def fitness(solution, idx):
global ginst
if idx is None:
return 0
genaction = GenAction(idx)
env = SnakeEnv(need_render=False,
use_simple=True,
set_life=int(GRID_WIDTH_NUM / 3))
obs = env.reset()
while True:
action = genaction(obs)
_, obs, done, _ = env(action)
if done:
break
bl = len(env.status.snake_body)
if bl < 10:
fscore = (env.n_moves**2) * (2**bl)
else:
fscore = env.n_moves**2
fscore *= 1024
fscore *= (bl - 9)
if fscore > ginst.max_score:
ginst.max_score = fscore
ginst.max_idx = idx
ginst.params = env.n_moves, bl
print('find new best: ', fscore, env.n_moves, bl)
return fscore
def __init__(self,
n_solutions=2000,
n_input=24,
n_output=4,
hiddens=[18, 18],
n_gen=0):
self.n_solutions = n_solutions
self.n_input = n_input
self.n_output = n_output
self.hiddens = hiddens
self.gann = pygad.gann.GANN(num_solutions=n_solutions,
num_neurons_input=n_input,
num_neurons_output=n_output,
num_neurons_hidden_layers=hiddens,
hidden_activations=["relu", "relu"],
output_activation="softmax")
global ginst
ginst.gann = self.gann
ginst.gene = 0
ginst.n_gen = n_gen
if not hasattr(ginst, 'env') or ginst.env is None:
ginst.env = SnakeEnv(need_render=True, use_simple=True)
ginst.env.reset()
ginst.env.render()
input()
ginst.env.reset()
ginst.env.render()
self.init_generation(n_gen=n_gen)
self.n_gen = n_gen
def show_gen(idx):
global ginst
genaction = GenAction(idx)
env = SnakeEnv(use_simple=True)
obs = env.reset()
c = 0
while True:
c += 1
action = genaction(obs)
_, obs, done, _ = env(action)
extra = '代数: {}'.format(ginst.n_gen)
if c % 3 == 0:
if ginst.n_gen:
env.render(extra)
else:
env.render()
# sleep(1 / ginst.n_gen)
print(done, env.status.direction, env.life)
if done:
break
sleep(1.5)
def run(display, retrain, num_episodes):
pygame.init()
env = SnakeEnv()
agent = QlearningAgent(env)
if not retrain:
try:
load_policy(agent)
except:
pass
for _ in tqdm(range(num_episodes)):
state = env.reset()
done = False
while not done:
for event in pygame.event.get():
if event.type == pygame.QUIT:
save_policy(agent)
pygame.quit()
sys.exit()
if display:
env.render()
action = agent.act(state)
next_state, reward, done = env.step(action)
agent.update_q_value(state, reward, action, next_state, done)
state = next_state
agent.epsilon = max(agent.epsilon * agent.epsilon_decay_rate,
agent.min_epsilon)
save_policy(agent)
def show_gen(idx):
global ginst
genaction = GenAction(idx)
env = SnakeEnv(use_simple=True)
obs = env.reset()
while True:
action = genaction(obs)
_, obs, done, _ = env(action)
env.render()
if done:
break
env.close()
def test_easy():
sum_opt = 0
sum_0 = 0
sum_1 = 0
env = SnakeEnv(0,[3,6])
policy_ref = [1]*97+[0]*3
policy_0 = [0]*100
policy_1 = [1]*100
for i in range(1000):
sum_opt += eval_game(env,policy_ref)
sum_0 += eval_game(env,policy_0)
sum_1 += eval_game(env,policy_1)
print('opt %f'%(sum_opt/1000))
print('0 %f'%(sum_0/1000))
print('1 %f'%(sum_1/1000))
def run():
env = tf_py_environment.TFPyEnvironment(SnakeEnv(step_limit=1000))
## Needs to be the same network from training
q_net = q_network.QNetwork(
env.observation_spec(),
env.action_spec(),
conv_layer_params=(),
fc_layer_params=(256, 100),
)
optimizer = tf.compat.v1.train.AdamOptimizer(learning_rate=1e-3)
global_counter = tf.compat.v1.train.get_or_create_global_step()
agent = dqn_agent.DqnAgent(
env.time_step_spec(),
env.action_spec(),
q_network=q_net,
optimizer=optimizer,
td_errors_loss_fn=common.element_wise_squared_loss,
train_step_counter=global_counter,
gamma=0.95,
epsilon_greedy=0.1,
n_step_update=1,
)
agent.initialize()
policy_checkpointer = common.Checkpointer(
ckpt_dir=os.path.join(train_dir, 'policy'),
policy=agent.policy,
global_step=global_counter,
)
policy_checkpointer.initialize_or_restore()
capture_run(
