train_drqn_ale.py

"""An example of training a Deep Recurrent Q-Network (DRQN).

DRQN is a DQN with a recurrent Q-network, described in
https://arxiv.org/abs/1507.06527.

To train DRQN for 50M timesteps on Breakout, run:
    python train_drqn_ale.py --recurrent

To train DQRN using a recurrent model on flickering 1-frame Breakout, run:
    python train_drqn_ale.py --recurrent --flicker --no-frame-stack
"""
import argparse

import gym
import gym.wrappers
import numpy as np
import torch
from torch import nn

import pfrl
from pfrl import experiments
from pfrl import explorers
from pfrl import utils
from pfrl import replay_buffers
from pfrl.q_functions import DiscreteActionValueHead

from pfrl.wrappers import atari_wrappers


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--env",
        type=str,
        default="BreakoutNoFrameskip-v4",
        help="OpenAI Atari domain to perform algorithm on.",
    )
    parser.add_argument(
        "--outdir",
        type=str,
        default="results",
        help=(
            "Directory path to save output files."
            " If it does not exist, it will be created."
        ),
    )
    parser.add_argument("--seed", type=int, default=0, help="Random seed [0, 2 ** 31)")
    parser.add_argument(
        "--gpu", type=int, default=0, help="GPU to use, set to -1 if no GPU."
    )
    parser.add_argument("--demo", action="store_true", default=False)
    parser.add_argument("--load", type=str, default=None)
    parser.add_argument(
        "--final-exploration-frames",
        type=int,
        default=10 ** 6,
        help="Timesteps after which we stop " + "annealing exploration rate",
    )
    parser.add_argument(
        "--final-epsilon",
        type=float,
        default=0.01,
        help="Final value of epsilon during training.",
    )
    parser.add_argument(
        "--eval-epsilon",
        type=float,
        default=0.001,
        help="Exploration epsilon used during eval episodes.",
    )
    parser.add_argument(
        "--steps",
        type=int,
        default=5 * 10 ** 7,
        help="Total number of timesteps to train the agent.",
    )
    parser.add_argument(
        "--max-frames",
        type=int,
        default=30 * 60 * 60,  # 30 minutes with 60 fps
        help="Maximum number of frames for each episode.",
    )
    parser.add_argument(
        "--replay-start-size",
        type=int,
        default=5 * 10 ** 4,
        help="Minimum replay buffer size before " + "performing gradient updates.",
    )
    parser.add_argument(
        "--target-update-interval",
        type=int,
        default=3 * 10 ** 4,
        help="Frequency (in timesteps) at which " + "the target network is updated.",
    )
    parser.add_argument("--demo-n-episodes", type=int, default=30)
    parser.add_argument("--eval-n-steps", type=int, default=125000)
    parser.add_argument(
        "--eval-interval",
        type=int,
        default=250000,
        help="Frequency (in timesteps) of evaluation phase.",
    )
    parser.add_argument(
        "--update-interval",
        type=int,
        default=4,
        help="Frequency (in timesteps) of network updates.",
    )
    parser.add_argument(
        "--log-level",
        type=int,
        default=20,
        help="Logging level. 10:DEBUG, 20:INFO etc.",
    )
    parser.add_argument(
        "--render",
        action="store_true",
        default=False,
        help="Render env states in a GUI window.",
    )
    parser.add_argument(
        "--monitor",
        action="store_true",
        default=False,
        help=(
            "Monitor env. Videos and additional information are saved as output files."
        ),
    )
    parser.add_argument("--lr", type=float, default=2.5e-4, help="Learning rate.")
    parser.add_argument(
        "--recurrent",
        action="store_true",
        default=False,
        help="Use a recurrent model. See the code for the model definition.",
    )
    parser.add_argument(
        "--flicker",
        action="store_true",
        default=False,
        help=(
            "Use so-called flickering Atari, where each"
            " screen is blacked out with probability 0.5."
        ),
    )
    parser.add_argument(
        "--no-frame-stack",
        action="store_true",
        default=False,
        help=(
            "Disable frame stacking so that the agent can only see the current screen."
        ),
    )
    parser.add_argument(
        "--episodic-update-len",
        type=int,
        default=10,
        help="Maximum length of sequences for updating recurrent models",
    )
    parser.add_argument(
        "--batch-size",
        type=int,
        default=32,
        help=(
            "Number of transitions (in a non-recurrent case)"
            " or sequences (in a recurrent case) used for an"
            " update."
        ),
    )
    args = parser.parse_args()

    import logging

    logging.basicConfig(level=args.log_level)

