Python logger.set_level方法代码示例

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def train(env_id, num_timesteps, seed):
    import baselines.common.tf_util as U
    sess = U.single_threaded_session()
    sess.__enter__()

    rank = MPI.COMM_WORLD.Get_rank()
    if rank == 0:
        logger.configure()
    else:
        logger.configure(format_strs=[])
        logger.set_level(logger.DISABLED)
    workerseed = seed + 10000 * MPI.COMM_WORLD.Get_rank()
    def policy_fn(name, ob_space, ac_space):
        return MlpPolicy(name=name, ob_space=ob_space, ac_space=ac_space,
            hid_size=32, num_hid_layers=2)
    env = make_mujoco_env(env_id, workerseed)
    trpo_mpi.learn(env, policy_fn, timesteps_per_batch=1024, max_kl=0.01, cg_iters=10, cg_damping=0.1,
        max_timesteps=num_timesteps, gamma=0.99, lam=0.98, vf_iters=5, vf_stepsize=1e-3)
    env.close() 

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def train(env_id, num_timesteps, seed):
    whoami  = mpi_fork(num_cpu)
    if whoami == "parent":
        return
    import baselines.common.tf_util as U
    logger.session().__enter__()
    sess = U.single_threaded_session()
    sess.__enter__()

    rank = MPI.COMM_WORLD.Get_rank()
    if rank != 0:
        logger.set_level(logger.DISABLED)
    workerseed = seed + 10000 * MPI.COMM_WORLD.Get_rank()
    set_global_seeds(workerseed)
    env = gym.make(env_id)
    def policy_fn(name, ob_space, ac_space):
        return MlpPolicy(name=name, ob_space=env.observation_space, ac_space=env.action_space,
            hid_size=32, num_hid_layers=2)
    env = bench.Monitor(env, osp.join(logger.get_dir(), "%i.monitor.json" % rank))
    env.seed(workerseed)
    gym.logger.setLevel(logging.WARN)

    trpo_mpi.learn(env, policy_fn, timesteps_per_batch=1024, max_kl=0.01, cg_iters=10, cg_damping=0.1,
        max_timesteps=num_timesteps, gamma=0.99, lam=0.98, vf_iters=5, vf_stepsize=1e-3)
    env.close() 

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def train(env_id, num_timesteps, seed):
    import baselines.common.tf_util as U
    sess = U.single_threaded_session()
    sess.__enter__()

    rank = MPI.COMM_WORLD.Get_rank()
    if rank != 0:
        logger.set_level(logger.DISABLED)
    workerseed = seed + 10000 * MPI.COMM_WORLD.Get_rank()
    set_global_seeds(workerseed)
    env = gym.make(env_id)
    def policy_fn(name, ob_space, ac_space):
        return MlpPolicy(name=name, ob_space=env.observation_space, ac_space=env.action_space,
            hid_size=32, num_hid_layers=2)
    env = bench.Monitor(env, logger.get_dir() and
        osp.join(logger.get_dir(), str(rank)))
    env.seed(workerseed)
    gym.logger.setLevel(logging.WARN)

    trpo_mpi.learn(env, policy_fn, timesteps_per_batch=1024, max_kl=0.01, cg_iters=10, cg_damping=0.1,
        max_timesteps=num_timesteps, gamma=0.99, lam=0.98, vf_iters=5, vf_stepsize=1e-3)
    env.close() 

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def train(env, seed, policy_fn, reward_giver, dataset, algo,
          g_step, d_step, policy_entcoeff, num_timesteps, save_per_iter,
          checkpoint_dir, log_dir, pretrained, BC_max_iter, task_name=None):

    pretrained_weight = None
    if pretrained and (BC_max_iter > 0):
        # Pretrain with behavior cloning
        from baselines.gail import behavior_clone
        pretrained_weight = behavior_clone.learn(env, policy_fn, dataset,
                                                 max_iters=BC_max_iter)

