Table Question Answering on ROBUT-WTQ Row Order Shuffling (test)
88.5Accuracy (Pre-perturbation)TAPEX
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| TAPEXEvaluation Protocol=Fine-tuned2023.06 | 88.5 | 83 | 86.2 | |
| OmniTabEvaluation Protocol=Fine-tuned2023.06 | 87.6 | 82.4 | 87.9 | |
| TableFormerEvaluation Protocol=Fine-tuned2023.06 | 85.7 | 85.2 | 96.9 | |
| TAPASEvaluation Protocol=Fine-tuned2023.06 | 84.8 | 80.1 | 91.1 | |
| GPT-3Evaluation Protocol=Few-shot, Sample Size=2002023.06 | 78.2 | 76.5 | 92.1 | |
| OmniTabEvaluation Protocol=Fine-tuned2023.06 | 60.6 | 51.2 | 77.8 | |
| TAPEXEvaluation Protocol=Fine-tuned2023.06 | 56.9 | 45.7 | 71.7 | |
| TableFormerEvaluation Protocol=Fine-tuned2023.06 | 51 | 50.9 | 97 | |
| TAPASEvaluation Protocol=Fine-tuned2023.06 | 48 | 40.6 | 74.8 | |
| GPT-3Evaluation Protocol=Few-shot2023.06 | 42.9 | 38.5 | 90.2 |