RuyiLuo
55 lines
2.3 KiB
Python
55 lines
2.3 KiB
Python
import tensorflow as tf
|
|
from tensorflow.keras.models import Model
|
|
from tensorflow.keras.layers import Flatten, Concatenate, Dense
|
|
from features import FeatureEncoder, get_linear_logits
|
|
from layers import FM, DNN, PredictLayer
|
|
|
|
|
|
def process_feature(linear_feature_columns, dnn_feature_columns, feature_encode):
|
|
"""
|
|
根据FeatureEncoder获取所有输入的Input层或者Embedding层，然后根据自己
|
|
实际场景的业务数据，对不同的特征进行处理.
|
|
"""
|
|
linear_input_sparse_dict = feature_encode.linear_sparse_feature_dict
|
|
group_embedding_dict = feature_encode.sparse_feature_dict
|
|
|
|
linear_input_sparse_list = list(linear_input_sparse_dict['default_group'].values())
|
|
linear_dense_dict = feature_encode.dense_feature_dict
|
|
linear_dense_list = list(linear_dense_dict.values())
|
|
|
|
dnn_emb_name = [fc.embedding_name for fc in dnn_feature_columns]
|
|
dnn_input_list = [v for k, v in group_embedding_dict['fm'].items()
|
|
if k in dnn_emb_name]
|
|
|
|
return linear_dense_list, linear_input_sparse_list, dnn_input_list
|
|
|
|
|
|
def DeepFM(linear_feature_columns, dnn_feature_columns, hidden_units=(32, 16, 1),
|
|
activation='relu', use_bias=True, dp_rate=0.2, use_bn=True, use_dp=True,
|
|
task='binary'):
|
|
# 构建所有特征的Input层和Embedding层
|
|
# 因为有些特征可能在linear层和dnn层会重复了
|
|
feature_columns = list(set(linear_feature_columns + dnn_feature_columns))
|
|
feature_encode = FeatureEncoder(feature_columns,
|
|
linear_sparse_feature=linear_feature_columns)
|
|
feature_input_layers_list = list(feature_encode.feature_input_layer_dict.values())
|
|
|
|
# 特征处理
|
|
linear_dense_list, linear_input_sparse_list, dnn_input_list = \
|
|
process_feature(linear_feature_columns, dnn_feature_columns, feature_encode)
|
|
|
|
fm_logits = FM()(dnn_input_list)
|
|
linear_logits = get_linear_logits(linear_dense_list, linear_input_sparse_list)
|
|
|
|
dnn_inputs = Flatten()(Concatenate(axis=1)(dnn_input_list))
|
|
dnn_logits = DNN(hidden_units, activation, use_bias, use_dp, dp_rate, dp_rate,
|
|
use_bn, get_logits=True)(dnn_inputs)
|
|
|
|
logits = linear_logits + fm_logits + dnn_logits
|
|
output = PredictLayer(task)(logits)
|
|
model = Model(feature_input_layers_list, output)
|
|
|
|
return model
|
|
|
|
|