Introduction to Value based Deep Reinforcement Learning
그로킹 심층 강화학습 중 8장 내용인 "가치기반 심층 강화학습 개요"에 대한 내용입니다.
Note: 실행을 위해 아래의 패키지들을 설치해주기 바랍니다.
!pip install tqdm numpy scikit-learn pyglet setuptools && \
!pip install gym asciinema pandas tabulate tornado==5.* PyBullet && \
!pip install git+https://github.com/pybox2d/pybox2d#egg=Box2D && \
!pip install git+https://github.com/mimoralea/gym-bandits#egg=gym-bandits && \
!pip install git+https://github.com/mimoralea/gym-walk#egg=gym-walk && \
!pip install git+https://github.com/mimoralea/gym-aima#egg=gym-aima && \
!pip install gym[atari]
!pip install torch torchvision
import warnings ; warnings.filterwarnings('ignore')
import os
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
import numpy as np
from IPython.display import display
from collections import namedtuple, deque
import matplotlib.pyplot as plt
import matplotlib.pylab as pylab
from itertools import cycle, count
from textwrap import wrap
import matplotlib
import subprocess
import os.path
import tempfile
import random
import base64
import pprint
import glob
import time
import json
import sys
import gym
import io
import os
import gc
import platform
from gym import wrappers
from subprocess import check_output
from IPython.display import HTML
LEAVE_PRINT_EVERY_N_SECS = 60
ERASE_LINE = '\x1b[2K'
EPS = 1e-6
RESULTS_DIR = os.path.join('.', 'gym-results')
SEEDS = (12, 34, 56, 78, 90)
%matplotlib inline
plt.style.use('fivethirtyeight')
params = {
'figure.figsize': (15, 8),
'font.size': 24,
'legend.fontsize': 20,
'axes.titlesize': 28,
'axes.labelsize': 24,
'xtick.labelsize': 20,
'ytick.labelsize': 20
}
pylab.rcParams.update(params)
np.set_printoptions(suppress=True)
torch.cuda.is_available()
def get_make_env_fn(**kargs):
def make_env_fn(env_name, seed=None, render=None, record=False,
unwrapped=False, monitor_mode=None,
inner_wrappers=None, outer_wrappers=None):
mdir = tempfile.mkdtemp()
env = None
if render:
try:
env = gym.make(env_name, render=render)
except:
pass
if env is None:
env = gym.make(env_name)
if seed is not None: env.seed(seed)
env = env.unwrapped if unwrapped else env
if inner_wrappers:
for wrapper in inner_wrappers:
env = wrapper(env)
env = wrappers.Monitor(
env, mdir, force=True,
mode=monitor_mode,
video_callable=lambda e_idx: record) if monitor_mode else env
if outer_wrappers:
for wrapper in outer_wrappers:
env = wrapper(env)
return env
return make_env_fn, kargs
def get_videos_html(env_videos, title, max_n_videos=5):
videos = np.array(env_videos)
if len(videos) == 0:
return
n_videos = max(1, min(max_n_videos, len(videos)))
idxs = np.linspace(0, len(videos) - 1, n_videos).astype(int) if n_videos > 1 else [-1,]
videos = videos[idxs,...]
strm = '<h2>{}</h2>'.format(title)
for video_path, meta_path in videos:
video = io.open(video_path, 'r+b').read()
encoded = base64.b64encode(video)
with open(meta_path) as data_file:
meta = json.load(data_file)
html_tag = """
<h3>{0}</h3>
<video width="960" height="540" controls>
<source src="data:video/mp4;base64,{1}" type="video/mp4" />
</video>"""
strm += html_tag.format('Episode ' + str(meta['episode_id']), encoded.decode('ascii'))
return strm
platform.system()
def get_gif_html(env_videos, title, subtitle_eps=None, max_n_videos=4):
videos = np.array(env_videos)
if len(videos) == 0:
return
n_videos = max(1, min(max_n_videos, len(videos)))
idxs = np.linspace(0, len(videos) - 1, n_videos).astype(int) if n_videos > 1 else [-1,]
videos = videos[idxs,...]
