TASK |
DATASET |
MODEL |
METRIC NAME |
METRIC VALUE |
GLOBAL RANK |
REMOVE |
Abstractive Text Summarization
|
CNN / Daily Mail
|
Transformer
|
ROUGE-1
|
39.50
|
# 50
|
|
Abstractive Text Summarization
|
CNN / Daily Mail
|
Transformer
|
ROUGE-2
|
16.06
|
# 50
|
|
Abstractive Text Summarization
|
CNN / Daily Mail
|
Transformer
|
ROUGE-L
|
36.63
|
# 45
|
|
Text Summarization
|
GigaWord
|
Transformer
|
ROUGE-1
|
37.57
|
# 21
|
|
Text Summarization
|
GigaWord
|
Transformer
|
ROUGE-2
|
18.90
|
# 21
|
|
Text Summarization
|
GigaWord
|
Transformer
|
ROUGE-L
|
34.69
|
# 23
|
|
Machine Translation
|
IWSLT2014 German-English
|
Transformer
|
BLEU score
|
34.44
|
# 26
|
|
Machine Translation
|
IWSLT2015 English-German
|
Transformer
|
BLEU score
|
28.50
|
# 2
|
|
Image-guided Story Ending Generation
|
LSMDC-E
|
Transformer
|
BLEU-1
|
15.35
|
# 3
|
|
Image-guided Story Ending Generation
|
LSMDC-E
|
Transformer
|
BLEU-2
|
4.49
|
# 4
|
|
Image-guided Story Ending Generation
|
LSMDC-E
|
Transformer
|
BLEU-3
|
1.82
|
# 2
|
|
Image-guided Story Ending Generation
|
LSMDC-E
|
Transformer
|
BLEU-4
|
0.76
|
# 2
|
|
Image-guided Story Ending Generation
|
LSMDC-E
|
Transformer
|
METEOR
|
11.43
|
# 3
|
|
Image-guided Story Ending Generation
|
LSMDC-E
|
Transformer
|
CIDEr
|
9.32
|
# 2
|
|
Image-guided Story Ending Generation
|
LSMDC-E
|
Transformer
|
ROUGE-L
|
19.16
|
# 4
|
|
Multimodal Machine Translation
|
Multi30K
|
Transformer
|
BLUE (DE-EN)
|
29.0
|
# 2
|
|
Natural Language Understanding
|
PDP60
|
Subword-level Transformer LM
|
Accuracy
|
58.3
|
# 10
|
|
Constituency Parsing
|
Penn Treebank
|
Transformer
|
F1 score
|
92.7
|
# 21
|
|
Supervised Only 3D Point Cloud Classification
|
ScanObjectNN
|
Transformer
|
Overall Accuracy (PB_T50_RS)
|
77.24
|
# 12
|
|
Supervised Only 3D Point Cloud Classification
|
ScanObjectNN
|
Transformer
|
GFLOPs
|
4.8
|
# 9
|
|
Supervised Only 3D Point Cloud Classification
|
ScanObjectNN
|
Transformer
|
Number of params (M)
|
22.1
|
# 12
|
|
Image-guided Story Ending Generation
|
VIST-E
|
Transformer
|
BLEU-1
|
17.18
|
# 4
|
|
Image-guided Story Ending Generation
|
VIST-E
|
Transformer
|
BLEU-2
|
6.29
|
# 3
|
|
Image-guided Story Ending Generation
|
VIST-E
|
Transformer
|
BLEU-3
|
3.07
|
# 3
|
|
Image-guided Story Ending Generation
|
VIST-E
|
Transformer
|
BLEU-4
|
2.01
|
# 3
|
|
Image-guided Story Ending Generation
|
VIST-E
|
Transformer
|
METEOR
|
6.91
|
# 3
|
|
Image-guided Story Ending Generation
|
VIST-E
|
Transformer
|
CIDEr
|
12.75
|
# 4
|
|
Image-guided Story Ending Generation
|
VIST-E
|
Transformer
|
ROUGE-L
|
18.23
|
# 4
|
|
Coreference Resolution
|
Winograd Schema Challenge
|
Subword-level Transformer LM
|
Accuracy
|
54.1
|
# 73
|
|
Machine Translation
|
WMT2014 English-French
|
Transformer Big
|
BLEU score
|
41.0
|
# 26
|
|
Machine Translation
|
WMT2014 English-French
|
Transformer Big
|
Hardware Burden
|
23G
|
# 1
|
|
Machine Translation
|
WMT2014 English-French
|
Transformer Big
|
Operations per network pass
|
2300000000.0G
|
# 1
|
|
Machine Translation
|
WMT2014 English-French
|
Transformer Base
|
BLEU score
|
38.1
|
# 39
|
|
Machine Translation
|
WMT2014 English-French
|
Transformer Base
|
Hardware Burden
|
23G
|
# 1
|
|
Machine Translation
|
WMT2014 English-French
|
Transformer Base
|
Operations per network pass
|
330000000.0G
|
# 1
|
|
Machine Translation
|
WMT2014 English-German
|
Transformer Base
|
BLEU score
|
27.3
|
# 52
|
|
Machine Translation
|
WMT2014 English-German
|
Transformer Base
|
Operations per network pass
|
330000000.0G
|
# 1
|
|
Machine Translation
|
WMT2014 English-German
|
Transformer Big
|
BLEU score
|
28.4
|
# 44
|
|
Machine Translation
|
WMT2014 English-German
|
Transformer Big
|
Hardware Burden
|
871G
|
# 1
|
|
Machine Translation
|
WMT2014 English-German
|
Transformer Big
|
Operations per network pass
|
2300000000.0G
|
# 1
|
|