def main():
parser = argparse.ArgumentParser(description="Halite II training")
parser.add_argument("--model_name", help="Name of the model")
parser.add_argument("--minibatch_size", type=int, help="Size of the minibatch", default=100)
parser.add_argument("--steps", type=int, help="Number of steps in the training", default=100)
parser.add_argument("--data", help="Data directory or zip file containing uncompressed games")
parser.add_argument("--cache", help="Location of the model we should continue to train")
parser.add_argument("--games_limit", type=int, help="Train on up to games_limit games", default=1000)
parser.add_argument("--seed", type=int, help="Random seed to make the training deterministic")
parser.add_argument("--bot_to_imitate", help="Name of the bot whose strategy we want to learn")
parser.add_argument("--dump_features_location", help="Location of hdf file where the features should be stored")
args = parser.parse_args()
# Make deterministic if needed
if args.seed is not None:
np.random.seed(args.seed)
nn = NeuralNet(cached_model=args.cache, seed=args.seed)
if args.data.endswith('.zip'):
raw_data = fetch_data_zip(args.data, args.games_limit)
else:
raw_data = fetch_data_dir(args.data, args.games_limit)
data_input, data_output = parse(raw_data, args.bot_to_imitate, args.dump_features_location)
data_size = len(data_input)
training_input, training_output = data_input[:int(0.85 * data_size)], data_output[:int(0.85 * data_size)]
validation_input, validation_output = data_input[int(0.85 * data_size):], data_output[int(0.85 * data_size):]
training_data_size = len(training_input)
# randomly permute the data
permutation = np.random.permutation(training_data_size)
training_input, training_output = training_input[permutation], training_output[permutation]
print("Initial, cross validation loss: {}".format(nn.compute_loss(validation_input, validation_output)))
curves = []
for s in range(args.steps):
start = (s * args.minibatch_size) % training_data_size
end = start + args.minibatch_size
training_loss = nn.fit(training_input[start:end], training_output[start:end])
if s % 200 == 0 or s == args.steps - 1:
validation_loss = nn.compute_loss(validation_input, validation_output)
print("Step: {}, cross validation loss: {}, training_loss: {}".format(s, validation_loss, training_loss))
curves.append((s, training_loss, validation_loss))
cf = pd.DataFrame(curves, columns=['step', 'training_loss', 'cv_loss'])
fig = cf.plot(x='step', y=['training_loss', 'cv_loss']).get_figure()
# Save the trained model, so it can be used by the bot
current_directory = os.path.dirname(os.path.abspath(__file__))
model_path = os.path.join(current_directory, os.path.pardir, "models", args.model_name + ".ckpt")
print("Training finished, serializing model to {}".format(model_path))
nn.save(model_path)
print("Model serialized")
curve_path = os.path.join(current_directory, os.path.pardir, "models", args.model_name + "_training_plot.png")
fig.savefig(curve_path)
def main():
parser = argparse.ArgumentParser(description="Halite II training")
parser.add_argument("--model_name", help="Name of the model")
parser.add_argument(
"--features_ready",
help="If data already preprocessed to features and saved")
parser.add_argument(
"--data",
help="Data directory or zip file containing uncompressed games")
parser.add_argument(
"--cache", help="Location of the model we should continue to train")
parser.add_argument("--games_limit",
type=int,
help="Train on up to games_limit games",
default=1000)
parser.add_argument("--seed",
type=int,
help="Random seed to make the training deterministic")
parser.add_argument("--bot_to_imitate",
help="Name of the bot whose strategy we want to learn")
parser.add_argument(
"--dump_features_location",
help="Location of hdf file where the features should be stored")
parser.add_argument(
