update hapi/text/text.py #60

jinyuKING · 2020-04-27T10:25:40Z

No description provided.

seq_tag

CLAassistant · 2020-04-27T10:25:46Z

All committers have signed the CLA.

guoshengCS · 2020-04-28T03:09:35Z

hapi/text/text.py

+    def forward(self, inputs):
+        hidden = self.h_0
+        res = []
+        for i in range(inputs.shape[1]):


Shape may be dummy for graph mode, for RNN please use RNN api in text.py

Shape may be dummy for graph mode, for RNN please use RNN api in text.py

这个GRU的接口不再写入text.py中了，使用text.py中原有的RNN相关接口搭建网络。

guoshengCS · 2020-04-28T10:09:09Z

hapi/text/text.py

            crf_decode = self.crf_decoding(input=emission, length=lengths)
            return crf_decode, lengths
+
+class SimpleConvPoolLayer(Layer):


As discussed, rename as Conv1DPoolLayer

As discussed, rename as Conv1DPoolLayer

Done. Thanks

hapi/text/text.py

guoshengCS · 2020-04-29T02:42:36Z

hapi/text/text.py

+        return x
+
+
+class SimCNNEncoder(Layer):


如讨论所述，可以重新命名为CNNEncoder

如讨论所述，可以重新命名为CNNEncoder

Done，Thanks

guoshengCS · 2020-04-29T02:44:55Z

hapi/text/text.py

+    def forward(self, input):
+        emb = self.emb_layer(input)
+        emb_reshape = fluid.layers.reshape(
+            emb, shape=[-1, self.channels, self.seq_len, self.hidden_dim])


感觉最好不要使用固定的seq_len，对于不同batch，seq_len应该允许可变

感觉最好不要使用固定的seq_len，对于不同batch，seq_len应该允许可变

已更改

guoshengCS · 2020-04-29T02:46:25Z

hapi/text/text.py

+        bow_emb = fluid.layers.reduce_sum(emb_reshape, dim=1)
+        return bow_emb
+
+class DynamicGRU(fluid.dygraph.Layer):


text.py中已有GRU的接口可以直接使用，对外统一提供一个吧，不然容易引起困惑

text.py中已有GRU的接口可以直接使用，对外统一提供一个吧，不然容易引起困惑

好的。

guoshengCS · 2020-04-29T02:47:47Z

hapi/text/text.py

+        super(SimBOWEncoder, self).__init__()
+        self.dict_size = dict_size
+        self.bow_dim = bow_dim
+        self.seq_len = seq_len


感觉BOWEncoder相对简单些可以直接放在模型中，这里可以不提供，另外这里似乎还是会有seq_len的问题

如果有其他模型需要使用的话这里也可以保留

如果有其他模型需要使用的话这里也可以保留

BOWEncoder只使用了Embedding，感觉由用户自己搭建比较好，不放在text.py中了

guoshengCS · 2020-04-29T02:48:18Z

hapi/text/text.py

+        res = fluid.layers.concat(res, axis=1)
+        return res
+
+class SimGRUEncoder(Layer):


同上，text.py中已有GRU的接口可以直接使用，对外统一提供一个吧

同上，text.py中已有GRU的接口可以直接使用，对外统一提供一个吧

好的

guoshengCS · 2020-04-29T02:49:41Z

hapi/text/text.py

+        gru = fluid.layers.tanh(gru)
+        return gru
+
+class SimLSTMEncoder(Layer):


可以Sim去掉，另外可以参考text.py中已有的GRUEncoder接口提供类似的实现

可以Sim去掉，另外可以参考text.py中已有的GRUEncoder接口提供类似的实现

好的

guoshengCS · 2020-04-29T02:50:00Z

hapi/text/text.py

+        self.emb_dim = emb_dim
+        self.lstm_dim = lstm_dim
+        self.hidden_dim = hidden_dim
+        self.seq_len = seq_len


