“RNN test”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
Test
Lr讨论 | 贡献
chinese data
第13行: 第13行:
  
 
[[chinese_data_gigword]]
 
[[chinese_data_gigword]]
 
==chinese data==
 
===prepare data===
 
*now data
 
** gigaword: /work2/xingchao/corpus/Chinese_corpus/gigaword
 
** bing parallel corpus:/nfs/disk/work/users/xingchao/bing_dict
 
** baidu:
 
** sougou:
 
* using data
 
** sample gigword about 344M
 
** dict:tencent11w
 
*train set
 
 
{| border="2px"
 
|+ Train Set Environment
 
|-
 
! Parameters  !! hidden !! class !! direct !! bbt !! bptt_block !! threads !!direct-order!!rand_seed!!nwords!!time(min)
 
|-
 
!set1
 
| 320 || 300 || 2000 || 2 || 20 || 1 || 4 || 1 || 10000||3380(56h)
 
|-
 
|}
 

2014年9月28日 (日) 11:25的版本

tool

  • LTSM/RNN training, GPU&deep supported [1]
  • RNNLM: RNN LM toolkit [2]
  • RWTHLM: RNN LTSM toolkit [3]
  • nplm: NN LM, large scale data [4]
  • RNN toolkit from microsoft [5]

paper

Steps

process dict and data

Test

wsj_data

chinese_data_gigword