Lue Fan
Lue Fan
Sorry that I missed your issue for such a long time. I am exploring another direction. Can you make sure your box annotations has the same definition with WOD? The...
V2还在期刊审稿中。你这个投会议和期刊都行吧。
SST did not train on Argo2. Only FSD and FSDv2 are trained on Argo2. I recommend using the FSDv2, which is easier to use. You can find the config here:...
Hello, thanks for your issue but it is your misunderstanding. We convert Argoverse 2 data into KITTI format. In KITTI format, the 'training' folder actually contains **both** the train split...
It depends on the domain gap between your data and the pre-training data. I personally believe tuning all parameters is better.
It is hard to estimate the domain gap without experiments. My experience is only using xyz as input without intensity is good for narrowing the domain gap. Different sensors have...
Hi, sorry for the late reply. You can follow this issue and our instructions in docs to use CTRL. https://github.com/tusen-ai/SST/issues/161 And our FSD detectors can automatically generate bin file after...