How to install torch-sparse 0.4.3 on google colab ? #206
Comments
|
Any chance you can post the full log? The shown error message is not very informative. |
|
I run this command : I get this error : >>>> Collecting torch_scatter==1.4.0 |
|
I post the full log
Le lun. 21 févr. 2022 à 18:19, Matthias Fey ***@***.***> a
écrit :
… Any chance you can post the full log? The shown error message is not very
informative.
—
Reply to this email directly, view it on GitHub
<#206 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/APFUVEEXXJMQAQX7ZOYANA3U4JX3FANCNFSM5O6VR6XA>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
|
Can you run and send me the log? |
|
|
Done.
Le mar. 22 févr. 2022 à 08:26, Matthias Fey ***@***.***> a
écrit :
… Can you run
pip uninstall torch-sparse
pip uninstall torch-sparse
pip install --verbose torch-sparse==0.4.3
and send me the log?
—
Reply to this email directly, view it on GitHub
<#206 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/APFUVEFW4UTC6AY2CXT4INDU4M3DRANCNFSM5O6VR6XA>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
|
Oh, I see. It cannot find cuSPARSE |
|
nvcc: NVIDIA (R) Cuda compiler driver |
|
Done it's cuda 11.1
it must be 10.1 ?
Le mer. 23 févr. 2022 à 09:54, Matthias Fey ***@***.***> a
écrit :
… Oh, I see. It cannot find cuSPARSE cusparseXcsrgemm routines. Which CUDA
version are you using (nvcc --version)?
—
Reply to this email directly, view it on GitHub
<#206 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/APFUVEA6DJZZATND5UOSMPLU4SOF7ANCNFSM5O6VR6XA>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
|
Yes, I assume that CUDA 11.1 is too new to compile such an old version. |
|
in Colab differents versions are installed , how can i set the version that i want? |
|
As far as I know, you cannot change the CUDA version in a Google Colab (at least Google didn't provide me with any meaningful answers update that). Any reason you want to rely on such an old PyTorch/ |
|
I need to run this code https://github.com/TianBian95/BiGCN
Le ven. 25 févr. 2022 à 10:53, Matthias Fey ***@***.***> a
écrit :
… As far as I know, you cannot change the CUDA version in a Google Colab (at
least Google didn't provide me with any meaningful answers update that).
Any reason you want to rely on such an old PyTorch/torch-sparse version?
—
Reply to this email directly, view it on GitHub
<#206 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/APFUVEFOFPWQ7ACSGDX4UWDU45GTLANCNFSM5O6VR6XA>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
|
I'm pretty sure you can use newer versions of |
* initial commit * update: add LightningNodeData * update * update * uppdate trainer * update * fix test * update * update * pragma: no cover * readd encoder * update * update * tracking * neighborloader support * update test * test validate and test * doc string * update * address review comments * add cfg option * add doc * update * update * update * typo * update * typo * typo * clean up doc * fix test * update * fix * fix test
|
TypeError: coalesce() got an unexpected keyword argument 'fill_value' I'm trying to replicate a code and I get this error on using the latest version of torch Sparse. Is there any support for the coalesce() function in the latest releases ? The CUDA installation in my system doesn't allow me to get a earlier version of torch-sparse. Apparently the function works well in the releases <0.4.4 |
|
This looks to be related to using |
|
I'm using 'torch-scatter==2.0.9', 'torch-sparse==0.6.13' |
|
This issue had no activity for 6 months. It will be closed in 2 weeks unless there is some new activity. Is this issue already resolved? |
Hi,
I developed my model with the environment blow:
torch==1.4.0 numpy==1.18.1
torch_scatter==1.4.0
torch_sparse==0.4.3
torch_cluster==1.4.5
torch_geometric==1.3.2
but i can't install torch-sparse==0.4.3
I try this commands :
!pip install torch-sparse==0.4.3
!pip install /content/pytorch_sparse-0.4.3.tar.gz
!pip install --verbose torch-sparse==0.4.3
error:
Processing ./pytorch_sparse-0.4.3.tar.gz
Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from torch-sparse==0.4.3) (1.4.1)
Requirement already satisfied: numpy>=1.13.3 in /usr/local/lib/python3.7/dist-packages (from scipy->torch-sparse==0.4.3) (1.18.1)
Building wheels for collected packages: torch-sparse
Building wheel for torch-sparse (setup.py) ... error
ERROR: Failed building wheel for torch-sparse
Running setup.py clean for torch-sparse
Failed to build torch-sparse
Installing collected packages: torch-sparse
Running setup.py install for torch-sparse ... error
ERROR: Command errored out with exit status 1: /usr/bin/python3 -u -c 'import io, os, sys, setuptools, tokenize; sys.argv[0] = '"'"'/tmp/pip-req-build-ihjhbd7q/setup.py'"'"'; file='"'"'/tmp/pip-req-build-ihjhbd7q/setup.py'"'"';f = getattr(tokenize, '"'"'open'"'"', open)(file) if os.path.exists(file) else io.StringIO('"'"'from setuptools import setup; setup()'"'"');code = f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, file, '"'"'exec'"'"'))' install --record /tmp/pip-record-c1qmqj51/install-record.txt --single-version-externally-managed --compile --install-headers /usr/local/include/python3.7/torch-sparse Check the logs for full command output.
Processing ./pytorch_scatter-1.4.0.tar.gz
The text was updated successfully, but these errors were encountered: