72 lines
4.1 KiB
Python
72 lines
4.1 KiB
Python
"""
|
|
Reference:
|
|
[1] Caruana R. Multitask learning[J]. Machine learning, 1997.(http://reports-archive.adm.cs.cmu.edu/anon/1997/CMU-CS-97-203.pdf)
|
|
"""
|
|
|
|
import tensorflow as tf
|
|
|
|
from deepctr.feature_column import build_input_features, input_from_feature_columns
|
|
from deepctr.layers.core import PredictionLayer, DNN
|
|
from deepctr.layers.utils import combined_dnn_input
|
|
|
|
def Shared_Bottom(dnn_feature_columns, num_tasks=None, task_types=None, task_names=None,
|
|
bottom_dnn_units=[128, 128], tower_dnn_units_lists=[[64,32], [64,32]],
|
|
l2_reg_embedding=0.00001, l2_reg_dnn=0, seed=1024, dnn_dropout=0,dnn_activation='relu', dnn_use_bn=False):
|
|
"""Instantiates the Shared_Bottom multi-task learning Network architecture.
|
|
|
|
:param dnn_feature_columns: An iterable containing all the features used by deep part of the model.
|
|
:param num_tasks: integer, number of tasks, equal to number of outputs, must be greater than 1.
|
|
:param task_types: list of str, indicating the loss of each tasks, ``"binary"`` for binary logloss or ``"regression"`` for regression loss. e.g. ['binary', 'regression']
|
|
:param task_names: list of str, indicating the predict target of each tasks
|
|
|
|
:param bottom_dnn_units: list,list of positive integer or empty list, the layer number and units in each layer of shared-bottom DNN
|
|
:param tower_dnn_units_lists: list, list of positive integer list, its length must be euqal to num_tasks, the layer number and units in each layer of task-specific DNN
|
|
|
|
:param l2_reg_embedding: float. L2 regularizer strength applied to embedding vector
|
|
:param l2_reg_dnn: float. L2 regularizer strength applied to DNN
|
|
:param seed: integer ,to use as random seed.
|
|
:param dnn_dropout: float in [0,1), the probability we will drop out a given DNN coordinate.
|
|
:param dnn_activation: Activation function to use in DNN
|
|
:param dnn_use_bn: bool. Whether use BatchNormalization before activation or not in DNN
|
|
:return: A Keras model instance.
|
|
"""
|
|
if num_tasks <= 1:
|
|
raise ValueError("num_tasks must be greater than 1")
|
|
if len(task_types) != num_tasks:
|
|
raise ValueError("num_tasks must be equal to the length of task_types")
|
|
|
|
for task_type in task_types:
|
|
if task_type not in ['binary', 'regression']:
|
|
raise ValueError("task must be binary or regression, {} is illegal".format(task_type))
|
|
|
|
if num_tasks != len(tower_dnn_units_lists):
|
|
raise ValueError("the length of tower_dnn_units_lists must be euqal to num_tasks")
|
|
|
|
features = build_input_features(dnn_feature_columns)
|
|
inputs_list = list(features.values())
|
|
|
|
sparse_embedding_list, dense_value_list = input_from_feature_columns(features, dnn_feature_columns, l2_reg_embedding,seed)
|
|
|
|
dnn_input = combined_dnn_input(sparse_embedding_list, dense_value_list)
|
|
shared_bottom_output = DNN(bottom_dnn_units, dnn_activation, l2_reg_dnn, dnn_dropout, dnn_use_bn, seed=seed)(dnn_input)
|
|
|
|
tasks_output = []
|
|
for task_type, task_name, tower_dnn in zip(task_types, task_names, tower_dnn_units_lists):
|
|
tower_output = DNN(tower_dnn, dnn_activation, l2_reg_dnn, dnn_dropout, dnn_use_bn, seed=seed, name='tower_'+task_name)(shared_bottom_output)
|
|
|
|
logit = tf.keras.layers.Dense(1, use_bias=False, activation=None)(tower_output)
|
|
output = PredictionLayer(task_type, name=task_name)(logit) #regression->keep, binary classification->sigmoid
|
|
tasks_output.append(output)
|
|
|
|
model = tf.keras.models.Model(inputs=inputs_list, outputs=tasks_output)
|
|
return model
|
|
|
|
|
|
if __name__ == "__main__":
|
|
from utils import get_mtl_data
|
|
dnn_feature_columns, train_model_input, test_model_input, y_list = get_mtl_data()
|
|
model = Shared_Bottom(dnn_feature_columns, num_tasks=2, task_types= ['binary', 'binary'],
|
|
task_names=['label_income','label_marital'], bottom_dnn_units=[16],
|
|
tower_dnn_units_lists=[[8],[8]])
|
|
model.compile("adam", loss=["binary_crossentropy", "binary_crossentropy"], metrics=['AUC'])
|
|
history = model.fit(train_model_input, y_list, batch_size=256, epochs=5, verbose=2, validation_split=0.0 ) |