incorrect annotation

[twangnh]

Branch

master branch https://mmdetection.readthedocs.io/en/latest/

📚 The doc issue

Hi @shaunyuan22 thanks for releasing the data, we have tried to use the original images and annotations for training, however, when visualizing the annotations, the bounding boxes does not corresponds to the objects (traffic light is correct, but vehicle is wrong), the data loading process is same as other COCO format dataset, could you please help?

Suggest a potential alternative/fix

No response

[bestux7]

Branch

master branch https://mmdetection.readthedocs.io/en/latest/

📚 The doc issue

Hi @shaunyuan22 thanks for releasing the data, we have tried to use the original images and annotations for training, however, when visualizing the annotations, the bounding boxes does not corresponds to the objects (traffic light is correct, but vehicle is wrong), the data loading process is same as other COCO format dataset, could you please help?

Suggest a potential alternative/fix

No response

I also find this problems, vehicle is all wrong.