Multimodal Spatial Reasoning on EmbSpatial
67.79AccuracySpecFlow
Evaluation Results
| Method | Links | |
|---|---|---|
| SpecFlowFLOPs(G)=17731.50, Lat.(s)=0.44, Mem.(GB)=42.572026.06 | 67.79 | |
| SparseVLMFLOPs(G)=20785.15, Lat.(s)=0.74, Mem.(GB)=51.932026.06 | 63.89 | |
| VoCoTFLOPs(G)=27211.34, Lat.(s)=0.74, Mem.(GB)=61.522026.06 | 59.76 | |
| LightFastVFLOPs(G)=23482.94, Lat.(s)=0.64, Mem.(GB)=55.112026.06 | 58.32 | |
| SoTFLOPs(G)=24784.89, Lat.(s)=0.68, Mem.(GB)=58.052026.06 | 56.14 | |
| PCCoTFLOPs(G)=17719.32, Lat.(s)=0.49, Mem.(GB)=43.842026.06 | 56.13 | |
| CODIFLOPs(G)=18584.17, Lat.(s)=0.44, Mem.(GB)=44.222026.06 | 55.74 | |
| HeimaFLOPs(G)=17607.64, Lat.(s)=0.42, Mem.(GB)=43.782026.06 | 54.66 |