Waymo converted to KITTI format.

[Xiangxu-0103]

PR #2065 may cause an error when converting waymo dataset to kitti format.

Error message:

tensorflow.python.framework.errors_impl.InternalError: CUDA runtime implicit initialization on GPU:0 failed. Status: out of memory

When decreasing num_workers to 64, the error can be:

tensorflow.python.framework.errors_impl.InternalError: Attempting to perform BLAS operation using StreamExecutor without BLAS support [Op:BatchMatMulV2]

When removing the lines added in PR #2065, everything is ok.

[ZCMax]

Have you encountered the stuck in torch.einsum when create dataset problem after removing the lines added in PR #2065

[Xiangxu-0103]

I didn't encounter the stuck problem.