os.path.join(root_dir, "snake" + str(global_counter.numpy()) + ".mp4"),
env, agent.policy)
def run_snake():
brain = DeepQNetwork(4, "")
snakeGame = SnakeEnv()
#先给一个向右走的决策输入，启动游戏
observation, reward, terminal,score =snakeGame.step(np.array([0, 0, 0, 1]))
observation = pre_process(observation)
brain.set_init_state(observation[:,:,0])
#开始正式游戏
i = 1 # 步数
while i<=500000:
i = i + 1
action = brain.choose_action()
next_observation, reward, terminal, score = snakeGame.step(action)
# print(reward)
next_observation = pre_process(next_observation)
brain.learn(next_observation, action, reward, terminal)
if(i%100) == 0:
print(i)
# 画loss和round step的曲线
brain.plot_cost()
snakeGame.plot_cost()
class Application:
def __init__(self, args):
self.args = args
self.env = SnakeEnv(args.snake_head_x, args.snake_head_y, args.food_x, args.food_y)
self.agent = Agent(self.env.get_actions(), args.Ne, args.C, args.gamma)
def execute(self):
if not self.args.human:
if self.args.train_eps != 0:
self.train()
self.eval()
self.show_games()
def train(self):
print("Train Phase:")
self.agent.train()
window = self.args.window
self.points_results = []
first_eat = True
start = time.time()
for game in range(1, self.args.train_eps + 1):
state = self.env.get_state()
dead = False
action = self.agent.choose_action(state, 0, dead)
while not dead:
state, points, dead = self.env.step(action)
# For debug convenience, you can check if your Q-table mathches ours for given setting of parameters
# (see Debug Convenience part on homework 4 web page)
if first_eat and points == 1:
self.agent.save_model(utils.CHECKPOINT)
first_eat = False
action = self.agent.choose_action(state, points, dead)
points = self.env.get_points()
self.points_results.append(points)
if game % self.args.window == 0:
print(
"Games:", len(self.points_results) - window, "-", len(self.points_results),
"Points (Average:", sum(self.points_results[-window:])/window,
"Max:", max(self.points_results[-window:]),
"Min:", min(self.points_results[-window:]),")",
)
self.env.reset()
print("Training takes", time.time() - start, "seconds")
self.agent.save_model(self.args.model_name)
def eval(self):
print("Evaling Phase:")
self.agent.eval()
self.agent.load_model(self.args.model_name)
points_results = []
start = time.time()
for game in range(1, self.args.test_eps + 1):
state = self.env.get_state()
dead = False
action = self.agent.choose_action(state, 0, dead)
while not dead:
state, points, dead = self.env.step(action)
action = self.agent.choose_action(state, points, dead)
points = self.env.get_points()
points_results.append(points)
self.env.reset()
print("Testing takes", time.time() - start, "seconds")
print("Number of Games:", len(points_results))
print("Average Points:", sum(points_results)/len(points_results))
print("Max Points:", max(points_results))
print("Min Points:", min(points_results))
def show_games(self):
print("Display Games")
self.env.display()
pygame.event.pump()
self.agent.eval()
points_results = []
end = False
for game in range(1, self.args.show_eps + 1):
state = self.env.get_state()
dead = False
action = self.agent.choose_action(state, 0, dead)
count = 0
while not dead:
count +=1
pygame.event.pump()
keys = pygame.key.get_pressed()
if keys[K_ESCAPE] or self.check_quit():
end = True
break
state, points, dead = self.env.step(action)
# Qlearning agent
if not self.args.human:
action = self.agent.choose_action(state, points, dead)
# for human player
else:
for event in pygame.event.get():
if event.type == pygame.KEYDOWN:
if event.key == pygame.K_UP:
action = 2
elif event.key == pygame.K_DOWN:
action = 3
elif event.key == pygame.K_LEFT:
action = 1
elif event.key == pygame.K_RIGHT:
action = 0
if end:
break
self.env.reset()