    if algo == 'trpo':
        from baselines.gail import trpo_mpi
        # Set up for MPI seed
        rank = MPI.COMM_WORLD.Get_rank()
        if rank != 0:
            logger.set_level(logger.DISABLED)
        workerseed = seed + 10000 * MPI.COMM_WORLD.Get_rank()
        set_global_seeds(workerseed)
        env.seed(workerseed)
        trpo_mpi.learn(env, policy_fn, reward_giver, dataset, rank,
                       pretrained=pretrained, pretrained_weight=pretrained_weight,
                       g_step=g_step, d_step=d_step,
                       entcoeff=policy_entcoeff,
                       max_timesteps=num_timesteps,
                       ckpt_dir=checkpoint_dir, log_dir=log_dir,
                       save_per_iter=save_per_iter,
                       timesteps_per_batch=1024,
                       max_kl=0.01, cg_iters=10, cg_damping=0.1,
                       gamma=0.995, lam=0.97,
                       vf_iters=5, vf_stepsize=1e-3,
                       task_name=task_name)
    else:
        raise NotImplementedError 

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def train(env_id, num_timesteps, seed, num_cpu):
    from baselines.trpo_mpi.nosharing_cnn_policy import CnnPolicy
    from baselines.trpo_mpi import trpo_mpi
    import baselines.common.tf_util as U
    whoami  = mpi_fork(num_cpu)
    if whoami == "parent":
        return
    rank = MPI.COMM_WORLD.Get_rank()
    sess = U.single_threaded_session()
    sess.__enter__()
    logger.session().__enter__()
    if rank != 0:
        logger.set_level(logger.DISABLED)


    workerseed = seed + 10000 * MPI.COMM_WORLD.Get_rank()
    set_global_seeds(workerseed)
    env = gym.make(env_id)
    def policy_fn(name, ob_space, ac_space): #pylint: disable=W0613
        return CnnPolicy(name=name, ob_space=env.observation_space, ac_space=env.action_space)
    env = bench.Monitor(env, osp.join(logger.get_dir(), "%i.monitor.json"%rank))
    env.seed(workerseed)
    gym.logger.setLevel(logging.WARN)

    env = wrap_train(env)
    num_timesteps /= 4 # because we're wrapping the envs to do frame skip
    env.seed(workerseed)

    trpo_mpi.learn(env, policy_fn, timesteps_per_batch=512, max_kl=0.001, cg_iters=10, cg_damping=1e-3,
        max_timesteps=num_timesteps, gamma=0.98, lam=1.0, vf_iters=3, vf_stepsize=1e-4, entcoeff=0.00)
    env.close() 

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def train(env_id, num_timesteps, seed, num_cpu):
    from baselines.pposgd import pposgd_simple, cnn_policy
    import baselines.common.tf_util as U
    whoami  = mpi_fork(num_cpu)
    if whoami == "parent": return
    rank = MPI.COMM_WORLD.Get_rank()
    sess = U.single_threaded_session()
    sess.__enter__()
    logger.session().__enter__()
    if rank != 0: logger.set_level(logger.DISABLED)
    workerseed = seed + 10000 * MPI.COMM_WORLD.Get_rank()
    set_global_seeds(workerseed)
    env = gym.make(env_id)
    def policy_fn(name, ob_space, ac_space): #pylint: disable=W0613
        return cnn_policy.CnnPolicy(name=name, ob_space=ob_space, ac_space=ac_space)
    env = bench.Monitor(env, osp.join(logger.get_dir(), "%i.monitor.json" % rank))
    env.seed(workerseed)
    gym.logger.setLevel(logging.WARN)

    env = wrap_train(env)
    num_timesteps /= 4 # because we're wrapping the envs to do frame skip
    env.seed(workerseed)

    pposgd_simple.learn(env, policy_fn,
        max_timesteps=num_timesteps,
        timesteps_per_batch=256,
        clip_param=0.2, entcoeff=0.01,
        optim_epochs=4, optim_stepsize=1e-3, optim_batchsize=64,
        gamma=0.99, lam=0.95,
        schedule='linear'
    )
    env.close() 

