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
PaddlePaddle
DeepSpeech
提交
44408e52
D
DeepSpeech
项目概览
PaddlePaddle
/
DeepSpeech
9 个月 前同步成功
通知
200
Star
8425
Fork
1598
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
245
列表
看板
标记
里程碑
合并请求
3
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
D
DeepSpeech
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
245
Issue
245
列表
看板
标记
里程碑
合并请求
3
合并请求
3
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
前往新版Gitcode,体验更适合开发者的 AI 搜索 >>
提交
44408e52
编写于
1月 14, 2022
作者:
J
Junkun
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
sync the variable name to others
上级
f866059b
变更
4
隐藏空白更改
内联
并排
Showing
4 changed file
with
9 addition
and
9 deletion
+9
-9
examples/ted_en_zh/st0/conf/tuning/decode.yaml
examples/ted_en_zh/st0/conf/tuning/decode.yaml
+1
-1
examples/ted_en_zh/st1/conf/tuning/decode.yaml
examples/ted_en_zh/st1/conf/tuning/decode.yaml
+1
-1
paddlespeech/s2t/exps/u2_st/model.py
paddlespeech/s2t/exps/u2_st/model.py
+2
-2
paddlespeech/s2t/models/u2_st/u2_st.py
paddlespeech/s2t/models/u2_st/u2_st.py
+5
-5
未找到文件。
examples/ted_en_zh/st0/conf/tuning/decode.yaml
浏览文件 @
44408e52
...
...
@@ -3,7 +3,7 @@ error_rate_type: char-bleu
decoding_method
:
fullsentence
# 'fullsentence', 'simultaneous'
beam_size
:
10
word_reward
:
0.7
maxlen
_
ratio
:
0.3
maxlenratio
:
0.3
decoding_chunk_size
:
-1
# decoding chunk size. Defaults to -1.
# <0: for decoding, use full chunk.
# >0: for decoding, use fixed chunk size as set.
...
...
examples/ted_en_zh/st1/conf/tuning/decode.yaml
浏览文件 @
44408e52
...
...
@@ -4,7 +4,7 @@ error_rate_type: char-bleu
decoding_method
:
fullsentence
# 'fullsentence', 'simultaneous'
beam_size
:
10
word_reward
:
0.7
maxlen
_
ratio
:
0.3
maxlenratio
:
0.3
decoding_chunk_size
:
-1
# decoding chunk size. Defaults to -1.
# <0: for decoding, use full chunk.
# >0: for decoding, use fixed chunk size as set.
...
...
paddlespeech/s2t/exps/u2_st/model.py
浏览文件 @
44408e52
...
...
@@ -408,7 +408,7 @@ class U2STTester(U2STTrainer):
decoding_method
=
decode_cfg
.
decoding_method
,
beam_size
=
decode_cfg
.
beam_size
,
word_reward
=
decode_cfg
.
word_reward
,
maxlen
_ratio
=
decode_cfg
.
maxlen_
ratio
,
maxlen
ratio
=
decode_cfg
.
maxlen
ratio
,
decoding_chunk_size
=
decode_cfg
.
decoding_chunk_size
,
num_decoding_left_chunks
=
decode_cfg
.
num_decoding_left_chunks
,
simulate_streaming
=
decode_cfg
.
simulate_streaming
)
...
...
@@ -436,7 +436,7 @@ class U2STTester(U2STTrainer):
decoding_method
=
decode_cfg
.
decoding_method
,
beam_size
=
decode_cfg
.
beam_size
,
word_reward
=
decode_cfg
.
word_reward
,
maxlen
_ratio
=
decode_cfg
.
maxlen_
ratio
,
maxlen
ratio
=
decode_cfg
.
maxlen
ratio
,
decoding_chunk_size
=
decode_cfg
.
decoding_chunk_size
,
num_decoding_left_chunks
=
decode_cfg
.
num_decoding_left_chunks
,
simulate_streaming
=
decode_cfg
.
simulate_streaming
)
...
...
paddlespeech/s2t/models/u2_st/u2_st.py
浏览文件 @
44408e52
...
...
@@ -264,7 +264,7 @@ class U2STBaseModel(nn.Layer):
speech_lengths
:
paddle
.
Tensor
,
beam_size
:
int
=
10
,
word_reward
:
float
=
0.0
,
maxlen
_
ratio
:
float
=
0.5
,
maxlenratio
:
float
=
0.5
,
decoding_chunk_size
:
int
=-
1
,
num_decoding_left_chunks
:
int
=-
1
,
simulate_streaming
:
bool
=
False
,
)
->
paddle
.
Tensor
:
...
...
@@ -274,7 +274,7 @@ class U2STBaseModel(nn.Layer):
speech_length (paddle.Tensor): (batch, )
beam_size (int): beam size for beam search
word_reward (float): word reward used in beam search
maxlen
_
ratio (float): max length ratio to bound the length of translated text
maxlenratio (float): max length ratio to bound the length of translated text
decoding_chunk_size (int): decoding chunk for dynamic chunk
trained model.
<0: for decoding, use full chunk.
...
...
@@ -297,7 +297,7 @@ class U2STBaseModel(nn.Layer):
num_decoding_left_chunks
,
simulate_streaming
)
# (B, maxlen, encoder_dim)
maxlen
=
max
(
int
(
encoder_out
.
shape
[
1
]
*
maxlen
_
ratio
),
5
)
maxlen
=
max
(
int
(
encoder_out
.
shape
[
1
]
*
maxlenratio
),
5
)
hyp
=
{
"score"
:
0.0
,
"yseq"
:
[
self
.
sos
],
"cache"
:
None
}
hyps
=
[
hyp
]
...
...
@@ -475,7 +475,7 @@ class U2STBaseModel(nn.Layer):
decoding_method
:
str
,
beam_size
:
int
,
word_reward
:
float
=
0.0
,
maxlen
_
ratio
:
float
=
0.5
,
maxlenratio
:
float
=
0.5
,
decoding_chunk_size
:
int
=-
1
,
num_decoding_left_chunks
:
int
=-
1
,
simulate_streaming
:
bool
=
False
):
...
...
@@ -511,7 +511,7 @@ class U2STBaseModel(nn.Layer):
feats_lengths
,
beam_size
=
beam_size
,
word_reward
=
word_reward
,
maxlen
_ratio
=
maxlen_
ratio
,
maxlen
ratio
=
maxlen
ratio
,
decoding_chunk_size
=
decoding_chunk_size
,
num_decoding_left_chunks
=
num_decoding_left_chunks
,
simulate_streaming
=
simulate_streaming
)
...
...
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