points_results.append(points)
print("Game:", str(game)+"/"+str(self.args.show_eps), "Points:", points)
if len(points_results) == 0:
return
print("Average Points:", sum(points_results)/len(points_results))
def check_quit(self):
for event in pygame.event.get():
if event.type == pygame.QUIT:
return True
return False
def run():
tf_env = tf_py_environment.TFPyEnvironment(SnakeEnv())
eval_env = tf_py_environment.TFPyEnvironment(SnakeEnv(step_limit=50))
q_net = q_network.QNetwork(
tf_env.observation_spec(),
tf_env.action_spec(),
conv_layer_params=(),
fc_layer_params=(512, 256, 128),
)
optimizer = tf.compat.v1.train.AdamOptimizer(learning_rate=learning_rate)
global_counter = tf.compat.v1.train.get_or_create_global_step()
agent = dqn_agent.DqnAgent(
tf_env.time_step_spec(),
tf_env.action_spec(),
q_network=q_net,
optimizer=optimizer,
td_errors_loss_fn=common.element_wise_squared_loss,
train_step_counter=global_counter,
gamma=0.95,
epsilon_greedy=0.1,
n_step_update=1,
)
root_dir = os.path.join('/tf-logs', 'snake')
train_dir = os.path.join(root_dir, 'train')
eval_dir = os.path.join(root_dir, 'eval')
agent.initialize()
train_metrics = [
tf_metrics.NumberOfEpisodes(),
tf_metrics.EnvironmentSteps(),
tf_metrics.AverageReturnMetric(),
tf_metrics.AverageEpisodeLengthMetric(),
]
replay_buffer = tf_uniform_replay_buffer.TFUniformReplayBuffer(
data_spec=agent.collect_data_spec,
batch_size=tf_env.batch_size,
max_length=replay_buffer_max_length,
)
collect_driver = dynamic_step_driver.DynamicStepDriver(
tf_env,
agent.collect_policy,
observers=[replay_buffer.add_batch] + train_metrics,
num_steps=collect_steps_per_iteration,
)
train_checkpointer = common.Checkpointer(
ckpt_dir=train_dir,
agent=agent,
global_step=global_counter,
metrics=metric_utils.MetricsGroup(train_metrics, 'train_metrics'),
)
policy_checkpointer = common.Checkpointer(
ckpt_dir=os.path.join(train_dir, 'policy'),
policy=agent.policy,
global_step=global_counter,
)
rb_checkpointer = common.Checkpointer(
ckpt_dir=os.path.join(train_dir, 'replay_buffer'),
max_to_keep=1,
replay_buffer=replay_buffer,
)
train_checkpointer.initialize_or_restore()
rb_checkpointer.initialize_or_restore()
collect_driver.run = common.function(collect_driver.run)
agent.train = common.function(agent.train)
random_policy = random_tf_policy.RandomTFPolicy(tf_env.time_step_spec(),
tf_env.action_spec())
if replay_buffer.num_frames() >= initial_collect_steps:
logging.info("We loaded memories, not doing random seed")
else:
logging.info("Capturing %d steps to seed with random memories",
initial_collect_steps)
dynamic_step_driver.DynamicStepDriver(
tf_env,
random_policy,
observers=[replay_buffer.add_batch] + train_metrics,
num_steps=initial_collect_steps).run()
train_summary_writer = tf.summary.create_file_writer(train_dir)
train_summary_writer.set_as_default()
avg_returns = []
avg_return_metric = tf_metrics.AverageReturnMetric(
buffer_size=num_eval_episodes)
eval_metrics = [
avg_return_metric,
tf_metrics.AverageEpisodeLengthMetric(buffer_size=num_eval_episodes),
]
logging.info("Running initial evaluation")
results = metric_utils.eager_compute(
eval_metrics,
eval_env,
agent.policy,
num_episodes=num_eval_episodes,
train_step=global_counter,
summary_writer=tf.summary.create_file_writer(eval_dir),
summary_prefix='Metrics',
)
avg_returns.append(
(global_counter.numpy(), avg_return_metric.result().numpy()))
metric_utils.log_metrics(eval_metrics)
time_step = None
policy_state = agent.collect_policy.get_initial_state(tf_env.batch_size)
timed_at_step = global_counter.numpy()
time_acc = 0
dataset = replay_buffer.as_dataset(num_parallel_calls=3,
sample_batch_size=batch_size,
num_steps=2).prefetch(3)
iterator = iter(dataset)
@common.function
def train_step():
