YOLO-V3-Kitti
Train YOLO-V3-Kitti
DataSet: Kitti 2d object left color image
- Image download link:https://s3.eu-central-1.amazonaws.com/avg-kitti/data_object_image_2.zip
- label download link:https://s3.eu-central-1.amazonaws.com/avg-kitti/data_object_label_2.zip
Data Split
- Total:7481
- Train:Test:Val: 8:1:1
original class
class_name | count |
---|---|
DontCare | 15782 |
Person_sitting | 222 |
Pedestrian | 0 |
Car | 28742 |
Tram | 511 |
Misc | 973 |
Cyclist | 1627 |
Truck | 1094 |
Van | 2914 |
Custom class
Class name (string in label file) | Class ID (number in database) | comment |
---|---|---|
Car | 0 | |
Truck | 1 | include [Truck, Van, Tram] |
Person | 2 | include [Pedestrian, Person_sitting, cyclist] |
Condition
Parameters | value |
---|---|
batch | 32 |
subdivisions | 16 |
width | 416 |
height | 416 |
channels | 3 |
momentum | 0.9 |
decay | 0.0005 |
filters | 24 |
classes | 3 |
Train log
Path: darknet/backup/
epoch | weight_name |
---|---|
100 | 100.weights |
200 | 200.weights |
300 | 300.weights |
400 | 400.weights |
500 | 500.weights |
600 | 600.weights |
700 | 700.weights |
800 | 800.weights |
900 | 900.weights |
10000 | 10000.weights |
20000 | 20000.weights |
Ref:
yolo v3
If you use YOLOv3 in your work please cite our paper! @article{yolov3, title={YOLOv3: An Incremental Improvement}, author={Redmon, Joseph and Farhadi, Ali}, journal = {arXiv}, year={2018} }
Kitti v3
Citation When using this dataset in your research, we will be happy if you cite us: @INPROCEEDINGS{Geiger2012CVPR, author = {Andreas Geiger and Philip Lenz and Raquel Urtasun}, title = {Are we ready for Autonomous Driving? The KITTI Vision Benchmark Suite}, booktitle = {Conference on Computer Vision and Pattern Recognition (CVPR)}, year = {2012} }