Discuss how to write API document
Created by: ranqiu92
其他主流DL框架API文档格式见 #8834 (closed) 初步拟订PaddlePaddle 文档API 格式如下
API文档模块
API文档须包含以下几个模块(排列顺序为文档撰写顺序):
-
Python API Definition
API的代码定义。
-
Python Wrapper Location
API的Python源码链接。
-
Function Description
API的功能描述。描述该API的含义、作用或对输入所做的操作,及参考文献和对应链接(如果有),必要时给出公式,并解释公式中关键变量的含义。
-
Args Description
API参数介绍。按代码定义中的参数顺序逐个介绍,介绍内容包含数据类型、默认值(如果有)、含义等。
-
Returns
API返回值介绍。介绍返回值含义,必要时给出对应的形状。若返回值为包含多个参数的tuple,则按顺序逐个介绍各参数。
-
Exceptions(如果有)
可能抛出的异常或错误及可能的产生原因。
-
Note(如果有)
注意事项。
-
Examples
API的使用示例。
格式及示例
Python API Definition & Python Wrapper Location
-
格式:
[Python API Name][Python Wrapper URL]
[Python API Definition]
-
示例
fluid.layers.fc
fc(input,
size,
num_flatten_dims=1,
param_attr=None,
bias_attr=None,
act=None,
name=None,
main_program=None,
startup_program=None)
Function Description
-
格式
[Function Description]
[Formula]
[Symbols' Descriptions if necessary]
-
示例
Applies a linear or nonlinear transformation to the input data:
Y = Act(W^T * X + b)
where X is the input value, W is the weight, b is the bias and Act is the activation function.
Args Description
- 格式
[Arg's Name][(Data Type, Default Value)][Description]
- 示例
input (Variable): The input of the layer.
num_flatten_dims (int, default 1): The column number of the input.
Returns
- 格式
[Name][Shape]
- 示例
以dynamic_lstm为例
A tuple containing:
The hidden state of LSTM whose shape is (T X D).
The cell state of LSTM whose shape is (T X D).
Exceptions
- 格式
[Exception Type] [Condition]
- 示例
ValueError: If the rank of the input is less than 2.
Note
- 格式
[Note]
- 示例(暂无)
Examples
- 格式
[Python Code Snipper]
- 示例
data = fluid.layers.data(name='data', shape=[1], dtype='float32')
fc = fluid.layers.fc(input=data, size=10, act="tanh")
完整示例
fluid.layers.fc
fc(input,
size,
num_flatten_dims=1,
param_attr=None,
bias_attr=None,
act=None,
name=None)
Applies a linear or nonlinear transformation to the input data:
Y = Act(W^T * X + b)
where X is the input value, W is the weight, b is the bias and Act is the activation function.
Args
input (Variable): The input value, a tensor with rank at least 2.
size (int): The size of the output.
num_flatten_dims (int, default 1): The column number of the input.
param_attr (ParamAttr or Initializer, default None): The attributes of the weights of this layer.
bias_attr (ParamAttr or Initializer, default None): The attributes of the bias of this layer.
act (str, default None): Activation type of this layer.
name (str, default None): The name of this layer.
Returns
The tensor variable storing the transformation result.
Exceptions
ValueError: If the rank of the input is less than 2.
Examples
data = fluid.layers.data(name='data', shape=[1], dtype='float32')
fc = fluid.layers.fc(input=data, size=10, act="tanh")