strm = '<h2>{}</h2>'.format(title)
for video_path, meta_path in videos:
basename = os.path.splitext(video_path)[0]
gif_path = basename + '.gif'
if not os.path.exists(gif_path):
if platform.system() == 'Linux':
ps = subprocess.Popen(
('ffmpeg',
'-i', video_path,
'-r', '7',
'-f', 'image2pipe',
'-vcodec', 'ppm',
'-crf', '20',
'-vf', 'scale=512:-1',
'-'),
stdout=subprocess.PIPE,
universal_newlines=True)
output = subprocess.check_output(
('convert',
'-coalesce',
'-delay', '7',
'-loop', '0',
'-fuzz', '2%',
'+dither',
'-deconstruct',
'-layers', 'Optimize',
'-', gif_path),
stdin=ps.stdout)
ps.wait()
else:
ps = subprocess.Popen('ffmpeg -i {} -r 7 -f image2pipe \
-vcodec ppm -crf 20 -vf scale=512:-1 - | \
convert -coalesce -delay 7 -loop 0 -fuzz 2% \
+dither -deconstruct -layers Optimize \
- {}'.format(video_path, gif_path),
stdin=subprocess.PIPE,
shell=True)
ps.wait()
gif = io.open(gif_path, 'r+b').read()
encoded = base64.b64encode(gif)
with open(meta_path) as data_file:
meta = json.load(data_file)
html_tag = """
<h3>{0}</h3>
<img src="data:image/gif;base64,{1}" />"""
prefix = 'Trial ' if subtitle_eps is None else 'Episode '
sufix = str(meta['episode_id'] if subtitle_eps is None \
else subtitle_eps[meta['episode_id']])
strm += html_tag.format(prefix + sufix, encoded.decode('ascii'))
return strm
class DiscountedCartPole(gym.Wrapper):
def __init__(self, env):
gym.Wrapper.__init__(self, env)
def reset(self, **kwargs):
return self.env.reset(**kwargs)
def step(self, a):
o, r, d, _ = self.env.step(a)
(x, x_dot, theta, theta_dot) = o
pole_fell = x < -self.env.unwrapped.x_threshold \
or x > self.env.unwrapped.x_threshold \
or theta < -self.env.unwrapped.theta_threshold_radians \
or theta > self.env.unwrapped.theta_threshold_radians
r = -1 if pole_fell else 0
return o, r, d, _
class FCQ(nn.Module):
def __init__(self,
input_dim,
output_dim,
hidden_dims=(32,32),
activation_fc=F.relu):
super(FCQ, self).__init__()
self.activation_fc = activation_fc
self.input_layer = nn.Linear(input_dim, hidden_dims[0])
self.hidden_layers = nn.ModuleList()
for i in range(len(hidden_dims)-1):
hidden_layer = nn.Linear(hidden_dims[i], hidden_dims[i+1])
self.hidden_layers.append(hidden_layer)
self.output_layer = nn.Linear(hidden_dims[-1], output_dim)
device = "cpu"
if torch.cuda.is_available():
device = "cuda:0"
self.device = torch.device(device)
self.to(self.device)
def _format(self, state):
x = state
if not isinstance(x, torch.Tensor):
x = torch.tensor(x,
device=self.device,
dtype=torch.float32)
x = x.unsqueeze(0)
return x
def forward(self, state):
x = self._format(state)
x = self.activation_fc(self.input_layer(x))
for hidden_layer in self.hidden_layers:
x = self.activation_fc(hidden_layer(x))
x = self.output_layer(x)
return x
def numpy_float_to_device(self, variable):
variable = torch.from_numpy(variable).float().to(self.device)
return variable
def load(self, experiences):
states, actions, new_states, rewards, is_terminals = experiences
states = torch.from_numpy(states).float().to(self.device)
actions = torch.from_numpy(actions).long().to(self.device)
new_states = torch.from_numpy(new_states).float().to(self.device)
rewards = torch.from_numpy(rewards).float().to(self.device)
is_terminals = torch.from_numpy(is_terminals).float().to(self.device)
return states, actions, new_states, rewards, is_terminals
class GreedyStrategy():
