ResearchDatasetsSTVQASTVQAFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsScene Text-Centric Visual Question AnsweringSTVQA0.759Accuracy20Spatial Visual Question AnsweringSTVQA 7K (val)93Accuracy19Spatial ReasoningSTVQA 300 samples 7k (train)88.5Relative Score13Spatial ReasoningSTVQA-7k (test)79.3Relative Position Accuracy6Visual Question AnsweringSTVQA-7k86.4Relation Acc6