nl8590687
|
9cf877924d
|
纠正数据集标签的一些错误
|
2018-05-23 11:20:14 +08:00 |
nl8590687
|
5d153ceff1
|
添加model25,并且纠正数据集中的一些错误
|
2018-05-22 22:09:43 +08:00 |
nl8590687
|
0f629a27de
|
纠正一些数据标注错误
|
2018-05-22 14:20:48 +08:00 |
nl8590687
|
0bd0d0fd85
|
update readme and log
|
2018-05-21 19:11:22 +08:00 |
nl8590687
|
563d718fa2
|
修复dev集的数据路径错误
|
2018-05-21 16:25:45 +08:00 |
nl8590687
|
8e97019992
|
做了一些小修改
|
2018-05-21 12:48:54 +08:00 |
nl8590687
|
60710270e3
|
紧急修复train.wav.lst中部分路径不存在的问题
|
2018-05-20 19:46:45 +08:00 |
nl8590687
|
ec320960a3
|
删掉无用的代码,并做一些优化
|
2018-05-17 21:57:57 +08:00 |
nl8590687
|
a4734b8516
|
使thchs30的数据列表适用于2018年新版数据集格式,并切换特征提取算法
|
2018-05-15 21:43:11 +08:00 |
nl8590687
|
2685bbe32b
|
add train and test python3 script and modify readme
|
2018-05-11 18:42:58 +08:00 |
nl8590687
|
fb89aabbad
|
fix bugs and improve asrserver
|
2018-05-11 16:56:59 +08:00 |
nl8590687
|
08f7033d21
|
fix bugs and do some test
|
2018-05-08 11:35:18 +08:00 |
nl8590687
|
bf489b9a1d
|
fix bugs and modify dropout and update language model
|
2018-05-05 13:41:45 +08:00 |
nl8590687
|
e2cf387404
|
fix bugs
|
2018-05-02 11:51:44 +08:00 |
nl8590687
|
6609bfaa3d
|
add data set lists and labels and add model 24 and dropout for m22
|
2018-05-01 22:16:01 +08:00 |
nl8590687
|
a43b871e9e
|
fix bug and wish it wont boom again~
|
2018-04-28 16:52:07 +08:00 |
nl8590687
|
a12ff4d1ca
|
fix bugs
|
2018-04-28 16:25:21 +08:00 |
nl8590687
|
e95985d456
|
fix bug
|
2018-04-28 15:59:32 +08:00 |
nl8590687
|
09607561cc
|
fix bugs
|
2018-04-28 15:48:15 +08:00 |
nl8590687
|
1750e33d94
|
fix a lot of bugs
|
2018-04-28 14:44:57 +08:00 |
nl8590687
|
4f4fbff240
|
修复了内存泄漏的bug,添加和修改了部分数据
|
2018-04-28 14:10:30 +08:00 |
nl8590687
|
7ee1d8a811
|
fix a great deal bugs
|
2018-04-26 12:10:34 +08:00 |
nl8590687
|
1de67a6647
|
modify asrserver
|
2018-04-25 21:55:34 +08:00 |
nl8590687
|
7a862a540b
|
move m2 m21 to trash and test asrserver
|
2018-04-25 21:18:28 +08:00 |
nl8590687
|
23ba3a2592
|
补充语言模型和端到端测试脚本
|
2018-04-25 19:50:32 +08:00 |
nl8590687
|
ef4f2fbbcd
|
add tf gpu config and fix bug
|
2018-04-24 13:55:40 +08:00 |
nl8590687
|
6731c9c60a
|
add language model and server and fix bugs
|
2018-04-21 20:50:53 +08:00 |
nl8590687
|
338db8b7c3
|
fix bugs
|
2018-04-18 17:32:40 +08:00 |
nl8590687
|
a99247df09
|
add model 22, fix bug and cut some time cost
|
2018-04-16 22:05:01 +08:00 |
nl8590687
|
ab9a5b56d5
|
add dict.txt
|
2018-04-15 18:17:56 +08:00 |
nl8590687
|
a5a4ae89c0
|
fix bugs
|
2018-04-13 18:22:07 +08:00 |
nl8590687
|
f79b7389a0
|
try a new model
|
2018-04-13 17:23:38 +08:00 |
nl8590687
|
84aa72e917
|
fix bug
|
2018-04-13 15:45:44 +08:00 |
nl8590687
|
861b2aa7a5
|
move model 1 3 4 5 to trash
|
2018-04-13 15:08:21 +08:00 |
nl8590687
|
970d09c7bf
|
fix bugs
|
2018-04-12 22:05:00 +08:00 |
nl8590687
|
a528bbd449
|
fix bugs
|
2018-04-12 15:29:56 +08:00 |
nl8590687
|
69e20f9867
|
fix bugs and add update model2 codes
|
2018-04-12 15:13:24 +08:00 |
nl8590687
|
2b082b878d
|
modify model 1
|
2018-04-12 09:31:32 +08:00 |
nl8590687
|
bd142eaeb1
|
fix bug
|
2018-04-11 22:28:08 +08:00 |
nl8590687
|
aa6a8053f9
|
add test model function
|
2018-04-11 18:05:57 +08:00 |
nl8590687
|
e4104f091a
|
can not parallel and move it to trash
|
2018-04-11 16:46:46 +08:00 |
nl8590687
|
8175f3c8c7
|
try to parallel
|
2018-04-11 16:33:15 +08:00 |
nl8590687
|
c5996dc3a8
|
优化程序结构,以及减少IO次数
|
2018-04-11 15:05:01 +08:00 |
nl8590687
|
49de23dbae
|
add data preprocessing
|
2018-04-10 15:29:33 +08:00 |
nl8590687
|
cb69b5f798
|
a little change
|
2018-04-10 14:41:59 +08:00 |
nl8590687
|
60340059d2
|
add ctc decode function
|
2018-04-09 21:58:20 +08:00 |
nl8590687
|
7eeb5ade96
|
Merge branch 'master' of github.com:nl8590687/ASRT_SpeechRecognition
|
2018-04-09 17:18:25 +08:00 |
nl8590687
|
69c3f33101
|
modify model 5 to try
|
2018-04-09 17:17:41 +08:00 |
nl8590687
|
a6f563bced
|
Merge pull request #6 from ZJUGuoShuai/fix-dataset-link
Update invalid data set download link
|
2018-04-09 02:42:21 -05:00 |
Guo
|
6aaac30ec4
|
Update invalid data set download link
The old THCHS30 download link is invalid now and moved to a new mirror.
|
2018-04-09 09:27:25 +08:00 |