PPO import PPO
from agents.actor_critic_agents.DDPG import DDPG
from agents.actor_critic_agents.SAC import SAC
from agents.actor_critic_agents.TD3 import TD3
from agents.Trainer import Trainer
from agents.hierarchical_agents.DIAYN import DIAYN
from utilities.data_structures.Config import Config
config = Config()
config.seed = 1
config.environment = gym.make("Walker2d-v2")
config.num_episodes_to_run = 400
config.file_to_save_data_results = "data_and_graphs/Walker_Results_Data.pkl"
config.file_to_save_results_graph = "data_and_graphs/Walker_Results_Graph.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 3
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
config.load_model = False
actor_critic_agent_hyperparameters = {
"Actor": {
"learning_rate": 0.0003,
"linear_hidden_units": [64, 64],
"final_layer_activation": None,
import os import sys from os.path import dirname, abspath sys.path.append(dirname(dirname(abspath(__file__)))) import gym from agents.actor_critic_agents.SAC_Discrete import SAC_Discrete from agents.Trainer import Trainer from utilities.data_structures.Config import Config from environments.DMP_simulator_3d_static_circle import deep_mobile_printing_3d1r PALN_CHOICE = 1 # 0 dense 1 sparse PLAN_LIST = ["dense", "sparse"] PLAN_NAME = PLAN_LIST[PALN_CHOICE] config = Config() config.seed = 5 config.environment = deep_mobile_printing_3d1r(plan_choose=PALN_CHOICE) config.num_episodes_to_run = 5000 config.show_solution_score = False config.visualise_individual_results = False config.visualise_overall_agent_results = True config.standard_deviation_results = 1.0 config.runs_per_agent = 1 config.use_GPU = True config.GPU = "cuda:1" config.overwrite_existing_results_file = True config.randomise_random_seed = False config.save_model = False OUT_FILE_NAME = "SAC_3d_" + PLAN_NAME + "_seed_" + str(config.seed) config.save_model_path = "/mnt/NAS/home/WenyuHan/SNAC/SAC/3D/static/" + OUT_FILE_NAME + "/" config.file_to_save_data_results = "/mnt/NAS/home/WenyuHan/SNAC/SAC/3D/static/" + OUT_FILE_NAME + "/" + "Results_Data.pkl" config.file_to_save_results_graph = "/mnt/NAS/home/WenyuHan/SNAC/SAC/3D/static/" + OUT_FILE_NAME + "/" + "Results_Graph.png" if os.path.exists(config.save_model_path) == False:
import gym
from agents.policy_gradient_agents.PPO import PPO
from agents.actor_critic_agents.DDPG import DDPG
from agents.actor_critic_agents.SAC import SAC
from agents.actor_critic_agents.TD3 import TD3
from agents.Trainer import Trainer
from agents.hierarchical_agents.DIAYN import DIAYN
from utilities.data_structures.Config import Config
config = Config()
config.seed = 1
config.environment = gym.make("Hopper-v2")
config.num_episodes_to_run = 1000
config.file_to_save_data_results = "data_and_graphs/Hopper_Results_Data.pkl"
config.file_to_save_results_graph = "data_and_graphs/Hopper_Results_Graph.png"
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 3
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
config.load_model = False
actor_critic_agent_hyperparameters = {
"Actor": {
"learning_rate": 0.0003,
if __name__ == '__main__':
from utilities.data_structures.Config import Config
import gym
## envs import ##
from environments.carla_enviroments import env_v1_ObstacleAvoidance
# net = q_network_toa(n_action=4)
# net.to('cuda')
# input = torch.rand(size=(10, 3, 224, 224)).to('cuda')
# q1, q2 = net(input)
config = Config()
config.seed = 1
config.environment = gym.make("ObstacleAvoidance-v0")
config.num_episodes_to_run = 2000
config.file_to_save_data_results = "C:/my_project/Deep-Reinforcement-Learning-Algorithms-with-PyTorch/results/data_and_graphs/carla_obstacle_avoidance/data.pkl"
config.file_to_save_results_graph = "C:/my_project/Deep-Reinforcement-Learning-Algorithms-with-PyTorch/results/data_and_graphs/carla_obstacle_avoidance/data.png"
config.show_solution_score = False
config.visualise_individual_results = True
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = True
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = True
config.resume = False
config.resume_path = ''
from utilities.data_structures.Config import Config
from agents.DQN_agents.DDQN import DDQN
from agents.DQN_agents.DDQN_With_Prioritised_Experience_Replay import DDQN_With_Prioritised_Experience_Replay
from agents.DQN_agents.DQN import DQN
from agents.DQN_agents.DQN_With_Fixed_Q_Targets import DQN_With_Fixed_Q_Targets
from agents.policy_gradient_agents.REINFORCE import REINFORCE
## envs import ##
from environments.carla_enviroments import env_v1_ObstacleAvoidance
