search

mit-han-lab / torchsparse

[MICRO'23, MLSys'22] TorchSparse: Efficient Training and Inference Framework for Sparse Convolution on GPUs.
https://torchsparse.mit.edu
MIT License
1.16k stars 132 forks source link

Error: cannot import name 'PointTensor' from 'torchsparse' #248

Closed Amber0913 closed 7 months ago

Amber0913 commented 9 months ago

Is there an existing issue for this?

Current Behavior

Could you provide the wheels of v2.0.0？I encountered an installation error when using this statement pip install --upgrade git+https://github.com/mit-han-lab/torchsparse.git@v2.0.0.

Expected Behavior

No response

Environment

- GCC:
- NVCC:
- PyTorch:
- PyTorch CUDA:
- TorchSparse:

Anything else?

No response

zhijian-liu commented 9 months ago

We don't have pre-built wheels from v2.0.0. Could you provide the detailed installation error?

binaryannamolly commented 5 months ago

I have the same error. Error: cannot import name 'PointTensor' from 'torchsparse'

ImportError: cannot import name 'PointTensor' from 'torchsparse' (.../python3.10/site-packages/torchsparse/init.cpython-310-x86_64-linux-gnu.so)

My environment,

Ubuntu 22.04.2 LTS PyCharm 2023.2.3 (Professional Edition) Python 3.10 torch version 2.0.1+cu118 cuda version 11.8

Torchparse was installed using, python -c "$(curl -fsSL https://raw.githubusercontent.com/mit-han-lab/torchsparse/master/install.py)"

DJNing commented 4 months ago

Same issue here

binaryannamolly commented 4 months ago

@DJNing this worked for me, pip install --upgrade git+https://github.com/mit-han-lab/torchsparse.git@v1.4.0.

See also, https://github.com/mit-han-lab/torchsparse/issues/291#issuecomment-1955712236.