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def train(env_id, num_timesteps, seed):
    import baselines.common.tf_util as U
    sess = U.single_threaded_session()
    sess.__enter__()
    workerseed = seed + 10000 * MPI.COMM_WORLD.Get_rank()
    def policy_fn(name, ob_space, ac_space):
        return MlpPolicy(name=name, ob_space=ob_space, ac_space=ac_space,
            hid_size=32, num_hid_layers=2)

    # Create a new base directory like /tmp/openai-2018-05-21-12-27-22-552435
    log_dir = os.path.join(energyplus_logbase_dir(), datetime.datetime.now().strftime("openai-%Y-%m-%d-%H-%M-%S-%f"))
    if not os.path.exists(log_dir + '/output'):
        os.makedirs(log_dir + '/output')
    os.environ["ENERGYPLUS_LOG"] = log_dir
    model = os.getenv('ENERGYPLUS_MODEL')
    if model is None:
        print('Environment variable ENERGYPLUS_MODEL is not defined')
        os.exit()
    weather = os.getenv('ENERGYPLUS_WEATHER')
    if weather is None:
        print('Environment variable ENERGYPLUS_WEATHER is not defined')
        os.exit()

    rank = MPI.COMM_WORLD.Get_rank()
    if rank == 0:
        print('train: init logger with dir={}'.format(log_dir)) #XXX
        logger.configure(log_dir)
    else:
        logger.configure(format_strs=[])
        logger.set_level(logger.DISABLED)

    env = make_energyplus_env(env_id, workerseed)

    trpo_mpi.learn(env, policy_fn,
                   max_timesteps=num_timesteps,
                   #timesteps_per_batch=1*1024, max_kl=0.01, cg_iters=10, cg_damping=0.1,
                   timesteps_per_batch=16*1024, max_kl=0.01, cg_iters=10, cg_damping=0.1,
                   gamma=0.99, lam=0.98, vf_iters=5, vf_stepsize=1e-3)
    env.close() 

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def run(env_id, seed, noise_type, layer_norm, evaluation, **kwargs):
    # Configure things.
    rank = MPI.COMM_WORLD.Get_rank()
    if rank != 0:
        logger.set_level(logger.DISABLED)

    # Create envs.
    env = gym.make(env_id)
    env = bench.Monitor(env, logger.get_dir() and os.path.join(logger.get_dir(), str(rank)))

    if evaluation and rank==0:
        eval_env = gym.make(env_id)
        eval_env = bench.Monitor(eval_env, os.path.join(logger.get_dir(), 'gym_eval'))
        env = bench.Monitor(env, None)
    else:
        eval_env = None

    # Parse noise_type
    action_noise = None
    param_noise = None
    nb_actions = env.action_space.shape[-1]
    for current_noise_type in noise_type.split(','):
        current_noise_type = current_noise_type.strip()
        if current_noise_type == 'none':
            pass
        elif 'adaptive-param' in current_noise_type:
            _, stddev = current_noise_type.split('_')
            param_noise = AdaptiveParamNoiseSpec(initial_stddev=float(stddev), desired_action_stddev=float(stddev))
        elif 'normal' in current_noise_type:
            _, stddev = current_noise_type.split('_')
            action_noise = NormalActionNoise(mu=np.zeros(nb_actions), sigma=float(stddev) * np.ones(nb_actions))
        elif 'ou' in current_noise_type:
            _, stddev = current_noise_type.split('_')
            action_noise = OrnsteinUhlenbeckActionNoise(mu=np.zeros(nb_actions), sigma=float(stddev) * np.ones(nb_actions))
        else:
            raise RuntimeError('unknown noise type "{}"'.format(current_noise_type))

    # Configure components.
    memory = Memory(limit=int(1e6), action_shape=env.action_space.shape, observation_shape=env.observation_space.shape)
    critic = Critic(layer_norm=layer_norm)
    actor = Actor(nb_actions, layer_norm=layer_norm)