experience, _ = next(iterator)
return agent.train(experience)
for _ in range(num_iterations):
start_time = time.time()
time_step, policy_state = collect_driver.run(
time_step=time_step,
policy_state=policy_state,
)
for _ in range(train_steps_per_iteration):
train_loss = train_step()
time_acc += time.time() - start_time
step = global_counter.numpy()
if step % log_interval == 0:
logging.info("step = %d, loss = %f", step, train_loss.loss)
steps_per_sec = (step - timed_at_step) / time_acc
logging.info("%.3f steps/sec", steps_per_sec)
timed_at_step = step
time_acc = 0
for train_metric in train_metrics:
train_metric.tf_summaries(train_step=global_counter,
step_metrics=train_metrics[:2])
if step % train_checkpoint_interval == 0:
train_checkpointer.save(global_step=step)
if step % policy_checkpoint_interval == 0:
policy_checkpointer.save(global_step=step)
if step % rb_checkpoint_interval == 0:
rb_checkpointer.save(global_step=step)
if step % capture_interval == 0:
print("Capturing run:")
capture_run(os.path.join(root_dir, "snake" + str(step) + ".mp4"),
eval_env, agent.policy)
if step % eval_interval == 0:
print("EVALUTION TIME:")
results = metric_utils.eager_compute(
eval_metrics,
eval_env,
agent.policy,
num_episodes=num_eval_episodes,
train_step=global_counter,
summary_writer=tf.summary.create_file_writer(eval_dir),
summary_prefix='Metrics',
)
metric_utils.log_metrics(eval_metrics)
avg_returns.append(
(global_counter.numpy(), avg_return_metric.result().numpy()))
state = env.reset()
return_val = 0
while True:
act = agent.play(state)
state, reward, terminate, _ = env.step(act)
return_val += reward
if terminate:
break
return return_val
def test_agent(env, agent):
agent.learn()
print ('states->actions pi = {}'.format(agent.get_pi()))
sum_reward = 0
for i in range(100):
sum_reward += eval_policy(env, agent)
print ("avg reward = {}".format(sum_reward / 100.))
if __name__ == '__main__':
env = SnakeEnv(0, [3, 6])
# test_agent(env, TableAgent(env, PolicyIter(-1)))
# test_agent(env, TableAgent(env, ValueIter(-1)))
# test_agent(env, TableAgent(env, GeneralizedIter(1, 10)))
# test_agent(env, ModelFreeAgent(env, MonteCarlo(0.1)))
# test_agent(env, ModelFreeAgent(env, Sarsa(0.1)))
# test_agent(env, ModelFreeAgent(env, QLearning(0.1)))
# test_agent(env, DQNAgent(env)) # SB玩意不收敛
test_agent(env, OpenAiAgent(env))
def set_up_env(io):
env_py = SnakeEnv(io)
train_env = tf_py_environment.TFPyEnvironment(env_py)
eval_env = tf_py_environment.TFPyEnvironment(env_py)
return train_env, eval_env
def __init__(self, args):
self.args = args
self.env = SnakeEnv(args.snake_head_x, args.snake_head_y, args.food_x, args.food_y)
self.agent = Agent(self.env.get_actions(), args.Ne, args.C, args.gamma)
tf.keras.layers.Dense(ACTION_DIM, activation="linear")
])
model.compile(loss='mean_squared_error',
optimizer=tf.keras.optimizers.Adam(0.001))
if weights:
model.load_weights(weights)
return model
def act(state, epsilon=0.1, step=0):
"""预测动作"""
return np.argmax(model.predict(np.array([state]))[0])
env = SnakeEnv()
model = create_model("weights.hdf5")
for i in range(1000):
state = env.reset()
step = 0
while True:
env.render()
model.predict(np.array([state]))
done = True
#next_state, reward, done, _ = env.step(0)
step += 1
if done:
print('Game', i + 1, ' Score:', env.score)
break
env.close()
s_batch = np.array([replay[0] for replay in replay_batch])
next_s_batch = np.array([replay[2] for replay in replay_batch])
Q = self.model.predict(s_batch)
Q_next = self.target_model.predict(next_s_batch)
# 使用公式更新训练集中的Q值
for i, replay in enumerate(replay_batch):
_, a, _, reward = replay
Q[i][a] = (1 - lr) * Q[i][a] + lr * (reward + factor * np.amax(Q_next[i]))
# 传入网络进行训练
self.model.fit(s_batch, Q, verbose=0)