def __init__(self):
self.exploratory_action_taken = False
def select_action(self, model, state):
with torch.no_grad():
q_values = model(state).cpu().detach().data.numpy().squeeze()
return np.argmax(q_values)
class EGreedyStrategy():
def __init__(self, epsilon=0.1):
self.epsilon = epsilon
self.exploratory_action_taken = None
def select_action(self, model, state):
self.exploratory_action_taken = False
with torch.no_grad():
q_values = model(state).cpu().detach().data.numpy().squeeze()
if np.random.rand() > self.epsilon:
action = np.argmax(q_values)
else:
action = np.random.randint(len(q_values))
self.exploratory_action_taken = action != np.argmax(q_values)
return action
class NFQ():
def __init__(self,
value_model_fn,
value_optimizer_fn,
value_optimizer_lr,
training_strategy_fn,
evaluation_strategy_fn,
batch_size,
epochs):
self.value_model_fn = value_model_fn
self.value_optimizer_fn = value_optimizer_fn
self.value_optimizer_lr = value_optimizer_lr
self.training_strategy_fn = training_strategy_fn
self.evaluation_strategy_fn = evaluation_strategy_fn
self.batch_size = batch_size
self.epochs = epochs
def optimize_model(self, experiences):
states, actions, rewards, next_states, is_terminals = experiences
batch_size = len(is_terminals)
max_a_q_sp = self.online_model(next_states).detach().max(1)[0].unsqueeze(1)
target_q_s = rewards + self.gamma * max_a_q_sp * (1 - is_terminals)
q_sa = self.online_model(states).gather(1, actions)
td_errors = q_sa - target_q_s
value_loss = td_errors.pow(2).mul(0.5).mean()
self.value_optimizer.zero_grad()
value_loss.backward()
self.value_optimizer.step()
def interaction_step(self, state, env):
action = self.training_strategy.select_action(self.online_model, state)
new_state, reward, is_terminal, info = env.step(action)
is_truncated = 'TimeLimit.truncated' in info and info['TimeLimit.truncated']
is_failure = is_terminal and not is_truncated
experience = (state, action, reward, new_state, float(is_failure))
self.experiences.append(experience)
self.episode_reward[-1] += reward
self.episode_timestep[-1] += 1
self.episode_exploration[-1] += int(self.training_strategy.exploratory_action_taken)
return new_state, is_terminal
def train(self, make_env_fn, make_env_kargs, seed, gamma,
max_minutes, max_episodes, goal_mean_100_reward):
training_start, last_debug_time = time.time(), float('-inf')
self.checkpoint_dir = tempfile.mkdtemp()
self.make_env_fn = make_env_fn
self.make_env_kargs = make_env_kargs
self.seed = seed
self.gamma = gamma
env = self.make_env_fn(**self.make_env_kargs, seed=self.seed)
torch.manual_seed(self.seed) ; np.random.seed(self.seed) ; random.seed(self.seed)
nS, nA = env.observation_space.shape[0], env.action_space.n
self.episode_timestep = []
self.episode_reward = []
self.episode_seconds = []
self.evaluation_scores = []
self.episode_exploration = []
self.online_model = self.value_model_fn(nS, nA)
self.value_optimizer = self.value_optimizer_fn(self.online_model,
self.value_optimizer_lr)
self.training_strategy = training_strategy_fn()
self.evaluation_strategy = evaluation_strategy_fn()
self.experiences = []
result = np.empty((max_episodes, 5))
result[:] = np.nan
training_time = 0
for episode in range(1, max_episodes + 1):
episode_start = time.time()
state, is_terminal = env.reset(), False
self.episode_reward.append(0.0)
self.episode_timestep.append(0.0)
self.episode_exploration.append(0.0)
