Copyright 2021 The TF-Agents Authors.
View on TensorFlow.org
|
Run in Google Colab
|
View source on GitHub
|
 Download notebook
|
Introduction
This example shows how to train a Categorical DQN (C51) agent on the Cartpole environment using the TF-Agents library.
Make sure you take a look through the DQN tutorial as a prerequisite. This tutorial will assume familiarity with the DQN tutorial; it will mainly focus on the differences between DQN and C51.
Setup
If you haven't installed tf-agents yet, run:
sudo apt-get updatesudo apt-get install -y xvfb ffmpegpip install 'imageio==2.4.0'pip install pyvirtualdisplaypip install tf-agents
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import base64
import imageio
import IPython
import matplotlib
import matplotlib.pyplot as plt
import PIL.Image
import pyvirtualdisplay
import tensorflow as tf
from tf_agents.agents.categorical_dqn import categorical_dqn_agent
from tf_agents.drivers import dynamic_step_driver
from tf_agents.environments import suite_gym
from tf_agents.environments import tf_py_environment
from tf_agents.eval import metric_utils
from tf_agents.metrics import tf_metrics
from tf_agents.networks import categorical_q_network
from tf_agents.policies import random_tf_policy
from tf_agents.replay_buffers import tf_uniform_replay_buffer
from tf_agents.trajectories import trajectory
from tf_agents.utils import common
tf.compat.v1.enable_v2_behavior()
# Set up a virtual display for rendering OpenAI gym environments.
display = pyvirtualdisplay.Display(visible=0, size=(1400, 900)).start()
Hyperparameters
env_name = "CartPole-v1" # @param {type:"string"}
num_iterations = 15000 # @param {type:"integer"}
initial_collect_steps = 1000 # @param {type:"integer"}
collect_steps_per_iteration = 1 # @param {type:"integer"}
replay_buffer_capacity = 100000 # @param {type:"integer"}
fc_layer_params = (100,)
batch_size = 64 # @param {type:"integer"}
learning_rate = 1e-3 # @param {type:"number"}
gamma = 0.99
log_interval = 200 # @param {type:"integer"}
num_atoms = 51 # @param {type:"integer"}
min_q_value = -20 # @param {type:"integer"}
max_q_value = 20 # @param {type:"integer"}
n_step_update = 2 # @param {type:"integer"}
num_eval_episodes = 10 # @param {type:"integer"}
eval_interval = 1000 # @param {type:"integer"}
Environment
Load the environment as before, with one for training and one for evaluation. Here we use CartPole-v1 (vs. CartPole-v0 in the DQN tutorial), which has a larger max reward of 500 rather than 200.
train_py_env = suite_gym.load(env_name)
eval_py_env = suite_gym.load(env_name)
train_env = tf_py_environment.TFPyEnvironment(train_py_env)
eval_env = tf_py_environment.TFPyEnvironment(eval_py_env)
Agent
C51 is a Q-learning algorithm based on DQN. Like DQN, it can be used on any environment with a discrete action space.
The main difference between C51 and DQN is that rather than simply predicting the Q-value for each state-action pair, C51 predicts a histogram model for the probability distribution of the Q-value:
By learning the distribution rather than simply the expected value, the algorithm is able to stay more stable during training, leading to improved final performance. This is particularly true in situations with bimodal or even multimodal value distributions, where a single average does not provide an accurate picture.
In order to train on probability distributions rather than on values, C51 must perform some complex distributional computations in order to calculate its loss function. But don't worry, all of this is taken care of for you in TF-Agents!
To create a C51 Agent, we first need to create a CategoricalQNetwork. The API of the CategoricalQNetwork is the same as that of the QNetwork, except that there is an additional argument num_atoms. This represents the number of support points in our probability distribution estimates. (The above image includes 10 support points, each represented by a vertical blue bar.) As you can tell from the name, the default number of atoms is 51.
categorical_q_net = categorical_q_network.CategoricalQNetwork(
train_env.observation_spec(),
train_env.action_spec(),
num_atoms=num_atoms,
fc_layer_params=fc_layer_params)
We also need an optimizer to train the network we just created, and a train_step_counter variable to keep track of how many times the network was updated.
