Brady Zhou
Brady Zhou
Can you test out with smaller batch size, like 32 or 16 and see if those OOM?
In the train loop, try changing ``` optim.zero_grad() loss_mean.backward() optim.step() ``` to ``` loss_mean.backward() optim.step() optim.zero_grad() ```
@rohanb2018 could be related to https://github.com/carla-simulator/carla/issues/2132
one way I got around this previously was just to remove the torchvision import, since torchvision is only used for the `ToTensor` transform, and replace all the `ToTensor` with the...
note - another easier way to get around this is to just find the first instance of `import carla` and simply add `import torchvision` right before that many libraries have...
At the very very top of benchmark_agent can you try import torch import torchvision
Do you mean adding an additional label to regress to the orientation of objects in the scene?
Derrick is correct - all of the models in this work were trained for single classes for closer comparison to prior works like Lift-Splat and FIERY
Sorry for the delay - try this [notebook](https://github.com/bradyz/cross_view_transformers/blob/master/scripts/stitch.ipynb) to visualize merge predictions from two models
Thank you! the code for this is messy so contact me via email and I'll send the implementation!