May I ask how your OPV2V dataset is processed

[yaoyaoling123]

May I ask if your OPV2V dataset is organized in this way after downloading? I am currently working on
“python opencood/tools/train.py -y None --model_dir opencood/logs/HEAL_m1_based/stage2/m2_alignto_m1”
The error message is as follows: “visibility_map = np.asarray(cv2.cvtColor(cav_contents[0]["bev_visibility.png"], cv2.COLOR_BGR2GRAY))
cv2.error: OpenCV(4.5.1) /tmp/pip-req-build-jhawztrk/opencv/modules/imgproc/src/color.cpp:182: error: (-215:Assertion failed) !_src.empty() in function 'cvtColor”
However, the original OPV2V dataset does not contain the “bev_visibility.png” file. It seems to be present in the OPV2V-H dataset, but the root directory pointed to by the config. yaml file is indeed the OPV2V dataset folder.

[yifanlu0227]

https://ucla.app.box.com/v/UCLA-MobilityLab-OPV2V/folder/279976559690

Hi, the bev_visibility.png is provided by the OPV2V authors in additional-001.zip. You need to download them and extract them together with OPV2V

[yaoyaoling123]

Thank you for your prompt and detailed reply. I have successfully started training according to your steps!