Natural Language Inference On Qnli
Metrics
Accuracy
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | Accuracy |
|---|---|
| fnet-mixing-tokens-with-fourier-transforms | 85% |
| informer-transformer-likes-informed-attention | 91.89% |
| nystromformer-a-nystrom-based-algorithm-for | 88.7% |
| q-bert-hessian-based-ultra-low-precision | 93.0 |
| debertav3-improving-deberta-using-electra | 96% |
| charformer-fast-character-transformers-via | 91.0% |
| Model 7 | 95.4% |
| smart-robust-and-efficient-fine-tuning-for | - |
| data2vec-a-general-framework-for-self-1 | 91.1% |
| spanbert-improving-pre-training-by | 94.3% |
| how-to-train-bert-with-an-academic-budget | 90.6 |
| adversarial-self-attention-for-language | 93.6% |
| a-statistical-framework-for-low-bitwidth | 94.5 |
| trans-blstm-transformer-with-bidirectional | 94.08% |
| exploring-the-limits-of-transfer-learning | 90.3% |
| adversarial-self-attention-for-language | 91.4% |
| smart-robust-and-efficient-fine-tuning-for | 99.2% |
| exploring-the-limits-of-transfer-learning | 93.7% |
| exploring-the-limits-of-transfer-learning | 96.7% |
| ernie-enhanced-language-representation-with | 91.3% |
| ernie-20-a-continual-pre-training-framework | 94.6% |
| bert-pre-training-of-deep-bidirectional | 92.7% |
| xlnet-generalized-autoregressive-pretraining | 94.9% |
| ernie-20-a-continual-pre-training-framework | 92.9% |
| roberta-a-robustly-optimized-bert-pretraining | 98.9% |
| 190910351 | 87.7% |
| albert-a-lite-bert-for-self-supervised | 99.2% |
| entailment-as-few-shot-learner | 94.5% |
| deberta-decoding-enhanced-bert-with | 95.3% |
| 190910351 | 90.4% |
| q8bert-quantized-8bit-bert | 93.0 |
| lm-cppf-paraphrasing-guided-data-augmentation | 70.2% |
| big-bird-transformers-for-longer-sequences | 92.2% |
| structbert-incorporating-language-structures | 99.2% |
| smart-robust-and-efficient-fine-tuning-for | 99.2% |
| exploring-the-limits-of-transfer-learning | 94.8% |
| sensebert-driving-some-sense-into-bert | 90.6% |
| llm-int8-8-bit-matrix-multiplication-for | 94.7% |
| distilbert-a-distilled-version-of-bert | 90.2% |
| smart-robust-and-efficient-fine-tuning-for | - |
| squeezebert-what-can-computer-vision-teach | 90.1% |
| exploring-the-limits-of-transfer-learning | 96.3% |
| clear-contrastive-learning-for-sentence | 93.4% |