def test_initialize_trainers(mock_communicator, mock_launcher):
open_name = 'unitytrainers.trainer_controller' + '.open'
with mock.patch('yaml.load') as mock_load:
with mock.patch(open_name, create=True) as _:
mock_communicator.return_value = MockCommunicator(
discrete_action=True, visual_inputs=1)
tc = TrainerController(' ', ' ', 1, None, True, True, False, 1, 1,
1, 1, '', "tests/test_unitytrainers.py",
False)
# Test for PPO trainer
mock_load.return_value = dummy_config
config = tc._load_config()
tf.reset_default_graph()
with tf.Session() as sess:
tc._initialize_trainers(config, sess)
assert (len(tc.trainers) == 1)
assert (isinstance(tc.trainers['RealFakeBrain'], PPOTrainer))
# Test for Behavior Cloning Trainer
mock_load.return_value = dummy_bc_config
config = tc._load_config()
tf.reset_default_graph()
with tf.Session() as sess:
tc._initialize_trainers(config, sess)
assert (isinstance(tc.trainers['RealFakeBrain'],
BehavioralCloningTrainer))
# Test for proper exception when trainer name is incorrect
mock_load.return_value = dummy_bad_config
config = tc._load_config()
tf.reset_default_graph()
with tf.Session() as sess:
with pytest.raises(UnityEnvironmentException):
tc._initialize_trainers(config, sess)
def test_initialization():
with mock.patch('subprocess.Popen'):
with mock.patch('socket.socket') as mock_socket:
with mock.patch('glob.glob') as mock_glob:
mock_glob.return_value = ['FakeLaunchPath']
mock_socket.return_value.accept.return_value = (mock_socket, 0)
mock_socket.recv.return_value.decode.return_value = dummy_start
tc = TrainerController(' ', ' ', 1, None, True, True, False, 1,
1, 1, 1, '', "tests/test_unitytrainers.py")
assert(tc.env.brain_names[0] == 'RealFakeBrain')
def test_load_config(mock_communicator, mock_launcher):
open_name = 'unitytrainers.trainer_controller' + '.open'
with mock.patch('yaml.load') as mock_load:
with mock.patch(open_name, create=True) as _:
mock_load.return_value = dummy_config
mock_communicator.return_value = MockCommunicator(
discrete_action=True, visual_inputs=1)
mock_load.return_value = dummy_config
tc = TrainerController(' ', ' ', 1, None, True, True, False, 1, 1,
1, 1, '', '', False)
config = tc._load_config()
assert (len(config) == 1)
assert (config['default']['trainer'] == "ppo")
def test_load_config():
open_name = 'unitytrainers.trainer_controller' + '.open'
with mock.patch('yaml.load') as mock_load:
with mock.patch(open_name, create=True) as _:
with mock.patch('subprocess.Popen'):
with mock.patch('socket.socket') as mock_socket:
with mock.patch('glob.glob') as mock_glob:
mock_load.return_value = dummy_config
mock_glob.return_value = ['FakeLaunchPath']
mock_socket.return_value.accept.return_value = (mock_socket, 0)
mock_socket.recv.return_value.decode.return_value = dummy_start
mock_load.return_value = dummy_config
tc = TrainerController(' ', ' ', 1, None, True, True, False, 1,
1, 1, 1, '','')
config = tc._load_config()
assert(len(config) == 1)
assert(config['default']['trainer'] == "ppo")
def run_training(sub_id, use_seed, options):
# Docker Parameters
if options['--docker-target-name'] == 'Empty':
docker_target_name = ''
else:
docker_target_name = options['--docker-target-name']
# General parameters
run_id = options['--run-id']
num_runs = int(options['--num-runs'])
seed = int(options['--seed'])
load_model = options['--load']
train_model = options['--train']
save_freq = int(options['--save-freq'])
env_path = options['<env>']
keep_checkpoints = int(options['--keep-checkpoints'])
worker_id = int(options['--worker-id'])
curriculum_file = str(options['--curriculum'])
if curriculum_file == "None":
curriculum_file = None
lesson = int(options['--lesson'])
fast_simulation = not bool(options['--slow'])
no_graphics = options['--no-graphics']
# Constants
# Assumption that this yaml is present in same dir as this file
base_path = os.path.dirname(__file__)
TRAINER_CONFIG_PATH = os.path.abspath(
os.path.join(base_path, "trainer_config.yaml"))
if env_path is None and num_runs > 1:
raise TrainerError(
"It is not possible to launch more than one concurrent training session "
"when training from the editor")
tc = TrainerController(env_path, run_id + "-" + str(sub_id), save_freq,
curriculum_file, fast_simulation, load_model,
train_model, worker_id + sub_id, keep_checkpoints,
lesson, use_seed, docker_target_name,
TRAINER_CONFIG_PATH, no_graphics)
tc.start_learning()
def test_initialize_trainers():
open_name = 'unitytrainers.trainer_controller' + '.open'
with mock.patch('yaml.load') as mock_load:
with mock.patch(open_name, create=True) as _:
with mock.patch('subprocess.Popen'):
with mock.patch('socket.socket') as mock_socket:
with mock.patch('glob.glob') as mock_glob:
mock_glob.return_value = ['FakeLaunchPath']
mock_socket.return_value.accept.return_value = (mock_socket, 0)
mock_socket.recv.return_value.decode.return_value = dummy_start
tc = TrainerController(' ', ' ', 1, None, True, True, False, 1,
1, 1, 1, '', "tests/test_unitytrainers.py")
# Test for PPO trainer
mock_load.return_value = dummy_config