env = SnakeEnv()
episodes = 1000 # 训练次数
agent = DQN()
for i in range(episodes):
state = env.reset()
while True:
# env.render(speed=0)
action = agent.act(state)
next_state, reward, done, _ = env.step(action)
agent.remember(state, action, next_state, reward)
agent.train()
state = next_state
if done:
print('Game', i + 1, ' Score:', env.score)
break
import sys
from snake_env import SnakeEnv
import numpy as np
from DQNetwork import DQNetwork
from Results import Results, test
from Options import Options
import pickle
import time
opt = Options().parse()
nrow, ncol = opt.gridsize, opt.gridsize
n_channels = opt.n_ch
if n_channels == 1:
env = SnakeEnv(nrow, ncol, colors='gray')
elif n_channels == 3:
env = SnakeEnv(nrow, ncol, colors='rgb')
n_train = opt.n_train
n_episodes = opt.n_episodes
n_batch = opt.n_batch
imax = opt.imax
min_epsilon = opt.min_epsilon
N_memory = opt.n_memory
model = DQNetwork(4, (n_channels, nrow, ncol), conv=opt.conv)
res = Results()
loadModel = opt.load
def play(record=0, no_render=False):
env = SnakeEnv(need_render=not no_render, alg='最短路径')
obs = env.reset()
env.render()
input()
x, y = [], []
directions = {
(-1, 0): env.right,
(1, 0): env.left,
(0, -1): env.down,
(0, 1): env.up
}
need_record = True if record else False
new_dst = None
origin_dst = None
# counter = 20
use_random = False
while True:
if not record and not no_render:
env.render()
src = np.where(obs == 2)
src = int(src[1]), int(src[0])
dst = np.where(obs == -1)
dst = int(dst[0]), int(dst[1])
if new_dst is not None:
paths = bfs(obs, start=src, dst=new_dst)
else:
paths = bfs(obs, start=src, dst=dst)
if paths is None:
# origin_dst = dst
# new_dst = (
# np.random.randint(0, obs.shape[0]),
# np.random.randint(0, obs.shape[1]),
# )
# counter -= 1
# if counter <= 0:
# print('score: ', env.status.score)
# new_dst = None
# origin_dst = None
# counter = 20
# obs = env.reset()
# continue
use_random = True
else:
new_dst = None
if new_dst is not None and paths[1] == new_dst:
new_dst = None
if origin_dst is not None:
dst = origin_dst
origin_dst = None
# counter = 20
continue
# if counter <= 0 or paths is None or len(paths) <= 1:
# print('score: ', env.status.score)
# obs = env.reset()
# continue
if use_random:
action = np.random.randint(0, 4)
use_random = False
else:
dst = paths[1]
dire = src[0] - dst[0], src[1] - dst[1]
action = directions[dire]
# import ipdb
# ipdb.set_trace()
if need_record:
x.append(obs)
y.append(action)
if len(y) >= record:
return x, y
if len(y) % 1000 == 0:
print(len(y))
_, obs, done, _ = env(action)
# counter = 20
if done:
print(env.status.score)
sleep(1.5)
break
if not record and not no_render:
env.render()
env.close()
def draw_graph():
print(GRID_HEIGHT_NUM, GRID_WIDTH_NUM)
# input()
graph = build_graph(row=GRID_HEIGHT_NUM, col=GRID_WIDTH_NUM)
total_graph = deepcopy(graph)
env = SnakeEnv(set_life=100000, alg='HC + BFS', no_sight_disp=True)
env.reset()
sleep(1)
graph, flags = deletion(graph, env)
for sp in graph:
for ep in graph[sp]:
if flags[(sp, ep)]:
# print(sp, ep)
env.draw_connection(sp, ep, width=4)
# env.render()
import pygame
pygame.display.update()
pre_len = None
while True:
sd_len = destroy(graph, total_graph, flags, env=env)
print('sd: ', sd_len)
if pre_len is not None and pre_len == sd_len:
global MAX_DEPTH
print('+1')
MAX_DEPTH += 1
pre_len = sd_len
show_graph(graph, flags, env)
if not sd_len:
break
sleep(1)
show_graph(graph, flags, env)
counter = 0
while not connector(graph, total_graph, flags, env):
counter += 1
print('counter: ', counter)
sleep(1)
for sp in graph:
for ep in graph[sp]:
if flags[(sp, ep)]:
env.draw_connection(sp, ep, color=(0xff, 0xff, 0), width=4)
import pygame
show_graph(graph, flags, env)
circle = get_list_circle(graph)
