Update the annotations of layers.py #5517

ranqiu92 · 2017-11-09T11:12:45Z

lcy-seso

Thank you for this refining.

lcy-seso · 2017-11-14T10:02:18Z

python/paddle/trainer_config_helpers/layers.py

    :type input: LayerOutput
-    :param state: State Layer. :math:`c_{t-1}`
+    :param state: The state of a lstm.


The state of the LSTM unit.

lcy-seso · 2017-11-14T10:04:07Z

python/paddle/trainer_config_helpers/layers.py

    :type state: LayerOutput
    :param act: Activation type. TanhActivation is the default.
    :type act: BaseActivation
-    :param gate_act: Gate Activation Type. SigmoidActivation is the default.
+    :param gate_act: Activation type of the gate. SigmoidActivation is the default.


the default --> the default activation.

lcy-seso · 2017-11-14T10:04:24Z

python/paddle/trainer_config_helpers/layers.py

    :type gate_act: BaseActivation
-    :param state_act: State Activation Type. TanhActivation is the default.
+    :param state_act: Activation type of the state. TanhActivation is the default.


the default --> the default activation.

lcy-seso · 2017-11-14T10:05:44Z

python/paddle/trainer_config_helpers/layers.py

@@ -3573,30 +3573,29 @@ def lstm_step_layer(input,


    This layer has two outputs. Default output is :math:`h_t`. The other
-    output is :math:`o_t`, whose name is 'state' and can use
+    output is :math:`o_t`, whose name is 'state' and users can use


Default output --> The default output

lcy-seso · 2017-11-14T10:07:52Z

python/paddle/trainer_config_helpers/layers.py

+                 it will be set to one-third of the dimension of the input automatically.
+    :type size: int
+    :param act: Activation type of this layer's output. SigmoidActivation
+                is the default.


the default --> the default activation.

lcy-seso · 2017-11-14T10:15:54Z

python/paddle/trainer_config_helpers/layers.py

@@ -3910,7 +3928,7 @@ def recurrent_group(step, input, reverse=False, name=None, targetInlink=None):
    Recurrent layer group is an extremely flexible recurrent unit in
    PaddlePaddle. As long as the user defines the calculation done within a
    time step, PaddlePaddle will iterate such a recurrent calculation over
-    sequence input. This is extremely usefull for attention based model, or
+    sequence input. This is extremely useful for attention-based models, or


把这个 extremely 去掉~ 加上这个副词比较奇怪。

lcy-seso · 2017-11-14T10:16:34Z

python/paddle/trainer_config_helpers/layers.py

-    :param step: recurrent one time step function.The input of this function is
-                 input of the group. The return of this function will be
-                 recurrent group's return value.
+    :param step: A step function which will be executed every step. The input


will be executed every step --> 这句话表意不明确，需要重新组织一下。

lcy-seso · 2017-11-14T10:17:13Z

python/paddle/trainer_config_helpers/layers.py

-                 The recurrent group scatter a sequence into time steps. And
-                 for each time step, will invoke step function, and return
-                 a time step result. Then gather each time step of output into
+                 The recurrent group scatters a sequence into time steps. And


The recurrent group scatters a sequence into separate time steps.

recurrent group 的这一段英文注释的表意还是难以被用户理解（英文描述的质量还是有些问题），这个PR可以先merge目前的修改，后面仍需针对性的完善一下。

lcy-seso · 2017-11-14T10:21:44Z

python/paddle/trainer_config_helpers/layers.py

@@ -4230,8 +4251,9 @@ def rnn_step(input):
    - machine translation : demo/seqToseq/translation/gen.conf \
                            demo/seqToseq/seqToseq_net.py

-    :param name: Name of the recurrent unit that generates sequences.
-    :type name: base string
+    :param name: The name of the recurrent unit that generates sequences.


that generates sequences. --> that is responsible for generating sequences.

lcy-seso · 2017-11-14T10:22:33Z

python/paddle/trainer_config_helpers/layers.py

    :type layer_attr: ExtraLayerAttribute
-    :param coeff: The coefficient affects the gradient in the backward.
+    :param coeff: The weight of the gradient in the back propagation.
+                  1.0 is the default.


the default --> the default value.

lcy-seso

LGTM

Update the annotations of layers.py

0e73967

ranqiu92 requested a review from lcy-seso November 9, 2017 11:12

ranqiu and others added 3 commits November 10, 2017 17:32

Update doc of layers.py

ea8e050

Merge branch 'develop' into doc

9202671

Update the annotations of layers

1baeebc

lcy-seso reviewed Nov 14, 2017

View reviewed changes

ranqiu added 2 commits November 14, 2017 19:10

Update annotations of layers.py according to comments

de2bc5d

Update doc of layers.py

ac29d00

lcy-seso approved these changes Nov 15, 2017

View reviewed changes

lcy-seso merged commit 374e168 into PaddlePaddle:develop Nov 15, 2017

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update the annotations of layers.py #5517

Update the annotations of layers.py #5517

ranqiu92 commented Nov 9, 2017

lcy-seso left a comment

lcy-seso Nov 14, 2017

ranqiu92 Nov 14, 2017

lcy-seso Nov 14, 2017

ranqiu92 Nov 14, 2017

lcy-seso Nov 14, 2017

ranqiu92 Nov 14, 2017

lcy-seso Nov 14, 2017

lcy-seso Nov 14, 2017

ranqiu92 Nov 14, 2017

lcy-seso Nov 14, 2017

ranqiu92 Nov 14, 2017

lcy-seso Nov 14, 2017

ranqiu92 Nov 14, 2017

lcy-seso Nov 14, 2017

lcy-seso Nov 14, 2017

ranqiu92 Nov 14, 2017

lcy-seso Nov 14, 2017

ranqiu92 Nov 14, 2017

lcy-seso left a comment

Update the annotations of layers.py #5517

Update the annotations of layers.py #5517

Conversation

ranqiu92 commented Nov 9, 2017

lcy-seso left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lcy-seso left a comment

Choose a reason for hiding this comment