Visual Question Answering On Iconqa
評価指標
Reasoning (Alg.)
Reasoning (Com.)
Reasoning (Cou.)
Reasoning (Est.)
Reasoning (Fra.)
Reasoning (Geo.)
Reasoning (Mea.)
Reasoning (Pat.)
Reasoning (Pro.)
Reasoning (Sce.)
Reasoning (Sen.)
Reasoning (Spa.)
Reasoning (Tim.)
Sub-tasks (Blank)
Sub-tasks (Img.)
Sub-tasks (Txt.)
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
モデル名 | Reasoning (Alg.) | Reasoning (Com.) | Reasoning (Cou.) | Reasoning (Est.) | Reasoning (Fra.) | Reasoning (Geo.) | Reasoning (Mea.) | Reasoning (Pat.) | Reasoning (Pro.) | Reasoning (Sce.) | Reasoning (Sen.) | Reasoning (Spa.) | Reasoning (Tim.) | Sub-tasks (Blank) | Sub-tasks (Img.) | Sub-tasks (Txt.) | Paper Title | Repository |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
DFAF | 50.27 | 81.69 | 70.68 | 99.02 | 77.60 | 81.80 | 98.83 | 56.60 | 85.70 | 67.01 | 84.11 | 51.42 | 67.72 | 78.28 | 77.72 | 72.17 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
Q-Only | 28.02 | 48.19 | 33.63 | 40.46 | 33.06 | 38.03 | 38.07 | 33.66 | 40.76 | 35.37 | 45.25 | 37.14 | 48.09 | 28.45 | 41.64 | 36.86 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
ViLBERT | 50.62 | 75.60 | 71.05 | 99.22 | 74.09 | 80.05 | 99.07 | 62.78 | 70.94 | 58.52 | 81.78 | 49.46 | 66.72 | 77.08 | 76.66 | 70.47 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
I-Only | 31.73 | 45.26 | 37.64 | 62.29 | 32.48 | 38.71 | 64.02 | 36.29 | 37.51 | 35.47 | 45.25 | 37.52 | 47.37 | 46.65 | 41.56 | 36.02 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
Random | 11.12 | 41.20 | 18.38 | 3.62 | 34.84 | 30.30 | 0.36 | 34.81 | 38.81 | 34.25 | 45.16 | 36.49 | 35.82 | 0.29 | 41.70 | 36.87 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
ViLT | 50.55 | 84.95 | 71.13 | 99.02 | 75.81 | 82.61 | 98.91 | 59.22 | 87.65 | 66.72 | 86.10 | 53.38 | 69.99 | 79.27 | 79.67 | 72.69 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
UNITER | 49.18 | 83.67 | 71.01 | 99.41 | 78.37 | 81.31 | 99.38 | 60.81 | 87.84 | 61.25 | 86.10 | 48.34 | 69.77 | 78.53 | 78.71 | 72.39 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
MCAN | 47.32 | 82.73 | 68.94 | 99.08 | 76.20 | 79.86 | 98.99 | 54.79 | 84.87 | 62.49 | 83.25 | 49.70 | 68.00 | 74.52 | 77.36 | 71.25 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
BAN | 47.46 | 82.12 | 67.56 | 97.06 | 73.77 | 79.99 | 96.50 | 55.67 | 82.45 | 66.92 | 82.12 | 53.20 | 66.50 | 75.54 | 76.33 | 70.82 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
Patch-TRM | 56.73 | 87.00 | 77.81 | 98.24 | 82.13 | 81.87 | 97.98 | 68.75 | 95.73 | 62.39 | 92.49 | 55.62 | 77.98 | 83.62 | 82.66 | 75.19 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
ViT | 51.10 | 82.12 | 70.84 | 98.95 | 77.41 | 82.60 | 98.76 | 58.46 | 86.07 | 68.80 | 84.72 | 54.64 | 68.66 | 78.92 | 79.15 | 72.34 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning | |
Top-Down | 50.00 | 80.65 | 65.01 | 99.54 | 72.43 | 80.07 | 99.46 | 55.01 | 83.75 | 58.22 | 84.54 | 45.78 | 68.28 | 73.03 | 75.92 | 68.51 | IconQA: A New Benchmark for Abstract Diagram Understanding and Visual Language Reasoning |
0 of 12 row(s) selected.