I am trying to get the pretrained faster rcnn model to work, but I keep running into this error. /opt/conda/conda-bld/pytorch_1616554793803/work/aten/src/ATen/native/cuda/IndexKernel.cu:142: operator(): block: [0,0,0], thread: [0,0,0] Assertion index >= -sizes[i] && index < sizes[i] && "index out …

Right, as per @ptrblck 's advice, you could try upgrading: [image] RuntimeError: CUDA error: device-side assert triggered when i tried TorchVision Object Detection Finetuning Tutorial vision Thank you Piotr, I seem to be able to use fasterrcnn now with pytorch 1.9.0.dev202…

This almost always means that the shape of your model output (should be batchx num categories) does not match your value range for targets (should be 0 to num categories - 1). Best regards Thomas

Hi Thomas, thank you for your reply. I don’t quite understand what you mean. Are you talking about the loss dictionary that is returned by the model() call? What do you mean with batchx num categories? How do I check the shape of the output, since the call is canceled before it returns. The finetu…

CUDA assertion error when calculating loss values

vision

tom (Thomas V) July 26, 2021, 4:52pm 4

Right, as per @ptrblck 's advice, you could try upgrading: