Natural Language Inference On Wnli

Accuracy

평가 결과

이 벤치마크에서 각 모델의 성능 결과

비교 표

모델 이름	Accuracy
albert-a-lite-bert-for-self-supervised	91.8
a-hybrid-neural-network-model-for-commonsense	89
structbert-incorporating-language-structures	89.7
squeezebert-what-can-computer-vision-teach	65.1
xlnet-generalized-autoregressive-pretraining	92.5
exploring-the-limits-of-transfer-learning	78.8
a-surprisingly-robust-trick-for-winograd	71.9
roberta-a-robustly-optimized-bert-pretraining	89
a-hybrid-neural-network-model-for-commonsense	83.6
distilbert-a-distilled-version-of-bert	44.4
finetuned-language-models-are-zero-shot	70.4
ernie-20-a-continual-pre-training-framework	67.8
exploring-the-limits-of-transfer-learning	85.6
a-surprisingly-robust-trick-for-winograd	74.7
finetuned-language-models-are-zero-shot	74.6
exploring-the-limits-of-transfer-learning	89.7
deberta-decoding-enhanced-bert-with	94.5
rwkv-reinventing-rnns-for-the-transformer-era	49.3
모델 19	95.9
exploring-the-limits-of-transfer-learning	69.2
bert-pre-training-of-deep-bidirectional	65.1
a-surprisingly-robust-trick-for-winograd	70.5
exploring-the-limits-of-transfer-learning	93.2