def main():
config = get_configuration(print_diagnostics=True,
with_neptune=True,
nesting_prefixes=("net_", "env_"))
# You can set locally e.g. POLO_SUPERVISED_OUTPUT=/tmp/supervised_output/
output_dir = os.environ.get("POLO_SUPERVISED_OUTPUT", os.getcwd())
print("Output directory: {}".format(output_dir))
experiment = SupervisedExperiment(output_dir=output_dir, **config)
experiment.run_experiment()
def main():
"""Follows example from mrunner experiment_gin.py
Requires BASE_PATH and GAME to be passed in config
:return:
"""
params = get_configuration(print_diagnostics=True, with_neptune=True, inject_parameters_to_gin=True)
LOG_PATH = os.path.join(params.BASE_PATH, 'tests', params.GAME)
runner = RolloutsRunner(LOG_PATH,create_rainbow_rollouts_agent)
runner.run_experiment()
def main():
with_neptune = True
if len(os.environ.get('NEPTUNE_API_TOKEN')) == 0:
print('empty token, run without neptune', file=sys.stderr)
with_neptune = False
params = get_configuration(print_diagnostics=True,
with_neptune=with_neptune)
np.random.seed(None)
learn_BitFlipper_HER(10)
def main(argv):
if len(argv) > 1:
raise app.UsageError('Too many command-line arguments.')
if FLAGS.run_mode == 'actor':
if not FLAGS.is_local:
get_configuration(config_file=FLAGS.mrunner_config,
inject_parameters_to_FLAGS=True)
actor.actor_loop(env.create_environment)
elif FLAGS.run_mode == 'learner':
if not FLAGS.is_local:
get_configuration(config_file=FLAGS.mrunner_config,
print_diagnostics=True,
with_neptune=True,
inject_parameters_to_FLAGS=True)
experiment = neptune.get_experiment()
experiment.append_tag(tag=FLAGS.nonce)
neptune_tensorboard.integrate_with_tensorflow()
learner.learner_loop(env.create_environment, create_agent,
create_optimizer)
elif FLAGS.run_mode == 'visualize':
visualize.visualize(env.create_environment, create_agent,
create_optimizer)
else:
raise ValueError('Unsupported run mode {}'.format(FLAGS.run_mode))
def main():
try:
params = get_configuration(print_diagnostics=True, with_neptune=True)
except TypeError:
print(' ************************************************\n',
' NEPTUNE DISABLED \n',
'************************************************')
sleep(2)
# env = gym.make("CartPole-v1")
# env = gym.make("MountainCar-v0")
# env = make_env_BitFlipper(n=5, space_seed=None)
# env = make_env_GoalBitFlipper(n=5, space_seed=None)
env = make_env_GoalRubik(step_limit=100, shuffles=100)
model = HER(env)
# model.learn(100000 * 16 * 50)
model.learn(120000000)
def main():
params = get_configuration(print_diagnostics=True, with_neptune=True, inject_parameters_to_gin=True)
LOG_PATH = os.path.join(params.BASE_PATH, 'tests', params.GAME)
runner = RainbowBasicRunner(LOG_PATH, create_agent)
runner.run_experiment()
import gym
from mrunner.helpers.client_helper import get_configuration
from spinup.sac import sac
def get_get_env(env_name):
def get_env():
return gym.make(env_name)
return get_env
def main(task, seed, steps, replay_size, batch_size, hidden_sizes):
sac(get_get_env(task),
seed=seed,
steps=steps,
replay_size=replay_size,
batch_size=batch_size,
actor_kwargs=dict(hidden_sizes=hidden_sizes),
critic_kwargs=dict(hidden_sizes=hidden_sizes))
if __name__ == '__main__':
config = get_configuration(print_diagnostics=True, with_neptune=True)
experiment_id = config.pop('experiment_id')
main(**config)
import gym
import neptune
from stable_baselines import DQN
from stable_baselines.logger import KVWriter
from stable_baselines.logger import Logger
from mrunner.helpers.client_helper import get_configuration
class NeptuneFormat(KVWriter):
def writekvs(self, kvs):
for key, value in sorted(kvs.items()):
neptune.send_metric(key, value)
if __name__ == "__main__":
params = get_configuration(with_neptune=True,
inject_parameters_to_gin=True)
Logger.CURRENT = Logger(folder=None, output_formats=[NeptuneFormat()])
# Create environment
env = gym.make('LunarLander-v2')
# Instantiate the agent
model = DQN('MlpPolicy',
env,
learning_rate=params["learning_rate"],
prioritized_replay=params["prioritized_replay"],
verbose=params["verbose"])
# Train the agent
model.learn(total_timesteps=params["total_timesteps"],
log_interval=params["log_interval"])