Note that one other significant difference from vanilla DqnAgent is that we now need to specify min_q_value and max_q_value as arguments. These specify the most extreme values of the support (in other words, the most extreme of the 51 atoms on either side). Make sure to choose these appropriately for your particular environment. Here we use -20 and 20.
optimizer = tf.compat.v1.train.AdamOptimizer(learning_rate=learning_rate)
train_step_counter = tf.compat.v2.Variable(0)
agent = categorical_dqn_agent.CategoricalDqnAgent(
train_env.time_step_spec(),
train_env.action_spec(),
categorical_q_network=categorical_q_net,
optimizer=optimizer,
min_q_value=min_q_value,
max_q_value=max_q_value,
n_step_update=n_step_update,
td_errors_loss_fn=common.element_wise_squared_loss,
gamma=gamma,
train_step_counter=train_step_counter)
agent.initialize()
One last thing to note is that we also added an argument to use n-step updates with $n$ = 2. In single-step Q-learning ($n$ = 1), we only compute the error between the Q-values at the current time step and the next time step using the single-step return (based on the Bellman optimality equation). The single-step return is defined as:
$G_t = R_{t + 1} + \gamma V(s_{t + 1})$
where we define $V(s) = \max_a{Q(s, a)}$.
N-step updates involve expanding the standard single-step return function $n$ times:
$G_t^n = R_{t + 1} + \gamma R_{t + 2} + \gamma^2 R_{t + 3} + \dots + \gamma^n V(s_{t + n})$
N-step updates enable the agent to bootstrap from further in the future, and with the right value of $n$, this often leads to faster learning.
Although C51 and n-step updates are often combined with prioritized replay to form the core of the Rainbow agent, we saw no measurable improvement from implementing prioritized replay. Moreover, we find that when combining our C51 agent with n-step updates alone, our agent performs as well as other Rainbow agents on the sample of Atari environments we've tested.
Metrics and Evaluation
The most common metric used to evaluate a policy is the average return. The return is the sum of rewards obtained while running a policy in an environment for an episode, and we usually average this over a few episodes. We can compute the average return metric as follows.
def compute_avg_return(environment, policy, num_episodes=10):
total_return = 0.0
for _ in range(num_episodes):
time_step = environment.reset()
episode_return = 0.0
while not time_step.is_last():
action_step = policy.action(time_step)
time_step = environment.step(action_step.action)
episode_return += time_step.reward
total_return += episode_return
avg_return = total_return / num_episodes
return avg_return.numpy()[0]
random_policy = random_tf_policy.RandomTFPolicy(train_env.time_step_spec(),
train_env.action_spec())
compute_avg_return(eval_env, random_policy, num_eval_episodes)
# Please also see the metrics module for standard implementations of different
# metrics.
21.1
Data Collection
As in the DQN tutorial, set up the replay buffer and the initial data collection with the random policy.
replay_buffer = tf_uniform_replay_buffer.TFUniformReplayBuffer(
data_spec=agent.collect_data_spec,
batch_size=train_env.batch_size,
max_length=replay_buffer_capacity)
def collect_step(environment, policy):
time_step = environment.current_time_step()
action_step = policy.action(time_step)
next_time_step = environment.step(action_step.action)
traj = trajectory.from_transition(time_step, action_step, next_time_step)
# Add trajectory to the replay buffer
replay_buffer.add_batch(traj)
for _ in range(initial_collect_steps):
collect_step(train_env, random_policy)
# This loop is so common in RL, that we provide standard implementations of
# these. For more details see the drivers module.
# Dataset generates trajectories with shape [BxTx...] where
# T = n_step_update + 1.
dataset = replay_buffer.as_dataset(
num_parallel_calls=3, sample_batch_size=batch_size,
num_steps=n_step_update + 1).prefetch(3)
iterator = iter(dataset)
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.7/site-packages/tensorflow/python/autograph/impl/api.py:382: ReplayBuffer.get_next (from tf_agents.replay_buffers.replay_buffer) is deprecated and will be removed in a future version. Instructions for updating: Use `as_dataset(..., single_deterministic_pass=False) instead.
