Ignite616

batch % im2col_step_ == 0INTERNAL ASSERT FAILED

1

good job！ When I train bevformer_tiny on two a800s, samples_per_gpu=16. i got an error `RuntimeError: batch % im2col_step_ == 0INTERNAL ASSERT FAILED at "/tmp/mmcv/mmcv/ops/csrc/pytorch/cuda/ms_deform_attn_cuda.cu":255, please report a bug to PyTorch....

what is the role of using pre-trained models during training？cascade_mask_rcnn_r18.

2

When I train, I use the pre-trained model provided by fast-bev. But there are many warnings, size mismatch. And the training speed is still very slow, which is 1/2 of...

Why use sweeps data？

1

I use fast-bev officially provided tags. I don't have local sweeps data. The data in sweepeps is not labeled, why is it used in the training stage? I wonder it....

Great job. I see that you have provided bbox_head_2d attribute in fastbev.py, how to use this interface? I think bbox_head_2d can be added in config, but I don't know which...