Closed perrineedham closed 1 month ago
Hi @perrineedham, I'm also seeing this issue. I believe it originates from the fact that there is no key for torch 24 (2.4) in the pytorch, cuda version mapping. If you downgrade your version to 2.0, it should install correctly. Alternatively, maybe it's as simple as adding 2.4 to this dictionary.
From install.py torch_cuda_mapping = dict( [ ("torch19", ["11.1"]), ("torch110", ["11.1", "11.3"]), ("torch111", ["11.3", "11.5"]), ("torch112", ["11.3", "11.6"]), ("torch113", ["11.6", "11.7"]), ("torch20", ["11.7", "11.8"]), ] )
Is there an existing issue for this?
Have you followed all the steps in the FAQ?
Current Behavior
python3 -c "$(curl -fsSL https://raw.githubusercontent.com/mit-han-lab/torch sparse/master/install.py)" Traceback (most recent call last): File "<string>", line 40, in <module> KeyError: 'torch24'
http://pypi.hanlab.ai/simple/torchsparse
I am running with torch 2.4 and cuda 12.4. Can you advise on how to install?
Error Line
Downloading seems to work without a problem when using
FORCE_CUDA=1 pip install --no-cache-dir git+https://github.com/mit-han-lab/torchsparse.git
however, I want to install via wheels for easier integration into our codebase.Environment
Full Error Log
No response