Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
wmsofts
Paddle
提交
8c416653
P
Paddle
项目概览
wmsofts
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
通知
1
Star
0
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
0
列表
看板
标记
里程碑
合并请求
0
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
P
Paddle
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
0
Issue
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
Pages
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
体验新版 GitCode,发现更多精彩内容 >>
未验证
提交
8c416653
编写于
12月 08, 2022
作者:
L
liu zhengxi
提交者:
GitHub
12月 08, 2022
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
refine bsd doc (#48882)
上级
93acef3b
变更
1
隐藏空白更改
内联
并排
Showing
1 changed file
with
8 addition
and
9 deletion
+8
-9
python/paddle/nn/decode.py
python/paddle/nn/decode.py
+8
-9
未找到文件。
python/paddle/nn/decode.py
浏览文件 @
8c416653
...
@@ -26,8 +26,6 @@ __all__ = []
...
@@ -26,8 +26,6 @@ __all__ = []
class
Decoder
:
class
Decoder
:
"""
"""
:api_attr: Static Graph
Decoder is the base class for any decoder instance used in `dynamic_decode`.
Decoder is the base class for any decoder instance used in `dynamic_decode`.
It provides interface for output generation for one time step, which can be
It provides interface for output generation for one time step, which can be
used to generate sequences.
used to generate sequences.
...
@@ -146,13 +144,14 @@ class BeamSearchDecoder(Decoder):
...
@@ -146,13 +144,14 @@ class BeamSearchDecoder(Decoder):
Please refer to `Beam search <https://en.wikipedia.org/wiki/Beam_search>`_
Please refer to `Beam search <https://en.wikipedia.org/wiki/Beam_search>`_
for more details.
for more details.
**NOTE** When decoding with beam search, the `inputs` and `states` of cell
Note:
would be tiled to `beam_size` (unsqueeze and tile), resulting to shapes like
When decoding with beam search, the `inputs` and `states` of cell
`[batch_size * beam_size, ...]` , which is built into `BeamSearchDecoder` and
would be tiled to `beam_size` (unsqueeze and tile), resulting to shapes like
done automatically. Thus any other tensor with shape `[batch_size, ...]` used
`[batch_size * beam_size, ...]` , which is built into `BeamSearchDecoder` and
in `cell.call` needs to be tiled manually first, which can be completed by using
done automatically. Thus any other tensor with shape `[batch_size, ...]` used
:code:`BeamSearchDecoder.tile_beam_merge_with_batch` . The most common case
in `cell.call` needs to be tiled manually first, which can be completed by using
for this is the encoder output in attention mechanism.
:code:`BeamSearchDecoder.tile_beam_merge_with_batch` . The most common case
for this is the encoder output in attention mechanism.
Returns:
Returns:
BeamSearchDecoder: An instance of decoder which can be used in
\
BeamSearchDecoder: An instance of decoder which can be used in
\
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录