Table Question Answering on ROBUT-WTQ Column Order Shuffling (test)
89Accuracy (Pre-perturbation)TAPEX
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TAPEXEvaluation Protocol=Fine-tuned2023.06 | 89 | 87.4 | 94.4 | |
| OmniTabEvaluation Protocol=Fine-tuned2023.06 | 87.6 | 85.3 | 90.4 | |
| TAPASEvaluation Protocol=Fine-tuned2023.06 | 85.6 | 83.9 | 93 | |
| TableFormerEvaluation Protocol=Fine-tuned2023.06 | 84.9 | 84.8 | 99.3 | |
| GPT-3Evaluation Protocol=Few-shot, Sample Size=2002023.06 | 77.5 | 76.9 | 92.1 | |
| OmniTabEvaluation Protocol=Fine-tuned2023.06 | 58.4 | 56 | 89.2 | |
| TAPEXEvaluation Protocol=Fine-tuned2023.06 | 54.4 | 48.5 | 81.4 | |
| TableFormerEvaluation Protocol=Fine-tuned2023.06 | 51.2 | 51 | 99.1 | |
| TAPASEvaluation Protocol=Fine-tuned2023.06 | 45.7 | 42.5 | 86.5 | |
| GPT-3Evaluation Protocol=Few-shot2023.06 | 40.9 | 40 | 93.3 |