def test_subprocess_env_endtoend(num_envs):
env_manager = SubprocessEnvManager(simple_env_factory,
EngineConfig.default_config(), num_envs)
trainer_config = generate_config(PPO_CONFIG)
# Run PPO using env_manager
_check_environment_trains(
simple_env_factory(0, []),
trainer_config,
env_manager=env_manager,
success_threshold=None,
)
# Note we can't check the env's rewards directly (since they're in separate processes) so we
# check the StatsReporter's debug stat writer's last reward.
assert isinstance(StatsReporter.writers[0], DebugWriter)
assert all(val > 0.99
for val in StatsReporter.writers[0].get_last_rewards().values())
env_manager.close()
def test_subprocess_env_endtoend(num_envs):
def simple_env_factory(worker_id, config):
env = SimpleEnvironment(["1D"], use_discrete=True)
return env
env_manager = SubprocessEnvManager(simple_env_factory,
EngineConfig.default_config(), num_envs)
trainer_config = generate_config(PPO_CONFIG,
override_vals={"max_steps": 5000})
# Run PPO using env_manager
_check_environment_trains(
simple_env_factory(0, []),
trainer_config,
env_manager=env_manager,
success_threshold=None,
)
# Note we can't check the env's rewards directly (since they're in separate processes) so we
# check the StatsReporter's debug stat writer's last reward.
assert isinstance(StatsReporter.writers[0], DebugWriter)
assert all(val > 0.7
for val in StatsReporter.writers[0].get_last_rewards().values())
env_manager.close()