    # Seed everything to make things reproducible.
    seed = seed + 1000000 * rank
    logger.info('rank {}: seed={}, logdir={}'.format(rank, seed, logger.get_dir()))
    tf.reset_default_graph()
    set_global_seeds(seed)
    env.seed(seed)
    if eval_env is not None:
        eval_env.seed(seed)

    # Disable logging for rank != 0 to avoid noise.
    if rank == 0:
        start_time = time.time()
    training.train(env=env, eval_env=eval_env, param_noise=param_noise,
        action_noise=action_noise, actor=actor, critic=critic, memory=memory, **kwargs)
    env.close()
    if eval_env is not None:
        eval_env.close()
    if rank == 0:
        logger.info('total runtime: {}s'.format(time.time() - start_time)) 

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def run(seed, noise_type, layer_norm, evaluation, **kwargs):
    # Configure things.
    rank = MPI.COMM_WORLD.Get_rank()
    if rank != 0:
        logger.set_level(logger.DISABLED)

    # Create envs.
    env = gymify_osim_env(Arm3dEnv(visualize = True))
    env = bench.Monitor(env, logger.get_dir() and os.path.join(logger.get_dir(), str(rank)))

    if evaluation and rank==0:
        eval_env = gymify_osim_env(Arm3dEnv(visualize = True))
        eval_env = bench.Monitor(eval_env, os.path.join(logger.get_dir(), 'gym_eval'))
        env = bench.Monitor(env, None)
    else:
        eval_env = None

    # Parse noise_type
    action_noise = None
    param_noise = None
    nb_actions = env.action_space.shape[-1]
    for current_noise_type in noise_type.split(','):
        current_noise_type = current_noise_type.strip()
        if current_noise_type == 'none':
            pass
        elif 'adaptive-param' in current_noise_type:
            _, stddev = current_noise_type.split('_')
            param_noise = AdaptiveParamNoiseSpec(initial_stddev=float(stddev), desired_action_stddev=float(stddev))
        elif 'normal' in current_noise_type:
            _, stddev = current_noise_type.split('_')
            action_noise = NormalActionNoise(mu=np.zeros(nb_actions), sigma=float(stddev) * np.ones(nb_actions))
        elif 'ou' in current_noise_type:
            _, stddev = current_noise_type.split('_')
            action_noise = OrnsteinUhlenbeckActionNoise(mu=np.zeros(nb_actions), sigma=float(stddev) * np.ones(nb_actions))
        else:
            raise RuntimeError('unknown noise type "{}"'.format(current_noise_type))

    # Configure components.
    memory = Memory(limit=int(1e6), action_shape=env.action_space.shape, observation_shape=env.observation_space.shape)
    critic = Critic(layer_norm=layer_norm)
    actor = Actor(nb_actions, layer_norm=layer_norm)

    # Seed everything to make things reproducible.
    seed = seed + 1000000 * rank
    logger.info('rank {}: seed={}, logdir={}'.format(rank, seed, logger.get_dir()))
    tf.reset_default_graph()
    set_global_seeds(seed)
    env.seed(seed)
    if eval_env is not None:
        eval_env.seed(seed)

    # Disable logging for rank != 0 to avoid noise.
    if rank == 0:
        start_time = time.time()
    training.train(env=env, eval_env=eval_env, param_noise=param_noise,
        action_noise=action_noise, actor=actor, critic=critic, memory=memory, **kwargs)
    env.close()
    if eval_env is not None:
        eval_env.close()
    if rank == 0:
        logger.info('total runtime: {}s'.format(time.time() - start_time)) 

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def run(seed, noise_type, layer_norm, evaluation, **kwargs):
    # Configure things.
    rank = MPI.COMM_WORLD.Get_rank()
    if rank != 0:
        logger.set_level(logger.DISABLED)

    # Create envs.
    env = gymify_osim_env(L2RunEnv(visualize = True))
    env = bench.Monitor(env, logger.get_dir() and os.path.join(logger.get_dir(), str(rank)))

    if evaluation and rank==0:
        eval_env = gymify_osim_env(L2RunEnv(visualize = True))
        eval_env = bench.Monitor(eval_env, os.path.join(logger.get_dir(), 'gym_eval'))
        env = bench.Monitor(env, None)
    else:
        eval_env = None

