galileo_framework_python_base_supervised_BaseSupervised.3.md

date	section	title
2021年 九月 1日 星期三	3	galileo.framework.python.base_supervised.BaseSupervised

NAME

galileo.framework.python.base_supervised.BaseSupervised - base supervised model for graph

SYNOPSIS

\

继承自 metaclass , 以及 ABCMeta .

被 galileo.framework.pytorch.python.supervised.Supervised , 以及 galileo.framework.tf.python.supervised.Supervised 继承.

Public 成员函数

def __init__ (self, label_dim=None, num_classes=None, *args, **kwargs)\

def encoder (self, inputs)
supervised feature encoder

def dense_encoder (self, inputs)
a dense layer after encoder

def loss_and_metrics (self, labels, logits)
compute loss and metrics

def convert_ids_tensor (self, inputs)
convert ids tensor

def convert_features_tensor (self, inputs)
convert features tensor

def convert_labels_tensor (self, inputs)
convert labels tensor

def unpack_sample (self, inputs)
unpack sample

def __call__ (self, inputs, **kwargs)\

Public 属性

label_dim\

num_classes\

详细描述

base supervised model for graph

参数

label_dim label dim
num_classes num of class

成员函数说明

def galileo.framework.python.base_supervised.BaseSupervised.__call__ ( self, inputs, ** kwargs)

参数

inputs dict of tensors
contains target, features and labels

参见

unpack_sample

def galileo.framework.python.base_supervised.BaseSupervised.unpack_sample ( self, inputs)

unpack sample

参数

inputs dict, keys of dict:
case 1: (features, target) for save embedding
case 2: (features, labels) for train and evaluate

返回

features, target_or_labels, only_embedding

作者

由 Doyxgen 通过分析 Galileo 的 源代码自动生成.