还请注意下seq_len的问题

还请注意下seq_len的问题

好的。

guoshengCS · 2020-04-29T02:50:46Z

另外还请在新增API时加上文档

jinyuKING · 2020-05-06T10:29:08Z

另外还请在新增API时加上文档

已添加

hapi/text/text.py

guoshengCS · 2020-05-08T05:16:37Z

hapi/text/text.py

+        bias_attr (ParamAttr|bool|None): The parameter attribute for the bias of conv.If it is set to False, no bias will be added to the output units.
+                If it is set to None or one attribute of ParamAttr, conv2d will create ParamAttr as bias_attr. If the Initializer of the bias_attr is not 
+                set, the bias is initialized zero. Default: None.
+        name(str, optional): The default value is None. Normally there is no need for user to set this property. Default: None


没有name这个参数

guoshengCS · 2020-05-08T05:20:21Z

hapi/text/text.py

+
+class Conv1dPoolLayer(Layer):
+    """
+        This interface is used to construct a callable object of the ``Conv1DPoolLayer`` class.The ``Conv1DPoolLayer`` is composed of a ``Conv2D``  and a ``Pool2D``  .


这里感觉不要说如何实现，用了Conv2D和Pool2D。而是描述做了什么，做了Conv1D和Pool1D

guoshengCS · 2020-05-08T05:21:19Z

hapi/text/text.py

+            Variable
+
+    Example:
+    ```python


请参考其他添加示例代码的方法 https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/fluid/dygraph/nn.py#L143

另外，尽量保证行长度不超过80

guoshengCS · 2020-05-08T05:22:40Z

hapi/text/text.py

+
+class CNNEncoder(Layer):
+    """
+        This interface is used to construct a callable object of the ``CNNEncoder`` class.The ``CNNEncoder`` is composed of a ``Embedding``  and a ``Conv1dPoolLayer``  .


这里没有包含Embedding

guoshengCS · 2020-05-08T05:24:01Z

hapi/text/text.py

+class CNNEncoder(Layer):
+    """
+        This interface is used to construct a callable object of the ``CNNEncoder`` class.The ``CNNEncoder`` is composed of a ``Embedding``  and a ``Conv1dPoolLayer``  .
+        For more details, refer to code examples. The ``CNNEncoder`` layer calculates the output based on the input, dict_size and emb_dim, filter_size, num_filters, 


要描述清楚这个接口做了什么，按照这里的描述无法区分这个CNNEncoder是做了多层CNN的堆叠还是多个CNN的并行

guoshengCS · 2020-05-08T05:25:49Z

hapi/text/text.py

+    Return:
+            3-D Tensor, the result of input after  embedding and conv1dPoollayer
+
+    Return Type:


class文档注释可以不用添加return内容

guoshengCS · 2020-05-08T05:26:45Z

hapi/text/text.py

+            use_cudnn=self.use_cudnn
+        ) for i in range(layer_num)])
+
+    def forward(self, input):


为class的method添加文档，说明输入参数和返回内容，上面那个Conv1DPoolLayer类似

guoshengCS · 2020-05-08T05:27:44Z

另外，请使用pre-commit

guoshengCS · 2020-05-11T02:27:02Z

后续可以加入stack的CNNEncoder，以及支持param_attr传入

guoshengCS and others added 4 commits April 27, 2020 03:30

Merge pull request PaddlePaddle#54 from 0YuanZhang0/seq_tag

f3e8f30

seq_tag

Merge branch 'master' of https://github.com/jinyuKING/hapi

2ea7670

update text.py

06fd61f

Merge branch 'master' of https://github.com/jinyuKING/hapi

a4cb497

guoshengCS reviewed Apr 29, 2020

View reviewed changes

guoshengCS mentioned this pull request Apr 29, 2020

add emotion_detection and update senta #62

Open

update text.py

75dcc16

guoshengCS reviewed May 7, 2020

View reviewed changes

hapi/text/text.py Show resolved Hide resolved

update text.py

3e18097

guoshengCS reviewed May 8, 2020

View reviewed changes

update text.py

3b5e6d9

guoshengCS approved these changes May 11, 2020

View reviewed changes

update hapi/text/text.py #60

Are you sure you want to change the base?

update hapi/text/text.py #60

Uh oh!

Conversation

jinyuKING commented Apr 27, 2020

Uh oh!

CLAassistant commented Apr 27, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

guoshengCS commented Apr 29, 2020

Uh oh!

jinyuKING commented May 6, 2020

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

guoshengCS commented May 8, 2020

Uh oh!

guoshengCS commented May 11, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

CLAassistant commented Apr 27, 2020 •

edited

Loading