env_title = "ObstacleAvoidance-v0"
config = Config()
config.env_title = env_title
config.seed = 1
config.environment = gym.make(env_title)
config.num_episodes_to_run = 2000
config.show_solution_score = False
config.visualise_individual_results = True
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = True
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = True
config.log_loss = False
config.log_base = time.strftime("%Y%m%d%H%M%S", time.localtime())
config.save_model_freq = 300 ## save model per 300 episodes
config.retrain = False
import gym
from agents.Trainer import Trainer
from agents.actor_critic_agents.DDPG import DDPG
from agents.hierarchical_agents.HIRO import HIRO
from utilities.data_structures.Config import Config
config = Config()
config.seed = 1
config.environment = gym.make(
"Reacher-v2") # Reacher-v2 "InvertedPendulum-v2") #Pendulum-v0
config.num_episodes_to_run = 1500
config.file_to_save_data_results = None
config.file_to_save_results_graph = None
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 1
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
config.load_model = False
config.hyperparameters = {
"HIRO": {
"LOWER_LEVEL": {
"max_lower_level_timesteps": 5,
"Actor": {
"learning_rate": 0.001,
"linear_hidden_units": [20, 20],
"final_layer_activation": "TANH",
import gym
from models.Trainer import Trainer
from models.actor_critic_agents.DDPG import DDPG
from models.actor_critic_agents.TD3 import TD3
from models.policy_gradient_agents.PPO import PPO
from utilities.data_structures.Config import Config
config = Config()
config.seed = 1
config.environment = gym.make("MountainCarContinuous-v0")
config.num_episodes_to_run = 450
config.file_to_save_data_results = None
config.file_to_save_results_graph = None
config.show_solution_score = False
config.visualise_individual_results = False
config.visualise_overall_agent_results = True
config.standard_deviation_results = 1.0
config.runs_per_agent = 3
config.use_GPU = False
config.overwrite_existing_results_file = False
config.randomise_random_seed = True
config.save_model = False
config.hyperparameters = {
"Policy_Gradient_Agents": {
"learning_rate": 0.05,
"linear_hidden_units": [30, 15],
"final_layer_activation": "TANH",
"learning_iterations_per_round": 10,
"discount_rate": 0.9,
# from environments.Four_Rooms_Environment import Four_Rooms_Environment
# from agents.hierarchical_agents.SNN_HRL import SNN_HRL
# from agents.actor_critic_agents.TD3 import TD3
from agents.Trainer import Trainer
from utilities.data_structures.Config import Config
from agents.DQN_agents.DQN import DQN
import numpy as np
import torch
random.seed(1)
np.random.seed(1)
torch.manual_seed(1)
config = Config()
config.seed = 1
config.environment = Bit_Flipping_Environment(4)
config.num_episodes_to_run = 2000
config.file_to_save_data_results = None
config.file_to_save_results_graph = None
config.visualise_individual_results = False
config.visualise_overall_agent_results = False
config.randomise_random_seed = False
config.runs_per_agent = 1
config.use_GPU = False
config.hyperparameters = {
"DQN_Agents": {
"learning_rate": 0.005,
"batch_size": 64,
"buffer_size": 40000,
"epsilon": 0.1,
"epsilon_decay_rate_denominator": 200,
# symbol="SH123038"
# symbol="SH113561"
# symbol="SH128096"
# symbol="SH110058"
# symbol="SZ128105"
# symbol="SH128019"
# symbol="SZ128112"
# symbol="SH123052"
# symbol="SH113553"
# symbol="SH123032"
# config.environment = gym.make('stocks-v0', frame_bound=(50, 100), window_size=10)
# gym_anytrading.register_new('sz.000001')
# gym_anytrading.register_new('sh.600959')
print(symbol)
gym_anytrading.register_new_kzz(symbol)
config.environment = gym.make('kzz-v1')
# config.environment.update_df()
# column_list = ['turn', 'pctChg']
# column_list = ['turn', 'pctChg']
column_list = ["test2"]
# column_list = ["turn", "pctChg", "peTTM", "psTTM", "pcfNcfTTM", "pbMRQ"]
column_list_str = "_".join(column_list)
# config.environment.update_df(fn=None, column_list=column_list)
config.environment.update_df(fn=None, column_list=None)
# config.environment.update_df(fn=lambda df:df.head(100), column_list=column_list)
# config.environment = gym.make("CartPole-v0")
config.num_episodes_to_run = 50
# config.num_episodes_to_run = 450
config.file_to_save_data_results = "results/data_and_graphs/stocks_Results_Data.pkl"
config.file_to_save_results_graph = "results/data_and_graphs/stocks_Results_Graph.png"
config.show_solution_score = False