    # Set a random seed used in PFRL.
    utils.set_random_seed(args.seed)

    # Set different random seeds for train and test envs.
    train_seed = args.seed
    test_seed = 2 ** 31 - 1 - args.seed

    args.outdir = experiments.prepare_output_dir(args, args.outdir)
    print("Output files are saved in {}".format(args.outdir))

    def make_env(test):
        # Use different random seeds for train and test envs
        env_seed = test_seed if test else train_seed
        env = atari_wrappers.wrap_deepmind(
            atari_wrappers.make_atari(args.env, max_frames=args.max_frames),
            episode_life=not test,
            clip_rewards=not test,
            flicker=args.flicker,
            frame_stack=not args.no_frame_stack,
        )
        env.seed(int(env_seed))
        if test:
            # Randomize actions like epsilon-greedy in evaluation as well
            env = pfrl.wrappers.RandomizeAction(env, args.eval_epsilon)
        if args.monitor:
            env = gym.wrappers.Monitor(
                env, args.outdir, mode="evaluation" if test else "training"
            )
        if args.render:
            env = pfrl.wrappers.Render(env)
        return env

    env = make_env(test=False)
    eval_env = make_env(test=True)
    print("Observation space", env.observation_space)
    print("Action space", env.action_space)

    n_frames = env.observation_space.shape[0]
    n_actions = env.action_space.n
    if args.recurrent:
        # Q-network with LSTM
        q_func = pfrl.nn.RecurrentSequential(
            nn.Conv2d(n_frames, 32, 8, stride=4),
            nn.ReLU(),
            nn.Conv2d(32, 64, 4, stride=2),
            nn.ReLU(),
            nn.Conv2d(64, 64, 3, stride=1),
            nn.Flatten(),
            nn.ReLU(),
            nn.LSTM(input_size=3136, hidden_size=512),
            nn.Linear(512, n_actions),
            DiscreteActionValueHead(),
        )
        # Replay buffer that stores whole episodes
        rbuf = replay_buffers.EpisodicReplayBuffer(10 ** 6)
    else:
        # Q-network without LSTM
        q_func = nn.Sequential(
            nn.Conv2d(n_frames, 32, 8, stride=4),
            nn.ReLU(),
            nn.Conv2d(32, 64, 4, stride=2),
            nn.ReLU(),
            nn.Conv2d(64, 64, 3, stride=1),
            nn.Flatten(),
            nn.Linear(3136, 512),
            nn.ReLU(),
            nn.Linear(512, n_actions),
            DiscreteActionValueHead(),
        )
        # Replay buffer that stores transitions separately
        rbuf = replay_buffers.ReplayBuffer(10 ** 6)

    explorer = explorers.LinearDecayEpsilonGreedy(
        1.0,
        args.final_epsilon,
        args.final_exploration_frames,
        lambda: np.random.randint(n_actions),
    )

    opt = torch.optim.Adam(q_func.parameters(), lr=1e-4, eps=1e-4)

    def phi(x):
        # Feature extractor
        return np.asarray(x, dtype=np.float32) / 255

    agent = pfrl.agents.DoubleDQN(
        q_func,
        opt,
        rbuf,
        gpu=args.gpu,
        gamma=0.99,
        explorer=explorer,
        replay_start_size=args.replay_start_size,
        target_update_interval=args.target_update_interval,
        update_interval=args.update_interval,
        batch_accumulator="mean",
        phi=phi,
        minibatch_size=args.batch_size,
        episodic_update_len=args.episodic_update_len,
        recurrent=args.recurrent,
    )

    if args.load:
        agent.load(args.load)

    if args.demo:
        eval_stats = experiments.eval_performance(
            env=eval_env, agent=agent, n_steps=None, n_episodes=args.demo_n_episodes,
        )
        print(
            "n_runs: {} mean: {} median: {} stdev {}".format(
                args.demo_n_episodes,
                eval_stats["mean"],
                eval_stats["median"],
                eval_stats["stdev"],
            )
        )
    else:
        experiments.train_agent_with_evaluation(
            agent=agent,
            env=env,
            steps=args.steps,
            eval_n_steps=args.eval_n_steps,
            eval_n_episodes=None,
            eval_interval=args.eval_interval,
            outdir=args.outdir,
            eval_env=eval_env,
        )


if __name__ == "__main__":
    main()