Training the agent
The training loop involves both collecting data from the environment and optimizing the agent's networks. Along the way, we will occasionally evaluate the agent's policy to see how we are doing.
The following will take ~7 minutes to run.
try:
%%time
except:
pass
# (Optional) Optimize by wrapping some of the code in a graph using TF function.
agent.train = common.function(agent.train)
# Reset the train step
agent.train_step_counter.assign(0)
# Evaluate the agent's policy once before training.
avg_return = compute_avg_return(eval_env, agent.policy, num_eval_episodes)
returns = [avg_return]
for _ in range(num_iterations):
# Collect a few steps using collect_policy and save to the replay buffer.
for _ in range(collect_steps_per_iteration):
collect_step(train_env, agent.collect_policy)
# Sample a batch of data from the buffer and update the agent's network.
experience, unused_info = next(iterator)
train_loss = agent.train(experience)
step = agent.train_step_counter.numpy()
if step % log_interval == 0:
print('step = {0}: loss = {1}'.format(step, train_loss.loss))
if step % eval_interval == 0:
avg_return = compute_avg_return(eval_env, agent.policy, num_eval_episodes)
print('step = {0}: Average Return = {1:.2f}'.format(step, avg_return))
returns.append(avg_return)
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.7/site-packages/tensorflow/python/util/dispatch.py:206: calling foldr_v2 (from tensorflow.python.ops.functional_ops) with back_prop=False is deprecated and will be removed in a future version. Instructions for updating: back_prop=False is deprecated. Consider using tf.stop_gradient instead. Instead of: results = tf.foldr(fn, elems, back_prop=False) Use: results = tf.nest.map_structure(tf.stop_gradient, tf.foldr(fn, elems)) step = 200: loss = 3.073254108428955 step = 400: loss = 1.7349590063095093 step = 600: loss = 1.792757272720337 step = 800: loss = 1.8498374223709106 step = 1000: loss = 1.5893661975860596 step = 1000: Average Return = 85.80 step = 1200: loss = 1.5345051288604736 step = 1400: loss = 1.0927748680114746 step = 1600: loss = 1.3448643684387207 step = 1800: loss = 1.3334734439849854 step = 2000: loss = 1.2656627893447876 step = 2000: Average Return = 104.50 step = 2200: loss = 0.9804798364639282 step = 2400: loss = 1.0801945924758911 step = 2600: loss = 0.8012468814849854 step = 2800: loss = 1.2241151332855225 step = 3000: loss = 1.101938009262085 step = 3000: Average Return = 75.00 step = 3200: loss = 1.1817928552627563 step = 3400: loss = 0.9651275873184204 step = 3600: loss = 1.1266820430755615 step = 3800: loss = 0.840592622756958 step = 4000: loss = 0.9282000064849854 step = 4000: Average Return = 129.00 step = 4200: loss = 1.0468955039978027 step = 4400: loss = 1.0415769815444946 step = 4600: loss = 1.0660417079925537 step = 4800: loss = 0.9771052002906799 step = 5000: loss = 0.8960349559783936 step = 5000: Average Return = 126.60 step = 5200: loss = 0.925206184387207 step = 5400: loss = 0.6512923240661621 step = 5600: loss = 0.7966318130493164 step = 5800: loss = 0.7415628433227539 step = 6000: loss = 0.897962212562561 step = 6000: Average Return = 109.10 step = 6200: loss = 0.8640482425689697 step = 6400: loss = 1.038527250289917 step = 6600: loss = 1.017322063446045 step = 6800: loss = 0.7789400815963745 step = 7000: loss = 0.9924593567848206 step = 7000: Average Return = 122.80 step = 7200: loss = 0.5298917293548584 step = 7400: loss = 0.5959913730621338 step = 7600: loss = 0.6432434320449829 step = 7800: loss = 0.6911174058914185 step = 8000: loss = 0.5797613263130188 step = 8000: Average Return = 144.50 step = 8200: loss = 0.5593736171722412 step = 8400: loss = 0.6327507495880127 step = 8600: loss = 0.5596221089363098 step = 8800: loss = 0.6492353677749634 step = 9000: loss = 0.6443781852722168 step = 9000: Average Return = 135.80 step = 9200: loss = 0.8031193614006042 step = 9400: loss = 0.541293740272522 step = 9600: loss = 0.643537700176239 step = 9800: loss = 0.7450239062309265 step = 10000: loss = 0.5937005281448364 step = 10000: Average Return = 202.80 step = 10200: loss = 0.4400593042373657 step = 10400: loss = 0.7001190185546875 step = 10600: loss = 0.56056809425354 step = 10800: loss = 0.7413918375968933 step = 11000: loss = 0.6136442422866821 step = 11000: Average Return = 126.10 step = 11200: loss = 0.6289737224578857 step = 11400: loss = 0.6546649932861328 step = 11600: loss = 0.4087551534175873 step = 11800: loss = 0.6893659830093384 step = 12000: loss = 0.5798210501670837 step = 12000: Average Return = 152.90 step = 12200: loss = 0.4694007635116577 step = 12400: loss = 0.6825675368309021 step = 12600: loss = 0.675622284412384 step = 12800: loss = 0.6773465275764465 step = 13000: loss = 0.35596776008605957 step = 13000: Average Return = 141.70 step = 13200: loss = 0.6894246339797974 step = 13400: loss = 0.7381303310394287 step = 13600: loss = 0.675399661064148 step = 13800: loss = 0.6918181777000427 step = 14000: loss = 0.5304032564163208 step = 14000: Average Return = 146.10 step = 14200: loss = 0.6360524892807007 step = 14400: loss = 0.5732524394989014 step = 14600: loss = 0.4850385785102844 step = 14800: loss = 0.37867382168769836 step = 15000: loss = 0.46798139810562134 step = 15000: Average Return = 154.10
Visualization
Plots
We can plot return vs global steps to see the performance of our agent. In Cartpole-v1, the environment gives a reward of +1 for every time step the pole stays up, and since the maximum number of steps is 500, the maximum possible return is also 500.
steps = range(0, num_iterations + 1, eval_interval)
plt.plot(steps, returns)
plt.ylabel('Average Return')
plt.xlabel('Step')
plt.ylim(top=550)
(6.240000247955322, 550.0)
Videos
It is helpful to visualize the performance of an agent by rendering the environment at each step. Before we do that, let us first create a function to embed videos in this colab.
def embed_mp4(filename):
"""Embeds an mp4 file in the notebook."""
video = open(filename,'rb').read()
b64 = base64.b64encode(video)
tag = '''
<video width="640" height="480" controls>
<source src="data:video/mp4;base64,{0}" type="video/mp4">
Your browser does not support the video tag.
</video>'''.format(b64.decode())
return IPython.display.HTML(tag)
The following code visualizes the agent's policy for a few episodes:
num_episodes = 3
video_filename = 'imageio.mp4'
with imageio.get_writer(video_filename, fps=60) as video:
for _ in range(num_episodes):
time_step = eval_env.reset()
video.append_data(eval_py_env.render())
while not time_step.is_last():
action_step = agent.policy.action(time_step)
time_step = eval_env.step(action_step.action)
video.append_data(eval_py_env.render())
embed_mp4(video_filename)
WARNING:root:IMAGEIO FFMPEG_WRITER WARNING: input image is not divisible by macro_block_size=16, resizing from (400, 600) to (400, 608) to ensure video compatibility with most codecs and players. To prevent resizing, make your input image divisible by the macro_block_size or set the macro_block_size to None (risking incompatibility). You may also see a FFMPEG warning concerning speedloss due to data not being aligned.
C51 tends to do slightly better than DQN on CartPole-v1, but the difference between the two agents becomes more and more significant in increasingly complex environments. For example, on the full Atari 2600 benchmark, C51 demonstrates a mean score improvement of 126% over DQN after normalizing with respect to a random agent. Additional improvements can be gained by including n-step updates.
For a deeper dive into the C51 algorithm, see A Distributional Perspective on Reinforcement Learning (2017).