galileo_framework_tf_python_unsupervised_Unsupervised.3.md

date	section	title
2021年 九月 1日 星期三	3	galileo.framework.tf.python.unsupervised.Unsupervised

NAME

galileo.framework.tf.python.unsupervised.Unsupervised - unsupervised network embedding layer

SYNOPSIS

\

继承自 Model , 以及 galileo.framework.python.base_unsupervised.BaseUnsupervised .

Public 成员函数

def __init__ (self, loss_name='neg_cross_entropy', metric_names='mrr', is_add_metrics=True, *args, **kwargs)\

def target_encoder (self, inputs)
unsupervised target encoder

def context_encoder (self, inputs)
unsupervised context encoder

def compute_logits (self, target, context)
compute logits

def loss_and_metrics (self, logits, negative_logits)\

def convert_ids_tensor (self, inputs)
convert ids tensor

def convert_features_tensor (self, inputs)
convert features tensor

def call (self, inputs)\

def get_config (self)\

Public 属性

loss_name\

loss_obj\

metric_names\

is_add_metrics\

metric_objs\

详细描述

unsupervised network embedding layer

Methods that the subclass must implement:
target_encoder, context_encoder,

参数

loss_name loss name
metric_names metric names, default is mrr
is_add_metrics add loss and metrics layers for keras

成员函数说明

def galileo.framework.tf.python.unsupervised.Unsupervised.loss_and_metrics ( self, logits, negative_logits)

返回

a dict of loss and metrics or y_true and y_pred

重载 galileo.framework.python.base_unsupervised.BaseUnsupervised .

作者

由 Doyxgen 通过分析 Galileo 的 源代码自动生成.