pv_rcnn inference error

[tongfengqi]

------------Environment Information-------------
platform:
    Linux-4.18.0-147.5.2.15.h1109.eulerosv2r10.x86_64-x86_64-with-glibc2.31
    gcc (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0
    Python - 3.9.11 (main, Mar 29 2022, 19:08:29)  [GCC 7.5.0]

Science Toolkits:
    cv2 - 4.5.5
    numpy - 1.23.5
    numba - 0.60.0
    pandas - 1.5.1
    pillow - 9.3.0
    skimage - 0.24.0

PaddlePaddle:
    paddle(gpu) - 2.6.1
    paddle3d - 1.0.0
    paddleseg - 2.8.0
    FLAGS_cudnn_deterministic - Not set.
    FLAGS_cudnn_exhaustive_search - Not set.

CUDA:
    cudnn - 8500
    nvcc - Build cuda_11.7.r11.7/compiler.31442593_0

I trained the pv_rcnn model using the official KITTI dataset and the provided file "Paddle3D/config/pv_rcnn/pv_rcnn_005voxel_kitti. yml".
The training and evaluation were normal, but there was an error during inference, as shown below：

Traceback (most recent call last):
  File "/home/ma-user/work/Paddle3D/deploy/pv_rcnn/python/infer.py", line 198, in <module>
    main(args)
  File "/home/ma-user/work/Paddle3D/deploy/pv_rcnn/python/infer.py", line 191, in main
    box3d_lidar, label_preds, scores = run(predictor, points)
  File "/home/ma-user/work/Paddle3D/deploy/pv_rcnn/python/infer.py", line 168, in run
    predictor.run()
ValueError: (InvalidArgument) The axis value should be less than the rank of input, but received axes[0] = 0, rank of input is 0.
  [Hint: Expected axis < in_dims.size(), but received axis:0 >= in_dims.size():0.] (at ../paddle/phi/kernels/funcs/slice_utils.h:45)
  [operator < slice > error]

Where is the problem with my operation and how do I need to modify it？

PS : I can run the training, evaluation, export, and inference of the Centerpoint model in the same environment