Natural Language Inference On Wnli

Metrics

Accuracy

Results

Performance results of various models on this benchmark

Model Name	Accuracy	Paper Title	Repository
ALBERT	91.8	ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
HNNensemble	89	A Hybrid Neural Network Model for Commonsense Reasoning
StructBERTRoBERTa ensemble	89.7	StructBERT: Incorporating Language Structures into Pre-training for Deep Language Understanding	-
SqueezeBERT	65.1	SqueezeBERT: What can computer vision teach NLP about efficient neural networks?
XLNet	92.5	XLNet: Generalized Autoregressive Pretraining for Language Understanding
T5-Base 220M	78.8	Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
BERT-large 340M (fine-tuned on WSCR)	71.9	A Surprisingly Robust Trick for Winograd Schema Challenge
RoBERTa (ensemble)	89	RoBERTa: A Robustly Optimized BERT Pretraining Approach
HNN	83.6	A Hybrid Neural Network Model for Commonsense Reasoning
DistilBERT 66M	44.4	DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter
FLAN 137B (few-shot, k=4)	70.4	Finetuned Language Models Are Zero-Shot Learners
ERNIE 2.0 Large	67.8	ERNIE 2.0: A Continual Pre-training Framework for Language Understanding
T5-Large 770M	85.6	Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
BERTwiki 340M (fine-tuned on WSCR)	74.7	A Surprisingly Robust Trick for Winograd Schema Challenge
FLAN 137B (zero-shot)	74.6	Finetuned Language Models Are Zero-Shot Learners
T5-XL 3B	89.7	Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
DeBERTa	94.5	DeBERTa: Decoding-enhanced BERT with Disentangled Attention
RWKV-4-Raven-14B	49.3	RWKV: Reinventing RNNs for the Transformer Era
Turing NLR v5 XXL 5.4B (fine-tuned)	95.9	-	-
T5-Small 60M	69.2	Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer

0 of 23 row(s) selected.