object_detection_torch_vision
object detection using pytorch vision
Input Data
Input data: image + csv
csv must be following format
| filename | width | height | class | xmin | ymin | xmax | ymax |
|---|---|---|---|---|---|---|---|
| abc1.JPG | 1200 | 800 | A | 340 | 748 | 382 | 780 |
| example.JPG | 1200 | 800 | B | 140 | 248 | 182 | 780 |
| abc2.JPG | 1200 | 800 | A | 240 | 648 | 282 | 780 |
Requirement
- torch==1.13.1
- torchvision==0.14.1
- numpy==1.22.4
- pandas==1.1.4
- opencv-python==4.6.0.66
- albumentations==1.3.0
- matplotlib==3.6.2