# Register custom envs
import utils.import_envs # noqa: F401 pytype: disable=import-error
from utils.exp_manager import ExperimentManager
from utils.utils import ALGOS, StoreDict
seaborn.set()
if __name__ == "__main__": # noqa: C901
parser = argparse.ArgumentParser()
parser.add_argument(
"--algo",
help="RL Algorithm",
default="ppo",
type=str,
required=False,
choices=list(ALGOS.keys()),
)
parser.add_argument("--env", type=str, default="CartPole-v1", help="environment ID")
parser.add_argument(
"-tb", "--tensorboard-log", help="Tensorboard log dir", default="", type=str
)
parser.add_argument(
"-i",
"--trained-agent",
help="Path to a pretrained agent to continue training",
default="",
type=str,
)
parser.add_argument(
"--truncate-last-trajectory",
help="When using HER with online sampling the last trajectory "
import gym_fishing
import numpy as np
import seaborn
import torch as th
from stable_baselines3.common.utils import set_random_seed
# Register custom envs
import utils.import_envs # noqa: F401 pytype: disable=import-error
from utils.exp_manager import ExperimentManager
from utils.utils import ALGOS, StoreDict
seaborn.set()
if __name__ == "__main__": # noqa: C901
parser = argparse.ArgumentParser()
parser.add_argument("--algo", help="RL Algorithm", default="ppo", type=str, required=False, choices=list(ALGOS.keys()))
parser.add_argument("--env", type=str, default="fishing-v1", help="environment ID")
parser.add_argument("-tb", "--tensorboard-log", help="Tensorboard log dir", default="", type=str)
parser.add_argument("-i", "--trained-agent", help="Path to a pretrained agent to continue training", default="", type=str)
parser.add_argument(
"--truncate-last-trajectory",
help="When using HER with online sampling the last trajectory "
"in the replay buffer will be truncated after reloading the replay buffer.",
default=True,
type=bool,
)
parser.add_argument("-n", "--n-timesteps", help="Overwrite the number of timesteps", default=-1, type=int)
parser.add_argument("--num-threads", help="Number of threads for PyTorch (-1 to use default)", default=-1, type=int)
parser.add_argument("--log-interval", help="Override log interval (default: -1, no change)", default=-1, type=int)
parser.add_argument(
"--eval-freq", help="Evaluate the agent every n steps (if negative, no evaluation)", default=10000, type=int
from stable_baselines3.common.utils import set_random_seed
from utils.exp_manager import ExperimentManager
from utils.utils import ALGOS, StoreDict
seaborn.set()
if __name__ == "__main__": # noqa: C901
parser = argparse.ArgumentParser()
parser.add_argument(
"--algo",
help="RL Algorithm",
default="ppo",
type=str,
required=False,
choices=list(
ALGOS.keys()))
parser.add_argument(
"--env",
type=str,
default="CartPole-v1",
help="environment ID")
parser.add_argument(
"-tb",
"--tensorboard-log",
help="Tensorboard log dir",
default="",
type=str)
parser.add_argument(
"-i",
"--trained-agent",
help="Path to a pretrained agent to continue training",
import argparse
import os
import time
import gym
import numpy as np
from stable_baselines.common import set_global_seeds
from config import ENV_ID
from utils.utils import ALGOS, create_test_env, get_latest_run_id, get_saved_hyperparams
parser = argparse.ArgumentParser()
parser.add_argument('-f', '--folder', help='Log folder', type=str, default='logs')
parser.add_argument('--algo', help='RL Algorithm', default='sac',
type=str, required=False, choices=list(ALGOS.keys()))
parser.add_argument('-n', '--n-timesteps', help='number of timesteps', default=1000,
type=int)
parser.add_argument('--exp-id', help='Experiment ID (-1: no exp folder, 0: latest)', default=0,
type=int)
parser.add_argument('--verbose', help='Verbose mode (0: no output, 1: INFO)', default=1,
type=int)
parser.add_argument('--no-render', action='store_true', default=False,
help='Do not render the environment (useful for tests)')
parser.add_argument('--deterministic', action='store_true', default=False,
help='Use deterministic actions')
parser.add_argument('--norm-reward', action='store_true', default=False,
help='Normalize reward if applicable (trained with VecNormalize)')
parser.add_argument('--seed', help='Random generator seed', type=int, default=0)
parser.add_argument('--reward-log', help='Where to log reward', default='', type=str)
parser.add_argument('-vae', '--vae-path', help='Path to saved VAE', type=str, default='')