def create_rl_experiment(experiment_config):
"""Creates a new reinforcement learning `Experiment` instance.
Args:
experiment_config: the config to use for creating the experiment.
"""
agent = getters.get_agent(experiment_config.agent_config,
experiment_config.model_config,
experiment_config.run_config)
env = getters.get_environment(experiment_config.environment_config.module,
experiment_config.environment_config.env_id,
**experiment_config.environment_config.params)
train_hooks = getters.get_hooks(experiment_config.train_hooks_config)
eval_hooks = getters.get_hooks(experiment_config.eval_hooks_config)
experiment = RLExperiment(
agent=agent,
env=env,
train_steps=experiment_config.train_steps,
train_episodes=experiment_config.train_episodes,
first_update=experiment_config.first_update,
update_frequency=experiment_config.update_frequency,
eval_steps=experiment_config.eval_steps,
train_hooks=train_hooks,
eval_hooks=eval_hooks,
eval_delay_secs=experiment_config.eval_delay_secs,
continuous_eval_throttle_secs=experiment_config.continuous_eval_throttle_secs,
eval_every_n_steps=experiment_config.eval_every_n_steps,
delay_workers_by_global_step=experiment_config.delay_workers_by_global_step,
export_strategies=experiment_config.export_strategies,
train_steps_per_iteration=experiment_config.train_steps_per_iteration)
return experiment
def create_experiment(experiment_config):
"""Creates a new `Experiment` instance.
Args:
experiment_config: the config to use for creating the experiment.
"""
# Creates training input function
train_input_data_config = experiment_config.train_input_data_config
train_input_fn = create_input_data_fn(
pipeline_config=train_input_data_config.pipeline_config,
mode=ModeKeys.TRAIN,
scope='train_input_fn',
input_type=train_input_data_config.input_type,
x=train_input_data_config.x,
y=train_input_data_config.y)
# Creates eval_input_fn input function
eval_input_data_config = experiment_config.eval_input_data_config
eval_input_fn = create_input_data_fn(
pipeline_config=eval_input_data_config.pipeline_config,
mode=ModeKeys.EVAL,
scope='eval_input_fn',
input_type=eval_input_data_config.input_type,
x=eval_input_data_config.x,
y=eval_input_data_config.y)
estimator = getters.get_estimator(experiment_config.estimator_config,
experiment_config.model_config,
experiment_config.run_config)
train_hooks = getters.get_hooks(experiment_config.train_hooks_config)
eval_hooks = getters.get_hooks(experiment_config.eval_hooks_config)
experiment = Experiment(
estimator=estimator,
train_input_fn=train_input_fn,
eval_input_fn=eval_input_fn,
train_steps=experiment_config.train_steps,
eval_steps=experiment_config.eval_steps,
train_hooks=train_hooks,
eval_hooks=eval_hooks,
eval_delay_secs=experiment_config.eval_delay_secs,
continuous_eval_throttle_secs=experiment_config.
continuous_eval_throttle_secs,
eval_every_n_steps=experiment_config.eval_every_n_steps,
delay_workers_by_global_step=experiment_config.
delay_workers_by_global_step,
export_strategies=experiment_config.export_strategies,
train_steps_per_iteration=experiment_config.train_steps_per_iteration)
return experiment