"--dump_features_fn",
help="File name of file where the features should be stored")
args = parser.parse_args()
if args.data.endswith('.zip'):
raw_data = fetch_data_zip(args.data, args.games_limit)
data_input, data_output = parse(raw_data, args.bot_to_imitate,
args.dump_features_location,
args.dump_features_fn)
elif args.features_ready:
data_input, data_output = load_parsed(args.dump_features_location,
args.dump_features_fn)
else:
raw_data = fetch_data_dir(args.data, args.games_limit)
data_input, data_output = parse(raw_data, args.bot_to_imitate,
args.dump_features_location,
args.dump_features_fn)
input_data_size = data_input.shape[1:]
# Make deterministic if needed
if args.seed is not None:
np.random.seed(args.seed)
# Redirect sys.stdout to the file
stderr_fn = sys.stderr
stdout_fn = sys.stdout
sys.stdout = open('./LogSTDOUT.txt', 'w')
sys.stdout = open('./LogSTDERR.txt', 'w')
net = CNN_Net(input_size=input_data_size,
output_size=PLANET_MAX_NUM,
cached_model=args.cache,
cached_model_path="",
seed=args.seed)
sys.stdout.close()
sys.stderr.close()
sys.stderr = stderr_fn
sys.stdout = stdout_fn
net.train(data_input,
data_output,
validation_split=0.2,
n_epochs=28,
batch_size=100,
verbose=1,
model_version="v0")
def main():
parser = argparse.ArgumentParser(description="Halite II training")
parser.add_argument("--model_name",
help="Name of the model",
default='xgb_model')
parser.add_argument(
"--data",
help="Data directory or zip file containing uncompressed games")
parser.add_argument("--games_limit",
type=int,
help="Train on up to games_limit games")
parser.add_argument("--seed",
type=int,
help="Random seed to make the training deterministic",
default=53)
parser.add_argument("--bot_to_imitate",
help="Name of the bot whose strategy we want to learn")
parser.add_argument("--cache_train_data",
help="Location of where to store/read cache data")
parser.add_argument('--only_cache_data', action='store_true')
args = parser.parse_args()
# Make deterministic if needed
if args.seed is not None:
np.random.seed(args.seed)
random.seed(args.seed)
if args.cache_train_data:
base_cache_path = os.path.splitext(args.cache_train_data)[0]
ext = os.path.splitext(args.cache_train_data)[1]
path_data_input = base_cache_path + '_input' + ext
path_data_output = base_cache_path + '_output' + ext
if not args.cache_train_data or not os.path.exists(path_data_input):
if args.data.endswith('.zip'):
raw_data = fetch_data_zip(args.data, args.games_limit)
else:
raw_data = fetch_data_dir(args.data, args.games_limit)
# if no cache path or file not exists then calculate train data
data_input, data_output = parse(raw_data, args.bot_to_imitate)
if args.cache_train_data:
np.save(path_data_input, data_input)
np.save(path_data_output, data_output)
else:
print('Reading cache')
# cache path exists and file exists
data_input = np.load(path_data_input)
data_output = np.load(path_data_output)
if args.only_cache_data:
return
data_input_p = np.reshape(data_input, (data_input.shape[0], -1))
data_output_p = np.argmax(data_output, axis=1)
X_train, X_test, y_train, y_test = \
model_selection.train_test_split(data_input_p, data_output_p, train_size=0.85, random_state=args.seed,
shuffle=True)
classifier = xgboost.XGBRegressor(
objective='multi:softprob',
n_jobs=-1,
random_state=args.seed,
)
classifier.set_params(**{'num_class': data_output.shape[1]})
fit_res = classifier.fit(X_train,
y_train,
eval_set=[(X_train, y_train), (X_test, y_test)],
eval_metric='mlogloss')
# Save the trained model, so it can be used by the bot
current_directory = os.path.dirname(os.path.abspath(__file__))
model_path = os.path.join(current_directory, os.path.pardir, "models",
args.model_name + ".pickle")
print("Training finished, serializing model to {}".format(model_path))
with open(model_path, 'wb') as f:
pickle.dump(classifier, f)
print("Model serialized")