Text-to-Visualization on nvBench Clean
94.76Overall AccuracyDataVisT5
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DataVisT5Mode=Victim, Test Set=Clean2024.10 | 94.76 | 99.19 | 98.65 | 91.68 | — | |
| DataVisT5Mode=Original, Test Set=Clean2024.10 | 94.25 | 99.16 | 98.39 | 91.29 | 99.75 | |
| CodeT5Mode=Victim, Test Set=Clean2024.10 | 92.29 | 98.88 | 97.21 | 90.17 | — | |
| CodeT5Mode=Original, Test Set=Clean2024.10 | 89.88 | 99.23 | 95.66 | 88.5 | — | |
| RGVisNetMode=Victim, Test Set=Clean2024.10 | 84.82 | 96.92 | 96.18 | 83.67 | — | |
| RGVisNetMode=Original, Test Set=Clean2024.10 | 83.08 | 94.64 | 96.21 | 88.15 | — | |
| ncNetMode=Victim, Test Set=Clean2024.10 | 79.7 | 99.03 | 96.43 | 77.16 | — | |
| ncNetMode=Original, Test Set=Clean2024.10 | 79.57 | 98.77 | 95.82 | 77.29 | — | |
| TransformerMode=Original, Test Set=Clean2024.10 | 71.86 | 98.81 | 90.84 | 74.11 | — | |
| TransformerMode=Victim, Test Set=Clean2024.10 | 71.19 | 98.9 | 91.39 | 73.08 | — | |
| Seq2VisMode=Original, Test Set=Clean2024.10 | 66.85 | 97.59 | 85.06 | 72.41 | — | |
| Seq2VisMode=Victim, Test Set=Clean2024.10 | 66.85 | 97.94 | 83.39 | 73.05 | — |