Difference between revisions of "NLP Status Report 2017-6-5"

From cslt Wiki
Jump to: navigation, search
Line 2: Line 2:
 
!Date !! People !! Last Week !! This Week
 
!Date !! People !! Last Week !! This Week
 
|-
 
|-
| rowspan="6"|2017/5/31
+
| rowspan="6"|2017/6/5
 
|Jiyuan Zhang ||
 
|Jiyuan Zhang ||
 
||  
 
||  
Line 13: Line 13:
 
* big data baseline bleu = '''30.83'''
 
* big data baseline bleu = '''30.83'''
 
* Fixed three embeddings
 
* Fixed three embeddings
{| class="wikitable"
+
  Shrink output vocab from 30000 to 20000 and best result is 31.53
|-
+
  Train the model with 40 batch size and best result until now is 30.63
! alpha
+
! beta
+
! result (bleu)
+
|-
+
| 1
+
| 1
+
| 43.50
+
|-
+
| 4/3
+
| 2/3
+
| 43.58 (w/o retrained)
+
|-
+
| 2/3
+
| 4/3
+
| 41.22 (w/o retrained)
+
|-
+
| 2/3
+
| 4/3
+
| 42.36 (w/ retrained)
+
 
|}
 
|}
* experiments with '''constant''' initialized embedding:
+
* test more checkpoints on model trained with batch = 40
 
+
* train model with reverse output
 
||
 
||
  

Revision as of 05:44, 5 June 2017

Date People Last Week This Week
2017/6/5 Jiyuan Zhang
Aodong LI
  • Small data:
 Only make the English encoder's embedding constant -- 45.98
 Only initialize the English encoder's embedding and then finetune it -- 46.06
 Share the attention mechanism and then directly add them -- 46.20
  • big data baseline bleu = 30.83
  • Fixed three embeddings
 Shrink output vocab from 30000 to 20000 and best result is 31.53
 Train the model with 40 batch size and best result until now is 30.63
  • test more checkpoints on model trained with batch = 40
  • train model with reverse output

||

|- |Shiyue Zhang ||

||

|- |Shipan Ren ||

||

|-


|}