Depth Completion on Inference Latency Benchmark
1,139.89Inference Speed (ms)Depth-Pro
Evaluation results
| Method | Links | |
|---|---|---|
| Depth-ProBatch Size=1, Hardware=RTX 3090 GPU, Optimization=Single-frame2026.06 | 1,139.89 | |
| Pi3Batch Size=1, Hardware=RTX 3090 GPU, Optimization=Single-frame2026.06 | 191.22 | |
| DA2Model Variant=vitl, Batch Size=1, Hardware=RTX 3090 GPU, Optimization=Single-frame2026.06 | 101.47 | |
| DA3Model Variant=large, Batch Size=1, Hardware=RTX 3090 GPU, Optimization=Single-frame2026.06 | 59.73 | |
| AISPOBatch Size=1, Hardware=RTX 3090 GPU, Optimization=Single-frame2026.06 | 28.79 | |
| SwinDRNetBatch Size=1, Hardware=RTX 3090 GPU, Optimization=Single-frame2026.06 | 19.7 | |
| DFNetBatch Size=1, Hardware=RTX 3090 GPU, Optimization=Single-frame2026.06 | 5.19 |