3D Question Answering on SQA3D v1.0 (test)
59EM@1PAVE-7B (w/ 3D info)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PAVE-7B (w/ 3D info)Mode=Task-specific, TFLOPS=98.68, Total Params=8.2B, Trainable Params=170.3M2025.03 | 59 | 61.4 | |
| LLaVA-OV-7B-FTMode=Task-specific, TFLOPS=98.53, Total Params=8.2B, Trainable Params=161.5M2025.03 | 55.8 | 58.1 | |
| LLaVA-3D-7BMode=Task-specific2025.03 | 55.6 | 57.6 | |
| Scene-LLM-7BMode=Task-specific2025.03 | 54.2 | — | |
| LEO-7BMode=Task-specific2025.03 | 50 | 52.4 | |
| 3D-LLM-7BMode=Task-specific2025.03 | 49.79 | — | |
| VideoChat2-7BMode=Zero-shot2025.03 | 32.3 | — | |
| LLaVA-OV-7BMode=Zero-shot, TFLOPS=98.53, Total Params=8.2B2025.03 | 8.3 | 50.7 |