for step in count():
state, is_terminal = self.interaction_step(state, env)
if len(self.experiences) >= self.batch_size:
experiences = np.array(self.experiences)
batches = [np.vstack(sars) for sars in experiences.T]
experiences = self.online_model.load(batches)
for _ in range(self.epochs):
self.optimize_model(experiences)
self.experiences.clear()
if is_terminal:
gc.collect()
break
# stats
episode_elapsed = time.time() - episode_start
self.episode_seconds.append(episode_elapsed)
training_time += episode_elapsed
evaluation_score, _ = self.evaluate(self.online_model, env)
self.save_checkpoint(episode-1, self.online_model)
total_step = int(np.sum(self.episode_timestep))
self.evaluation_scores.append(evaluation_score)
mean_10_reward = np.mean(self.episode_reward[-10:])
std_10_reward = np.std(self.episode_reward[-10:])
mean_100_reward = np.mean(self.episode_reward[-100:])
std_100_reward = np.std(self.episode_reward[-100:])
mean_100_eval_score = np.mean(self.evaluation_scores[-100:])
std_100_eval_score = np.std(self.evaluation_scores[-100:])
lst_100_exp_rat = np.array(
self.episode_exploration[-100:])/np.array(self.episode_timestep[-100:])
mean_100_exp_rat = np.mean(lst_100_exp_rat)
std_100_exp_rat = np.std(lst_100_exp_rat)
wallclock_elapsed = time.time() - training_start
result[episode-1] = total_step, mean_100_reward, \
mean_100_eval_score, training_time, wallclock_elapsed
reached_debug_time = time.time() - last_debug_time >= LEAVE_PRINT_EVERY_N_SECS
reached_max_minutes = wallclock_elapsed >= max_minutes * 60
reached_max_episodes = episode >= max_episodes
reached_goal_mean_reward = mean_100_eval_score >= goal_mean_100_reward
training_is_over = reached_max_minutes or \
reached_max_episodes or \
reached_goal_mean_reward
elapsed_str = time.strftime("%H:%M:%S", time.gmtime(time.time() - training_start))
debug_message = 'el {}, ep {:04}, ts {:06}, '
debug_message += 'ar 10 {:05.1f}\u00B1{:05.1f}, '
debug_message += '100 {:05.1f}\u00B1{:05.1f}, '
debug_message += 'ex 100 {:02.1f}\u00B1{:02.1f}, '
debug_message += 'ev {:05.1f}\u00B1{:05.1f}'
debug_message = debug_message.format(
elapsed_str, episode-1, total_step, mean_10_reward, std_10_reward,
mean_100_reward, std_100_reward, mean_100_exp_rat, std_100_exp_rat,
mean_100_eval_score, std_100_eval_score)
print(debug_message, end='\r', flush=True)
if reached_debug_time or training_is_over:
print(ERASE_LINE + debug_message, flush=True)
last_debug_time = time.time()
if training_is_over:
if reached_max_minutes: print(u'--> reached_max_minutes \u2715')
if reached_max_episodes: print(u'--> reached_max_episodes \u2715')
if reached_goal_mean_reward: print(u'--> reached_goal_mean_reward \u2713')
break
final_eval_score, score_std = self.evaluate(self.online_model, env, n_episodes=100)
wallclock_time = time.time() - training_start
print('Training complete.')
print('Final evaluation score {:.2f}\u00B1{:.2f} in {:.2f}s training time,'
' {:.2f}s wall-clock time.\n'.format(
final_eval_score, score_std, training_time, wallclock_time))
env.close() ; del env
self.get_cleaned_checkpoints()
return result, final_eval_score, training_time, wallclock_time
def evaluate(self, eval_policy_model, eval_env, n_episodes=1):
rs = []
for _ in range(n_episodes):
s, d = eval_env.reset(), False
rs.append(0)
for _ in count():
a = self.evaluation_strategy.select_action(eval_policy_model, s)
