Crowd Counting On Nwpu Crowd Val
Metrics
MAE
RMSE
Results
Performance results of various models on this benchmark
Model Name | MAE | RMSE | Paper Title | Repository |
|---|---|---|---|---|
| CLIP-EBC (ViT-L/14) | 32.3 | 79.7 | CLIP-EBC: CLIP Can Count Accurately through Enhanced Blockwise Classification | |
| DMCount-EBC | 39.6 | 95.8 | CLIP-EBC: CLIP Can Count Accurately through Enhanced Blockwise Classification | |
| CSRNet-EBC | 42.9 | 100.1 | CLIP-EBC: CLIP Can Count Accurately through Enhanced Blockwise Classification | |
| CLIP-EBC (ResNet50) | 38.6 | 90.3 | CLIP-EBC: CLIP Can Count Accurately through Enhanced Blockwise Classification | |
| CLIP-EBC (ViT-B/16) | 36.6 | 81.7 | CLIP-EBC: CLIP Can Count Accurately through Enhanced Blockwise Classification |
0 of 5 row(s) selected.