Visual Question Answering on A-OKVQA v1.0 (test)
53.36AccuracySelective Decomposition
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Selective DecompositionVQA Model=BLIP-2 13B, Decomposer=FLAN-T5 (T), Decomposer Params=80M2023.10 | 53.36 | 114 | 18 | 14.91 | |
| Selective DecompositionVQA Model=BLIP-2 13B, Decomposer=FLAN-T5 (T), Decomposer Params=250M2023.10 | 53.36 | 35 | 9 | 22.96 | |
| Selective DecompositionVQA Model=BLIP-2 13B, Decomposer=FLAN-T5 (T), Decomposer Params=780M2023.10 | 53.36 | 52 | 14 | 17.93 | |
| Selective DecompositionVQA Model=BLIP-2 13B, Decomposer=FLAN-T5 (T), Decomposer Params=3B2023.10 | 53.36 | 105 | 14 | 17.93 | |
| Selective DecompositionVQA Model=BLIP-2 13B, Decomposer=Galactica (T*), Decomposer Params=7B2023.10 | 53.36 | 148 | 18 | 14.91 | |
| Selective DecompositionVQA Model=BLIP-2 13B, Decomposer=FLAN-T5 (T), Decomposer Params=11B2023.10 | 53.36 | 166 | 25 | 10.95 | |
| Selective DecompositionVQA Model=BLIP-2 13B, Decomposer=BLIP-2 (I+T), Decomposer Params=11B2023.10 | 53.36 | 192 | 25 | 10.95 | |
| Selective DecompositionVQA Model=BLIP-2 3B, Decomposer=FLAN-T5 (T), Decomposer Params=80M2023.10 | 49.69 | 79 | 16 | 13.57 | |
| Selective DecompositionVQA Model=BLIP-2 3B, Decomposer=FLAN-T5 (T), Decomposer Params=250M2023.10 | 49.69 | 44 | 11 | 17.58 | |
| Selective DecompositionVQA Model=BLIP-2 3B, Decomposer=FLAN-T5 (T), Decomposer Params=780M2023.10 | 49.69 | 87 | 18 | 12.42 | |
| Selective DecompositionVQA Model=BLIP-2 3B, Decomposer=FLAN-T5 (T), Decomposer Params=3B2023.10 | 49.69 | 114 | 31 | 8.13 | |
| Selective DecompositionVQA Model=BLIP-2 3B, Decomposer=BLIP-2 (I+T), Decomposer Params=3B2023.10 | 49.69 | 122 | 33 | 7.71 | |
| Selective DecompositionVQA Model=BLIP-2 3B, Decomposer=Galactica (T*), Decomposer Params=7B2023.10 | 49.69 | 175 | 31 | 8.13 | |
| Selective DecompositionVQA Model=BLIP-2 3B, Decomposer=FLAN-T5 (T), Decomposer Params=11B2023.10 | 49.69 | 70 | 11 | 17.58 |