    # Parse noise_type
    action_noise = None
    param_noise = None
    nb_actions = env.action_space.shape[-1]
    for current_noise_type in noise_type.split(','):
        current_noise_type = current_noise_type.strip()
        if current_noise_type == 'none':
            pass
        elif 'adaptive-param' in current_noise_type:
            _, stddev = current_noise_type.split('_')
            param_noise = AdaptiveParamNoiseSpec(initial_stddev=float(stddev), desired_action_stddev=float(stddev))
        elif 'normal' in current_noise_type:
            _, stddev = current_noise_type.split('_')
            action_noise = NormalActionNoise(mu=np.zeros(nb_actions), sigma=float(stddev) * np.ones(nb_actions))
        elif 'ou' in current_noise_type:
            _, stddev = current_noise_type.split('_')
            action_noise = OrnsteinUhlenbeckActionNoise(mu=np.zeros(nb_actions), sigma=float(stddev) * np.ones(nb_actions))
        else:
            raise RuntimeError('unknown noise type "{}"'.format(current_noise_type))

    # Configure components.
    memory = Memory(limit=int(1e6), action_shape=env.action_space.shape, observation_shape=env.observation_space.shape)
    critic = Critic(layer_norm=layer_norm)
    actor = Actor(nb_actions, layer_norm=layer_norm)

    # Seed everything to make things reproducible.
    seed = seed + 1000000 * rank
    logger.info('rank {}: seed={}, logdir={}'.format(rank, seed, logger.get_dir()))
    tf.reset_default_graph()
    set_global_seeds(seed)
    env.seed(seed)
    if eval_env is not None:
        eval_env.seed(seed)

    # Disable logging for rank != 0 to avoid noise.
    if rank == 0:
        start_time = time.time()
    training.train(env=env, eval_env=eval_env, param_noise=param_noise,
        action_noise=action_noise, actor=actor, critic=critic, memory=memory, **kwargs)
    env.close()
    if eval_env is not None:
        eval_env.close()
    if rank == 0:
        logger.info('total runtime: {}s'.format(time.time() - start_time)) 

# 需要导入模块: from baselines import logger [as 别名]
# 或者: from baselines.logger import set_level [as 别名]
def train(env, num_timesteps, seed, ckpt_dir=None,
          render=False, ckpt_freq=0, restore_dir=None, optim_stepsize=3e-4,
          schedule="linear", gamma=0.99, optim_epochs=10, optim_batchsize=64,
          horizon=2048):

    from baselines.common.fc_learning_utils import FlightLog
    from mpi4py import MPI
    from baselines import logger
    from baselines.ppo1.mlp_policy import MlpPolicy
    from baselines.common import set_global_seeds
    from baselines.ppo1 import pposgd_simple
    import baselines.common.tf_util as U
    sess = U.single_threaded_session()
    sess.__enter__()

    rank = MPI.COMM_WORLD.Get_rank()
    if rank == 0:
        logger.configure()
    else:
        logger.configure(format_strs=[])
    logger.set_level(logger.DISABLED)
    workerseed = seed + 1000000 * rank
    def policy_fn(name, ob_space, ac_space):
        return MlpPolicy(name=name, ob_space=ob_space, ac_space=ac_space,
                         hid_size=32, num_hid_layers=2)
    if render:
        env.render()
    env.seed(workerseed)
    set_global_seeds(workerseed)
    pposgd_simple.learn(env, policy_fn,
                        max_timesteps=num_timesteps,
                        timesteps_per_actorbatch=horizon,
                        clip_param=0.2, entcoeff=0.0,
                        optim_epochs=optim_epochs,
                        optim_stepsize=optim_stepsize,
                        optim_batchsize=optim_batchsize,
                        gamma=0.99, lam=0.95, schedule=schedule,
                        flight_log = None,
                        ckpt_dir = ckpt_dir,
                        restore_dir = restore_dir,
                        save_timestep_period= ckpt_freq
                        )
    env.close()