print(circle)
pos_encoder = {pos: i for i, pos in enumerate(circle)}
# pos_decoder = {i: pos for i, pos in enumerate(circle)}
pos_xy_decoder = {
i: (pos % GRID_WIDTH_NUM, pos // GRID_WIDTH_NUM)
for i, pos in enumerate(circle)
}
pos_xy_encoder = {(pos % GRID_WIDTH_NUM, pos // GRID_WIDTH_NUM): i
for i, pos in enumerate(circle)}
obs = env.reset()
c = 0
while True:
c += 1
if len(env.status.snake_body) < 15:
remainder = 20
elif len(env.status.snake_body) < 30:
remainder = 20
elif len(env.status.snake_body) < 60:
remainder = 30
elif len(env.status.snake_body) < 90:
remainder = 30
elif len(env.status.snake_body) < 120:
remainder = 40
elif len(env.status.snake_body) < 150:
remainder = 80
elif len(env.status.snake_body) < 300:
remainder = 100
elif len(env.status.snake_body) < (GRID_WIDTH_NUM * GRID_HEIGHT_NUM -
10):
remainder = 30
else:
remainder = 5
bfs_action, dst = dfs_policy(obs, env)
bfs_dst_idx = 100000000
if dst:
bfs_dst_idx = pos_xy_encoder[dst]
head = env.status.snake_body[0]
head_pos, tail_pos = pos_xy_encoder[head], pos_xy_encoder[
env.status.snake_body[-1]]
head_idx, tail_idx = pos_xy_encoder[head], pos_xy_encoder[
env.status.snake_body[-1]]
hc_next_pos = pos_xy_decoder[(head_pos + 1) % len(graph)]
directions = {
(-1, 0): env.right,
(1, 0): env.left,
(0, -1): env.down,
(0, 1): env.up
}
dire = head[0] - hc_next_pos[0], head[1] - hc_next_pos[1]
print(head, hc_next_pos, dst, dst not in env.status.snake_body[:-1])
print(head_idx, tail_idx, bfs_dst_idx)
action = directions[dire]
if not env.status.food_pos:
show_graph(graph,
flags,
env,
update=True,
width=1,
extra='倍速: {} X'.format(remainder * 5))
break
food_idx = pos_xy_encoder[env.status.food_pos]
if bfs_action:
print(food_idx, bfs_dst_idx, head_idx, tail_idx)
print(rel_pos(food_idx, tail_idx, len(graph)),
rel_pos(bfs_dst_idx, tail_idx, len(graph)),
rel_pos(head_idx, tail_idx, len(graph)),
rel_pos(tail_idx, tail_idx, len(graph)))
if rel_pos(food_idx, tail_idx, len(graph)) >= rel_pos(
bfs_dst_idx, tail_idx, len(graph)) >= rel_pos(
head_idx, tail_idx, len(graph)) >= rel_pos(
tail_idx, tail_idx, len(graph)):
action = bfs_action
pass
reward, obs, done, _ = env(action)
if done:
show_graph(graph,
flags,
env,
update=True,
width=1,
extra='倍速: {} X'.format(remainder * 5))
print(done)
break
# env.screen.blit(env.background, (0, 0))
if c % remainder == 0:
show_graph(graph,
flags,
env,
update=True,
width=1,
extra='倍速: {} X'.format(remainder * 5))
# env.render(blit=False)
show_graph(graph, flags, env, update=True, width=1)
sleep(10)
input()
import argparse
import gym
from stable_baselines.common.policies import CnnPolicy
from stable_baselines.common.vec_env import DummyVecEnv
from stable_baselines import PPO2
from snake_env import SnakeEnv
parser = argparse.ArgumentParser()
parser.add_argument('--mode', choices=['train', 'test'], default='test')
args = parser.parse_args()
env = SnakeEnv((20, 20), 'standard')
env = DummyVecEnv([lambda: env])
model = PPO2(CnnPolicy, env, verbose=1)
if args.mode == 'train':
model.learn(total_timesteps=20000)
model.save('policy_baseline_snake')
elif args.mode == 'test':
obs = env.reset()
model.load('policy_baseline_snake')
for i in range(1000):
action, _states = model.predict(obs)
obs, reward, done, info = env.step(action)
env.render()
if done:
# -*- coding: utf-8 -*-
import time
from snake_env import SnakeEnv
from simple_mlp import Policy
def play(env, policy):
obs = env.reset()
while True:
action = policy.predict(obs)
reward, obs, done, _ = env(action)
env.render()
if done:
obs = env.reset()
time.sleep(1)
# time.sleep(0.05)
if __name__ == '__main__':
policy = Policy(pre_trained='pretrained/mlp-v0.joblib')
env = SnakeEnv(alg='MLP')
env.reset()
env.render()
input()
play(env, policy)
