配置环境问题

 1.

Traceback (most recent call last):
  File "/home/dingfei/projects/hush-PCAM/pcam/scripts/eval_ori.py", line 7, in <module>
    import pcam.datasets.modelnet40 as m40
  File "/home/dingfei/projects/PCAM/pcam/datasets/modelnet40.py", line 19, in <module>
    import lightconvpoint.nn as lcp_nn
  File "/home/dingfei/projects/PCAM/lcp/lightconvpoint/nn/__init__.py", line 10, in <module>
    from .search_convpoint import SearchConvPoint
  File "/home/dingfei/projects/PCAM/lcp/lightconvpoint/nn/search_convpoint.py", line 2, in <module>
    import lightconvpoint.knn as nearest_neighbors
ImportError: /home/dingfei/projects/PCAM/lcp/lightconvpoint/knn.cpython-37m-x86_64-linux-gnu.so: undefined symbol: _ZN2at5zerosEN3c108ArrayRefIlEENS0_13TensorOptionsE

更新torch之后torchvision没有跟着更新，报错。

2.不能导入nni.nas.pytorch.utils。

升级到nni==2.4

3.在配置环境中执行了

pip install -e ./

命令的项目，如果更换工作目录，需要重新执行以下这些命令。 

4.IndexError: Caught IndexError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "/home/dingfei/anaconda3/envs/pcam/lib/python3.7/site-packages/torch/utils/data/_utils/worker.py", line 202, in _worker_loop
    data = fetcher.fetch(index)
  File "/home/dingfei/anaconda3/envs/pcam/lib/python3.7/site-packages/torch/utils/data/_utils/fetch.py", line 44, in fetch
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/home/dingfei/anaconda3/envs/pcam/lib/python3.7/site-packages/torch/utils/data/_utils/fetch.py", line 44, in <listcomp>
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/home/dingfei/projects/hush-PCAM/pcam/datasets/threedmatch_dataset_2.py", line 124, in __getitem__
    unique_xyz0_th = xyz0_th[sel0]
IndexError: tensors used as indices must be long, byte or bool tensors

解决：

4.

 安装pytorch_lightning==1.0