SpatialCLIP: Learning 3D-aware Image Representations from Spatially Discriminative Language
Open paper
Benchmark
Model
Rank
Results
visual-question-answering-on-mm-vet
Spatial-LLaVA-7B (extra data)
–
GPT-4 score: 30.7
visual-question-answering-vqa-on-realworldqa
Spatial-LLaVA-7B (extra data)
–
Accuracy: 54.2
Rank counts only results with a code link.