الرئيسية المنصة الوثائق الأخبار الأوراق البحثية الدروس مجموعات البيانات الموسوعة SOTA نماذج LLM لوحة الأداء GPU الفعاليات

العربية

Machine Translation On Wmt2014 German English

المقاييس

BLEU score

النتائج

نتائج أداء النماذج المختلفة على هذا المعيار القياسي

		Paper Title	Repository
Bi-SimCut	35.15	Bi-SimCut: A Simple Strategy for Boosting Neural Machine Translation
BiBERT	34.94	BERT, mBERT, or BiBERT? A Study on Contextualized Embeddings for Neural Machine Translation
SimCut	34.86	Bi-SimCut: A Simple Strategy for Boosting Neural Machine Translation
Mega	33.12	Mega: Moving Average Equipped Gated Attention
CMLM+LAT+4 iterations	32.04	Incorporating a Local Translation Mechanism into Non-autoregressive Translation
MAT+Knee	31.9	Wide-minima Density Hypothesis and the Explore-Exploit Learning Rate Schedule
CNAT	30.75	Non-Autoregressive Translation by Learning Target Categorical Codes
CMLM+LAT+1 iterations	29.91	Incorporating a Local Translation Mechanism into Non-autoregressive Translation
FlowSeq-large (NPD n = 30)	28.29	FlowSeq: Non-Autoregressive Conditional Sequence Generation with Generative Flow
FlowSeq-large (NPD n = 15)	27.71	FlowSeq: Non-Autoregressive Conditional Sequence Generation with Generative Flow
FlowSeq-large (IWD n=15)	27.16	FlowSeq: Non-Autoregressive Conditional Sequence Generation with Generative Flow
Denoising autoencoders (non-autoregressive)	25.43	Deterministic Non-Autoregressive Neural Sequence Modeling by Iterative Refinement
FlowSeq-large	25.4	FlowSeq: Non-Autoregressive Conditional Sequence Generation with Generative Flow
FlowSeq-base	23.36	FlowSeq: Non-Autoregressive Conditional Sequence Generation with Generative Flow
NAT +FT + NPD	23.20	Non-Autoregressive Neural Machine Translation
SMT + iterative backtranslation (unsupervised)	17.43	Unsupervised Statistical Machine Translation

0 of 16 row(s) selected.