from my_dueling_dqn import DuelingDQN
import numpy as np
import matplotlib.pyplot as plt
import tensorflow as tf
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
from matplotlib.figure import Figure
import tkinter as tk
root = tk.Tk()
root.title("matplotlib in TK")
f = Figure(figsize=(6, 6), dpi=100)
canvas = FigureCanvasTkAgg(f, master=root)
canvas.get_tk_widget().pack(side=tk.TOP, fill=tk.BOTH, expand=1)
env = freq_env()
MEMORY_SIZE = 5000
ACTION_SPACE = 10
sess = tf.Session()
with tf.variable_scope('natural'):
natural_DQN = DuelingDQN(n_actions=ACTION_SPACE,
n_features=160,
memory_size=MEMORY_SIZE,
e_greedy_increment=0.001,
sess=sess,
dueling=False)
with tf.variable_scope('dueling'):
dueling_DQN = DuelingDQN(n_actions=ACTION_SPACE,
n_features=160,
def main():
#env = gym.make(ENV_NAME)
env = freq_env()
agent = Agent(num_actions=env.actions)
root = tk.Tk()
root.title("matplotlib in TK")
f = Figure(figsize=(6, 6), dpi=100)
canvas = FigureCanvasTkAgg(f, master=root)
canvas.get_tk_widget().pack(side=tk.TOP, fill=tk.BOTH, expand=1)
waterfall_figure = f.add_subplot(111)
if TRAIN: # Train mode
if env.mode == 0:
for _ in range(NUM_EPISODES):
terminal = False
observation = env.waterfall_reset()
last_observation = observation
state = agent.get_initial_state(last_observation)
while not terminal:
last_observation = observation
action = agent.get_action(state)
observation, reward, terminal = env.step(
action, last_observation)
#waterfall_figure.imshow(observation[0, :, :, 0])
#canvas.draw()
processed_observation = preprocess(observation)
#print(observation==last_observation)
state = agent.run(state, action, reward, terminal,
processed_observation)
elif env.mode == 1:
for _ in range(NUM_EPISODES):
terminal = False
observation = env.waterfall_reset_combo()
last_observation = observation
state = agent.get_initial_state(last_observation)
while not terminal:
last_observation = observation
action = agent.get_action(state)
observation, reward, terminal = env.step_combo(
action, last_observation)
#waterfall_figure.imshow(observation[0, :, :, 0])
#canvas.draw()
processed_observation = preprocess(observation)
#print(observation==last_observation)
state = agent.run(state, action, reward, terminal,
processed_observation)
elif env.mode == 2:
for _ in range(NUM_EPISODES):
terminal = False
observation = env.waterfall_reset_dynamic()
last_observation = observation
state = agent.get_initial_state(last_observation)
while not terminal:
last_observation = observation
action = agent.get_action(state)
observation, reward, terminal = env.step_dynamic(
action, last_observation)
#waterfall_figure.imshow(observation[0, :, :, 0])
#canvas.draw()
processed_observation = preprocess(observation)
#print(observation==last_observation)
state = agent.run(state, action, reward, terminal,
processed_observation)
elif env.mode == 3:
for _ in range(NUM_EPISODES):
terminal = False
observation = env.waterfall_reset_intelligent()
last_observation = observation
state = agent.get_initial_state(last_observation)
while not terminal:
last_observation = observation
action = agent.get_action(state)
observation, reward, terminal = env.step_intelligent(
action, last_observation)
#waterfall_figure.imshow(observation[0, :, :, 0])
#canvas.draw()
processed_observation = preprocess(observation)
#print(observation==last_observation)
state = agent.run(state, action, reward, terminal,
processed_observation)
else: # Test mode
if env.mode == 0:
for _ in range(NUM_EPISODES_AT_TEST):
terminal = False
observation = env.waterfall_reset()
"""
for _ in range(random.randint(1, NO_OP_STEPS)):
last_observation = observation
observation, _, _, _ = env.step(0) # Do nothing
"""
last_observation = observation
state = agent.get_initial_state(last_observation)
while not terminal:
last_observation = observation
action = agent.get_action_at_test(state)
observation, reward, terminal = env.step(
action, last_observation)
waterfall_figure.imshow(observation[0, :, :, 0])
canvas.draw()
processed_observation = preprocess(observation)
state = processed_observation
elif env.mode == 1:
for _ in range(NUM_EPISODES_AT_TEST):
terminal = False
observation = env.waterfall_reset_combo()
"""
for _ in range(random.randint(1, NO_OP_STEPS)):
last_observation = observation
observation, _, _, _ = env.step(0) # Do nothing
"""
last_observation = observation
state = agent.get_initial_state(last_observation)
while not terminal:
last_observation = observation
action = agent.get_action_at_test(state)
observation, reward, terminal = env.step_combo(
action, last_observation)
waterfall_figure.imshow(observation[0, :, :, 0])
canvas.draw()
processed_observation = preprocess(observation)
state = processed_observation
elif env.mode == 2:
for _ in range(NUM_EPISODES_AT_TEST):
terminal = False
observation = env.waterfall_reset_dynamic()
"""
for _ in range(random.randint(1, NO_OP_STEPS)):
last_observation = observation
observation, _, _, _ = env.step(0) # Do nothing
"""
last_observation = observation
state = agent.get_initial_state(last_observation)
while not terminal:
last_observation = observation
action = agent.get_action_at_test(state)
observation, reward, terminal = env.step_dynamic(
action, last_observation)
waterfall_figure.imshow(observation[0, :, :, 0])
canvas.draw()
processed_observation = preprocess(observation)
state = processed_observation
elif env.mode == 3:
for _ in range(NUM_EPISODES_AT_TEST):
terminal = False
observation = env.waterfall_reset_intelligent()
"""
for _ in range(random.randint(1, NO_OP_STEPS)):
last_observation = observation
observation, _, _, _ = env.step(0) # Do nothing
"""
last_observation = observation
state = agent.get_initial_state(last_observation)
while not terminal:
last_observation = observation
action = agent.get_action_at_test(state)
observation, reward, terminal = env.step_intelligent(
action, last_observation)
waterfall_figure.imshow(observation[0, :, :, 0])
canvas.draw()
processed_observation = preprocess(observation)
state = processed_observation