s, r, d, _ = eval_env.step(a)
rs[-1] += r
if d: break
return np.mean(rs), np.std(rs)
def get_cleaned_checkpoints(self, n_checkpoints=5):
try:
return self.checkpoint_paths
except AttributeError:
self.checkpoint_paths = {}
paths = glob.glob(os.path.join(self.checkpoint_dir, '*.tar'))
paths_dic = {int(path.split('.')[-2]):path for path in paths}
last_ep = max(paths_dic.keys())
# checkpoint_idxs = np.geomspace(1, last_ep+1, n_checkpoints, endpoint=True, dtype=np.int)-1
checkpoint_idxs = np.linspace(1, last_ep+1, n_checkpoints, endpoint=True, dtype=np.int)-1
for idx, path in paths_dic.items():
if idx in checkpoint_idxs:
self.checkpoint_paths[idx] = path
else:
os.unlink(path)
return self.checkpoint_paths
def demo_last(self, title='Fully-trained {} Agent', n_episodes=3, max_n_videos=3):
env = self.make_env_fn(**self.make_env_kargs, monitor_mode='evaluation', render=True, record=True)
checkpoint_paths = self.get_cleaned_checkpoints()
last_ep = max(checkpoint_paths.keys())
self.online_model.load_state_dict(torch.load(checkpoint_paths[last_ep]))
self.evaluate(self.online_model, env, n_episodes=n_episodes)
env.close()
data = get_gif_html(env_videos=env.videos,
title=title.format(self.__class__.__name__),
max_n_videos=max_n_videos)
del env
return HTML(data=data)
def demo_progression(self, title='{} Agent progression', max_n_videos=5):
env = self.make_env_fn(**self.make_env_kargs, monitor_mode='evaluation', render=True, record=True)
checkpoint_paths = self.get_cleaned_checkpoints()
for i in sorted(checkpoint_paths.keys()):
self.online_model.load_state_dict(torch.load(checkpoint_paths[i]))
self.evaluate(self.online_model, env, n_episodes=1)
env.close()
data = get_gif_html(env_videos=env.videos,
title=title.format(self.__class__.__name__),
subtitle_eps=sorted(checkpoint_paths.keys()),
max_n_videos=max_n_videos)
del env
return HTML(data=data)
def save_checkpoint(self, episode_idx, model):
torch.save(model.state_dict(),
os.path.join(self.checkpoint_dir, 'model.{}.tar'.format(episode_idx)))
nfq_results = []
best_agent, best_eval_score = None, float('-inf')
for seed in SEEDS:
environment_settings = {
'env_name': 'CartPole-v1',
'gamma': 1.00,
'max_minutes': 20,
'max_episodes': 10000,
'goal_mean_100_reward': 475
}
value_model_fn = lambda nS, nA: FCQ(nS, nA, hidden_dims=(512,128))
# value_optimizer_fn = lambda net, lr: optim.Adam(net.parameters(), lr=lr)
value_optimizer_fn = lambda net, lr: optim.RMSprop(net.parameters(), lr=lr)
value_optimizer_lr = 0.0005
training_strategy_fn = lambda: EGreedyStrategy(epsilon=0.5)
# evaluation_strategy_fn = lambda: EGreedyStrategy(epsilon=0.05)
evaluation_strategy_fn = lambda: GreedyStrategy()
batch_size = 1024
epochs = 40
env_name, gamma, max_minutes, \
max_episodes, goal_mean_100_reward = environment_settings.values()
agent = NFQ(value_model_fn,
value_optimizer_fn,
value_optimizer_lr,
training_strategy_fn,
evaluation_strategy_fn,
batch_size,
epochs)
# make_env_fn, make_env_kargs = get_make_env_fn(
# env_name=env_name, addon_wrappers=[DiscountedCartPole,])
make_env_fn, make_env_kargs = get_make_env_fn(env_name=env_name)
result, final_eval_score, training_time, wallclock_time = agent.train(
make_env_fn, make_env_kargs, seed, gamma, max_minutes, max_episodes, goal_mean_100_reward)
nfq_results.append(result)
if final_eval_score > best_eval_score:
best_eval_score = final_eval_score
