Visual Question Answering on ToS-TextVQA
46.02Accuracy (Avg)Upper-bound (Q-Former)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Upper-bound (Q-Former)# Trainable Params=163.82 M2026.06 | 46.02 | — | — | |
| Upper-bound (PA)# Trainable Params=21.74 M2026.06 | 41.05 | — | — | |
| ECA# Trainable Params=21.74 M2026.06 | 38.13 | 2.36 | 19.3 | |
| LwF# Trainable Params=21.74 M2026.06 | 32.92 | -0.53 | 15.52 | |
| Vanilla (Q-Former)# Trainable Params=163.82 M2026.06 | 32.13 | -1.37 | 15.03 | |
| MoE-LoRA# Trainable Params=195.71 M2026.06 | 31.76 | -3.2 | 12.4 | |
| EWC# Trainable Params=21.74 M2026.06 | 30.21 | 0.54 | 11.67 | |
| CODA-Prompt# Trainable Params=22.13 M2026.06 | 26.13 | 1.35 | 8.83 | |
| Dual-Prompt# Trainable Params=19.82 M2026.06 | 25.64 | 1.75 | 8.14 | |
| Vanilla (PA)# Trainable Params=21.74 M2026.06 | 24.94 | -3.66 | 10.85 | |
| ECA# Trainable Params=31.51 M2026.06 | 22.61 | 10.21 | 9.27 | |
| EWC# Trainable Params=31.51 M2026.06 | 21.59 | 9.83 | 8.81 | |
| Vanilla (PA)# Trainable Params=31.51 M2026.06 | 17.1 | 5.3 | 7.27 | |
| ZeroShot# Trainable Params=0 M2026.06 | 14.83 | — | — | |
| MoE-LoRA# Trainable Params=208.05 M2026.06 | 8.04 | -3.33 | 6.02 | |
| LwF# Trainable Params=31.51 M2026.06 | 6.13 | -2.17 | 4.82 | |
| ModalPrompt# Trainable Params=19.96 M2026.06 | 4.06 | -2.62 | 4.26 | |
| ZeroShot# Trainable Params=0 M2026.06 | 2.49 | — | — |