def main():
# tag::e2e_processor[]
go_board_rows, go_board_cols = 19, 19
nb_classes = go_board_rows * go_board_cols
encoder = SevenPlaneEncoder((go_board_rows, go_board_cols))
processor = GoDataProcessor(encoder=encoder.name())
X, y = processor.load_go_data(num_samples=100)
# end::e2e_processor[]
# tag::e2e_model[]
input_shape = (encoder.num_planes, go_board_rows, go_board_cols)
model = Sequential()
network_layers = large.layers(input_shape)
for layer in network_layers:
model.add(layer)
model.add(Dense(nb_classes, activation='softmax'))
model.compile(loss='categorical_crossentropy',
optimizer='adadelta',
metrics=['accuracy'])
model.fit(X, y, batch_size=128, epochs=20, verbose=1)
# end::e2e_model[]
# tag::e2e_agent[]
deep_learning_bot = DeepLearningAgent(model, encoder)
model_file = h5py.File("../agents/deep_bot.h5", "w")
deep_learning_bot.serialize(model_file)
model_file.close()
def main():
parser = argparse.ArgumentParser()
parser.add_argument('--epochs', '-e', default=20, type=int,
help="Training epochs. Default is 20")
args = parser.parse_args()
# tag::e2e_processor[]
go_board_rows, go_board_cols = 19, 19
nb_classes = go_board_rows * go_board_cols
encoder = SevenPlaneEncoder((go_board_rows, go_board_cols))
processor = GoDataProcessor(encoder=encoder.name())
X, y = processor.load_go_data(num_samples=100)
# end::e2e_processor[]
# tag::e2e_model[]
input_shape = (go_board_rows, go_board_cols, encoder.num_planes)
model = Sequential()
network_layers = drew_example.layers(input_shape)
for layer in network_layers:
model.add(layer)
model.add(Dense(nb_classes, activation='softmax'))
model.compile(loss='categorical_crossentropy', optimizer='adadelta', metrics=['accuracy'])
model.fit(X, y, batch_size=128, epochs=args.epochs, verbose=1)
# end::e2e_model[]
# tag::e2e_agent[]
deep_learning_bot = DeepLearningAgent(model, encoder)
fpath = os.path.join(CODE_ROOT_DIR, "agents/deep_bot.h5")
model_file = h5py.File(fpath, "w")
deep_learning_bot.serialize(model_file)
model_file.close()
def main():
samp = 1000
epo = 1
# tag::e2e_processor[]
timestr = time.strftime("%Y%m%d-%H%M%S")
model_h5filename = "./agents/deep_bot_" + timestr + "_s" + str(
samp) + "e" + str(epo) + ".h5"
go_board_rows, go_board_cols = 19, 19
nb_classes = go_board_rows * go_board_cols
encoder = XPlaneEncoder((go_board_rows, go_board_cols))
data_dir = "data/" + str(encoder.num_planes) + "-planes"
processor = GoDataProcessor(encoder=encoder.name(),
data_directory=data_dir)
X, y = processor.load_go_data(num_samples=samp)
# end::e2e_processor[]
# tag::e2e_model[]
input_shape = (encoder.num_planes, go_board_rows, go_board_cols)
model = Sequential()
network_layers = large.layers(input_shape)
for layer in network_layers:
model.add(layer)
try:
with tf.device('/device:GPU:0'):
model.add(Dense(nb_classes, activation='softmax'))
model.compile(loss='categorical_crossentropy',
optimizer='adadelta',
metrics=['accuracy'])
model.fit(X, y, batch_size=128, epochs=epo, verbose=1)
# end::e2e_model[]
# tag::e2e_agent[]
deep_learning_bot = DeepLearningAgent(model, encoder)
deep_learning_bot.serialize(h5py.File(model_h5filename, "w"))
# end::e2e_agent[]
# tag::e2e_load_agent[]
model_file = h5py.File(model_h5filename, "r")
bot_from_file = load_prediction_agent(model_file)
web_app = get_web_app({'predict': bot_from_file})
web_app.run()
# end::e2e_load_agent[]
except RuntimeError as e:
print(e)
from dlgo.networks import large
from keras.models import Sequential
from keras.layers.core import Dense
from keras.callbacks import ModelCheckpoint
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
go_board_rows, go_board_cols = 19, 19
num_classes = go_board_rows * go_board_cols
num_games = 100
encoder = SevenPlaneEncoder((go_board_rows, go_board_cols))
processor = GoDataProcessor(encoder=encoder.name())
if __name__ == '__main__':
generator = processor.load_go_data('train', num_games, use_generator=True)
test_generator = processor.load_go_data('test',
num_games,
use_generator=True)
input_shape = (encoder.num_planes, go_board_rows, go_board_cols)
network_layers = large.layers(input_shape)
model = Sequential()
for layer in network_layers:
model.add(layer)
model.add(Dense(num_classes, activation='softmax'))
model.compile(loss='categorical_crossentropy',
optimizer='adagrad',