best_agent = agent
nfq_results = np.array(nfq_results)
best_agent.demo_progression()
NFQ Agent progression
Episode 0
Episode 691
Episode 1382
Episode 2073
Episode 2764
best_agent.demo_last()
Fully-trained NFQ Agent
Trial 0
Trial 1
Trial 2
nfq_max_t, nfq_max_r, nfq_max_s, \
nfq_max_sec, nfq_max_rt = np.max(nfq_results, axis=0).T
nfq_min_t, nfq_min_r, nfq_min_s, \
nfq_min_sec, nfq_min_rt = np.min(nfq_results, axis=0).T
nfq_mean_t, nfq_mean_r, nfq_mean_s, \
nfq_mean_sec, nfq_mean_rt = np.mean(nfq_results, axis=0).T
nfq_x = np.arange(len(nfq_mean_s))
# nfq_max_t, nfq_max_r, nfq_max_s, \
# nfq_max_sec, nfq_max_rt = np.nanmax(nfq_results, axis=0).T
# nfq_min_t, nfq_min_r, nfq_min_s, \
# nfq_min_sec, nfq_min_rt = np.nanmin(nfq_results, axis=0).T
# nfq_mean_t, nfq_mean_r, nfq_mean_s, \
# nfq_mean_sec, nfq_mean_rt = np.nanmean(nfq_results, axis=0).T
# nfq_x = np.arange(len(nfq_mean_s))
fig, axs = plt.subplots(5, 1, figsize=(15,30), sharey=False, sharex=True)
# NFQ
axs[0].plot(nfq_max_r, 'y', linewidth=1)
axs[0].plot(nfq_min_r, 'y', linewidth=1)
axs[0].plot(nfq_mean_r, 'y', label='NFQ', linewidth=2)
axs[0].fill_between(nfq_x, nfq_min_r, nfq_max_r, facecolor='y', alpha=0.3)
axs[1].plot(nfq_max_s, 'y', linewidth=1)
axs[1].plot(nfq_min_s, 'y', linewidth=1)
axs[1].plot(nfq_mean_s, 'y', label='NFQ', linewidth=2)
axs[1].fill_between(nfq_x, nfq_min_s, nfq_max_s, facecolor='y', alpha=0.3)
axs[2].plot(nfq_max_t, 'y', linewidth=1)
axs[2].plot(nfq_min_t, 'y', linewidth=1)
axs[2].plot(nfq_mean_t, 'y', label='NFQ', linewidth=2)
axs[2].fill_between(nfq_x, nfq_min_t, nfq_max_t, facecolor='y', alpha=0.3)
axs[3].plot(nfq_max_sec, 'y', linewidth=1)
axs[3].plot(nfq_min_sec, 'y', linewidth=1)
axs[3].plot(nfq_mean_sec, 'y', label='NFQ', linewidth=2)
axs[3].fill_between(nfq_x, nfq_min_sec, nfq_max_sec, facecolor='y', alpha=0.3)
axs[4].plot(nfq_max_rt, 'y', linewidth=1)
axs[4].plot(nfq_min_rt, 'y', linewidth=1)
axs[4].plot(nfq_mean_rt, 'y', label='NFQ', linewidth=2)
axs[4].fill_between(nfq_x, nfq_min_rt, nfq_max_rt, facecolor='y', alpha=0.3)
# ALL
axs[0].set_title('Moving Avg Reward (Training)')
axs[1].set_title('Moving Avg Reward (Evaluation)')
axs[2].set_title('Total Steps')
axs[3].set_title('Training Time')
axs[4].set_title('Wall-clock Time')
plt.xlabel('Episodes')
axs[0].legend(loc='upper left')
plt.show()
nfq_root_dir = os.path.join(RESULTS_DIR, 'nfq')
not os.path.exists(nfq_root_dir) and os.makedirs(nfq_root_dir)
np.save(os.path.join(nfq_root_dir, 'x'), nfq_x)
np.save(os.path.join(nfq_root_dir, 'max_r'), nfq_max_r)
np.save(os.path.join(nfq_root_dir, 'min_r'), nfq_min_r)
np.save(os.path.join(nfq_root_dir, 'mean_r'), nfq_mean_r)
np.save(os.path.join(nfq_root_dir, 'max_s'), nfq_max_s)
np.save(os.path.join(nfq_root_dir, 'min_s'), nfq_min_s )
np.save(os.path.join(nfq_root_dir, 'mean_s'), nfq_mean_s)
np.save(os.path.join(nfq_root_dir, 'max_t'), nfq_max_t)
np.save(os.path.join(nfq_root_dir, 'min_t'), nfq_min_t)
np.save(os.path.join(nfq_root_dir, 'mean_t'), nfq_mean_t)
np.save(os.path.join(nfq_root_dir, 'max_sec'), nfq_max_sec)
np.save(os.path.join(nfq_root_dir, 'min_sec'), nfq_min_sec)
np.save(os.path.join(nfq_root_dir, 'mean_sec'), nfq_mean_sec)
np.save(os.path.join(nfq_root_dir, 'max_rt'), nfq_max_rt)
np.save(os.path.join(nfq_root_dir, 'min_rt'), nfq_min_rt)
np.save(os.path.join(nfq_root_dir, 'mean_rt'), nfq_mean_rt)