def main(_): model = linear_model(output_dim=1) train_data = train_input_fn(batch_size=500) validate_data = test_input_fn(batch_size=1000) model.compile(optimizer=tf.keras.optimizers.SGD(learning_rate=0.001), loss=tf.keras.losses.mean_squared_error, metrics=tf.keras.metrics.RootMeanSquaredError()) model.fit(x=train_data, validation_data=validate_data, epochs=1) tf.keras.utils.plot_model(model, to_file="./lr.png", rankdir="BT")
def main(_): model = afm_model(embedding_size=10, l2_factor=0.5, hidden_unit=10, dropout=0.5) train_data = train_input_fn(batch_size=500) validate_data = test_input_fn(batch_size=1000) model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.001), loss=tf.keras.losses.mean_squared_error, metrics=tf.keras.metrics.RootMeanSquaredError()) model.fit(x=train_data, validation_data=validate_data, epochs=1) tf.keras.utils.plot_model(model, to_file="./afm.png", rankdir="BT")
def main(_): model = wide_deep_model(hidden_units=[128, 128, 128], dropout=0.5) train_data = train_input_fn(batch_size=500) validate_data = test_input_fn(batch_size=1000) model.compile(optimizer=tf.keras.optimizers.SGD(learning_rate=0.001), loss=tf.keras.losses.mean_squared_error, metrics=tf.keras.metrics.RootMeanSquaredError()) model.fit(x=train_data, validation_data=validate_data, epochs=1) tf.keras.utils.plot_model(model, to_file="./wide_deep.png", rankdir="BT")
def main(_): model = svd_plus_plus_model(average_score=3.5, embedding_size=5, l2_factor=0.5) train_data = train_input_fn(batch_size=500) validate_data = test_input_fn(batch_size=1000) model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.001), run_eagerly=True, loss=tf.keras.losses.mean_squared_error, metrics=tf.keras.metrics.RootMeanSquaredError()) model.fit(x=train_data, validation_data=validate_data, epochs=1) tf.keras.utils.plot_model(model, to_file="./svd++.png", rankdir="BT")
def main(_): model = youtube_net_model(embedding_size=128, hidden_units=[256, 256, 128], dropout=0.5) train_data = train_input_fn(batch_size=500, label_key="label") validate_data = test_input_fn(batch_size=1000, label_key="label") model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.001), loss=tf.keras.losses.mean_squared_error, metrics=tf.keras.metrics.RootMeanSquaredError()) model.fit(x=train_data, validation_data=validate_data, epochs=1) tf.keras.utils.plot_model(model, to_file="./youtube_net.png", rankdir="BT")
def main(_): model = din_model(hidden_units=[32, 32, 32], dropout=0.5, attention_hidden_unit=8) train_data = train_input_fn(batch_size=100) validate_data = test_input_fn(batch_size=1000) model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.01), loss=tf.keras.losses.mean_squared_error, metrics=tf.keras.metrics.RootMeanSquaredError()) model.fit(x=train_data, validation_data=validate_data, epochs=1) tf.keras.utils.plot_model(model, to_file="./din.png", rankdir="BT")