ResearchDatasetsCLEVR-styleCLEVR-styleFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsVisual Question AnsweringCLEVR-style synthetic 10 scenes (test)34Accuracy (Exist)2