config = tc._load_config()
tf.reset_default_graph()
with tf.Session() as sess:
tc._initialize_trainers(config, sess)
assert(len(tc.trainers) == 1)
assert(isinstance(tc.trainers['RealFakeBrain'], PPOTrainer))
# Test for Behavior Cloning Trainer
mock_load.return_value = dummy_bc_config
config = tc._load_config()
tf.reset_default_graph()
with tf.Session() as sess:
tc._initialize_trainers(config, sess)
assert(isinstance(tc.trainers['RealFakeBrain'], BehavioralCloningTrainer))
# Test for proper exception when trainer name is incorrect
mock_load.return_value = dummy_bad_config
config = tc._load_config()
tf.reset_default_graph()
with tf.Session() as sess:
with pytest.raises(UnityEnvironmentException):
tc._initialize_trainers(config, sess)
logger.info(options)
# Docker Parameters
if options['--docker-target-name'] == 'Empty':
docker_target_name = ''
else:
docker_target_name = options['--docker-target-name']
# General parameters
run_id = options['--run-id']
seed = int(options['--seed'])
load_model = options['--load']
train_model = options['--train']
save_freq = int(options['--save-freq'])
env_path = options['<env>']
keep_checkpoints = int(options['--keep-checkpoints'])
worker_id = int(options['--worker-id'])
curriculum_file = str(options['--curriculum'])
if curriculum_file == "None":
curriculum_file = None
lesson = int(options['--lesson'])
fast_simulation = not bool(options['--slow'])
# Constants
# Assumption that this yaml is present in same dir as this file
base_path = os.path.dirname(__file__)
TRAINER_CONFIG_PATH = os.path.abspath(os.path.join(base_path, "trainer_config.yaml"))
tc = TrainerController(env_path, run_id, save_freq, curriculum_file, fast_simulation, load_model, train_model,
worker_id, keep_checkpoints, lesson, seed, docker_target_name, TRAINER_CONFIG_PATH)
tc.start_learning()
def test_initialization(mock_communicator, mock_launcher):
mock_communicator.return_value = MockCommunicator(discrete_action=True,
visual_inputs=1)
tc = TrainerController(' ', ' ', 1, None, True, True, False, 1, 1, 1, 1,
'', "tests/test_unitytrainers.py", False)
assert (tc.env.brain_names[0] == 'RealFakeBrain')
from unitytrainers.trainer_controller import TrainerController
import os
import numpy as np
import time
tc = TrainerController('python/Balance', None, None, None, None, None, True,
0, 0, 0, 0, '', os.path.abspath(os.path.join(os.path.dirname(__file__), "trainer_config.yaml")))
class Net():
def forward(self):
return np.random.randint(0,180)
pop = 10
bots = []
env = tc.env
[bots.append(Net()) for i in range(pop)]
default_brain = env.external_brain_names[0]
brain = env.brains[default_brain]
rewards = [0] * pop
fittest = {}
max_gens = 1000
max_frames = 10000
state = env.reset(train_mode=True)[default_brain]
print()
for generation in range(max_gens):
print(" GENERATION "+str(generation))
print("----------------------------")
print()
for count, bot in enumerate(bots):
frame = 0
terminal = False
docker_target_name = options['--docker-target-name']
# General parameters
run_id = options['--run-id']
seed = int(options['--seed'])
load_model = options['--load']
train_model = options['--train']
save_freq = int(options['--save-freq'])
env_path = options['<env>']
keep_checkpoints = int(options['--keep-checkpoints'])
worker_id = int(options['--worker-id'])
curriculum_file = str(options['--curriculum'])
if curriculum_file == "None":
curriculum_file = None
lesson = int(options['--lesson'])
fast_simulation = not bool(options['--slow'])
no_graphics = options['--no-graphics']
trainer_config_path = options['--trainer-config-path']
if not trainer_config_path:
# Assumption that this yaml is present in same dir as this file
base_path = os.path.dirname(__file__)
trainer_config_path = trainer_config_path or os.path.abspath(
os.path.join(base_path, "trainer_config.yaml"))
tc = TrainerController(env_path, run_id, save_freq, curriculum_file,
fast_simulation, load_model, train_model, worker_id,
keep_checkpoints, lesson, seed, docker_target_name,
trainer_config_path, no_graphics)
tc.start_learning()
logger.info(options)
# Docker Parameters
if options['--docker-target-name'] == 'Empty':
docker_target_name = ''
else:
docker_target_name = options['--docker-target-name']
# General parameters
run_id = options['--run-id']
seed = int(options['--seed'])
load_model = options['--load']
train_model = options['--train']
save_freq = int(options['--save-freq'])
env_path = options['<env>']
keep_checkpoints = int(options['--keep-checkpoints'])
worker_id = int(options['--worker-id'])
curriculum_file = str(options['--curriculum'])
if curriculum_file == "None":
curriculum_file = None
lesson = int(options['--lesson'])
fast_simulation = not bool(options['--slow'])
# Constants
# Assumption that this yaml is present in same dir as this file
base_path = os.path.dirname(__file__)
TRAINER_CONFIG_PATH = os.path.abspath(
os.path.join(base_path, "trainer_config.yaml"))
tc = TrainerController(env_path, None, None, None, None, None, train_model,
0, keep_checkpoints, 0, 0, '', TRAINER_CONFIG_PATH)