SOTA Machine Translation on IWSLT En-De 2014 (test) and PapersWithCode

37.85BLEU

Baseline + relaxed cross attention

Updated 1mo ago

Evaluation Results

Method	Links
Baseline + relaxed cross attention 2022.09		37.85
Baseline + relaxed self-attention 2022.09		37.74
Baseline + relaxed self-attention 2022.09		37.71
Baseline + smoothed focus 2022.09		37.67
Baseline + relaxed self-attention 2022.09		37.67
Baseline + relaxed self-attention 2022.09		37.67
Baseline + relaxed cross attention 2022.09		37.64
Baseline (Shen et al. resimulated) 2022.09		37.62
Shen et al. 2022.09		37.6
Baseline + relaxed cross attention 2022.09		37.6
Baseline + relaxed cross attention 2022.09		37.57
Baseline + relaxed self-attention 2022.09		37.57
Baseline + relaxed cross attention 2022.09		37.56
Baseline + relaxed cross attention 2022.09		37.53
Baseline + smoothed focus 2022.09		37.52
Baseline + relaxed self-attention 2022.09		37.49
Baseline (Shen et al. resimulated) 2022.09		37.42
Baseline (Shen et al. resimulated) 2022.09		37.42
Baseline + smoothed focus 2022.09		37.42
Liang et al. 2022.09		37.25
Wu et al. [59] 2022.09		36.88
Admin 2021.02		36.1
GradInit 2021.02		36.1
GradInit 2021.02		36
Admin 2021.02		35.7
Standard 2021.02		35.6
GradInit 2021.02		35.6
T-FixUp 2021.02		35.5
Wu et al. [58] 2022.09		35.2
FixUp 2021.02		34.5
Fan et al. 2022.09		34.5
Vaswani et al. 2022.09		34.4
Admin 2021.02		33.7
Bi-SimCut 2022.06		31.16
Document-level + BERT 2020.02		31.02
DRDA 2024.06		30.92
DRDA 2024.06		30.84
R-Drop 2022.06		30.73
CipherDAug - 2 keys 2022.04		30.65
Data Diverse 2022.04		30.47
Sentence-level + BERT 2020.02		30.45
BiBERT 2022.06		30.45
UniDrop 2021.04		29.99
UniDrop 2022.06		29.99
MixReps+co-teaching 2021.04		29.93
Mixed Rep 2022.06		29.93
MAT 2021.04		29.9
Mixed-Repr. 2022.04		29.71
VAT 2022.06		29.45
NPRF enc-dec w/ RPE 2021.06		29.4
Standard enc + PRF dec 2021.06		29.3
WordDropout 2022.04		29.21
Standard enc-dec 2021.06		29.2
CipherDAug - 1 key 2022.04		29.14
RAML+SwitchOut 2022.04		29.11
RAML 2022.04		29.07
Transformer 2024.06		29.03
SwitchOut 2022.04		29
Our Document-level 2020.02		28.9
Transformer-base+Self-Explaining 2020.12		28.9
Transformer w/ LABO 2023.05		28.8
RAML+WordDropout 2022.04		28.78
Transformer 2022.06		28.7
Transformer 2021.04		28.67
BPE-Dropout 2022.04		28.63
Sentence-level baseline 2020.02		28.57
Transformer 2022.04		28.57
Transformer-base 2020.12		28.4
Transformer w/ LS 2023.05		28.3
Transformer** 2018.08		28.23
Transformer** 2018.08		28.07
Pervasive Attention 2018.08		27.99
Miculicich et al. (2018) 2020.02		27.94
Transformer w/ CP 2023.05		27.9
Transformer w/ AFL 2023.05		27.9
Transformer 2023.05		27.8
PRF enc-dec 2021.06		27.6
Pervasive Attention 2018.08		27.21
Transformer w/ FL 2023.05		26.8
ConvS2S** (MLE) 2018.08		26.73
CeMAT 2022.03		26.7
NPMT + language model 2018.08		25.36
RNNsearch* 2018.08		25.04
RIMs 2019.09		24.23
Transformer 2022.03		23.9
mRASP 2022.03		23.9
RIMs 2019.09		23.71
Transformer 2019.09		22.92
Transformer 2019.09		22.89
Mask-Predict 2022.03		22
LSTM 2019.09		21.32
LSTM 2019.09		20.37