Clark Haynes

You can grab the corners of each cuboid using the provided functions from the devkit, and then use code like the following to actually plot them (via whatever library you...

It seems like the codebase may have changed since the pretrained weights were saved. I found the following (manually changed) key names gets things to work: ``` pretrained = torch.load('/mnt/drive/MyDrive/Datasets/SMOKE/model_final.pth')['model']...

I _think_ I can get this working by simply having two KFs, never predicting on one of them, transferring state between them, then calling update individually based upon different sensor...