PyTorch Internal Error
See original GitHub issueDescription
What steps will reproduce the problem?
I have created a virtual environment and installed PyTorch library there. When I execute the statements
import torch
x = torch.rand(3)
the code works, and output is displayed but this internal error keeps popping up. I don’t know what’s the issue.
Traceback
Exception in comms call get_namespace_view:
File "D:\Anaconda\lib\site-packages\spyder_kernels\comms\commbase.py", line 314, in _comm_message
buffer = cloudpickle.loads(msg['buffers'][0],
ModuleNotFoundError: No module named 'torch'
Versions
- Spyder version: 5.1.5
- Python version: 3.9.12
- Qt version: 5.9.7
- PyQt5 version: 5.9.2
- Operating System: Windows 10
Dependencies
# Mandatory:
atomicwrites >=1.2.0 : 1.4.0 (OK)
chardet >=2.0.0 : 4.0.0 (OK)
cloudpickle >=0.5.0 : 2.0.0 (OK)
cookiecutter >=1.6.0 : 1.7.3 (OK)
diff_match_patch >=20181111 : 20200713 (OK)
intervaltree >=3.0.2 : 3.1.0 (OK)
IPython >=7.6.0 : 8.2.0 (OK)
jedi >=0.17.2;<0.19.0 : 0.18.1 (OK)
jsonschema >=3.2.0 : 4.4.0 (OK)
keyring >=17.0.0 : 23.4.0 (OK)
nbconvert >=4.0 : 6.4.4 (OK)
numpydoc >=0.6.0 : 1.2 (OK)
paramiko >=2.4.0 : 2.8.1 (OK)
parso >=0.7.0;<0.9.0 : 0.8.3 (OK)
pexpect >=4.4.0 : 4.8.0 (OK)
pickleshare >=0.4 : 0.7.5 (OK)
psutil >=5.3 : 5.8.0 (OK)
pygments >=2.0 : 2.11.2 (OK)
pylint >=2.5.0;<2.10.0 : 2.9.6 (OK)
pyls_spyder >=0.4.0 : 0.4.0 (OK)
pylsp >=1.2.2;<1.3.0 : 1.2.4 (OK)
pylsp_black >=1.0.0 : None (OK)
qdarkstyle =3.0.2 : 3.0.2 (OK)
qstylizer >=0.1.10 : 0.1.10 (OK)
qtawesome >=1.0.2 : 1.0.3 (OK)
qtconsole >=5.1.0 : 5.3.0 (OK)
qtpy >=1.5.0 : 2.0.1 (OK)
rtree >=0.9.7 : 0.9.7 (OK)
setuptools >=49.6.0 : 61.2.0 (OK)
sphinx >=0.6.6 : 4.4.0 (OK)
spyder_kernels >=2.1.1;<2.2.0 : 2.1.3 (OK)
textdistance >=4.2.0 : 4.2.1 (OK)
three_merge >=0.1.1 : 0.1.1 (OK)
watchdog >=0.10.3 : 2.1.6 (OK)
zmq >=17 : 22.3.0 (OK)
# Optional:
cython >=0.21 : 0.29.28 (OK)
matplotlib >=2.0.0 : 3.5.1 (OK)
numpy >=1.7 : 1.21.5 (OK)
pandas >=1.1.1 : 1.4.2 (OK)
scipy >=0.17.0 : 1.7.3 (OK)
sympy >=0.7.3 : 1.10.1 (OK)
Issue Analytics
- State:
- Created a year ago
- Comments:5 (3 by maintainers)
Top Results From Across the Web
RuntimeError: NCCL Error 3: internal error - PyTorch Forums
Hi, I am using DataParallel in a SGE enviroment with cudatoolkit 10.1 and NCCL version 2.7. I get the following error
Read more >sparse.mm: CUDA error: internal error when calling ... - GitHub
Consider the following script: A = torch.sparse_coo_tensor( indices=[ [1500, 1505, 1506], [8347, 8347, 8347], ], values = [1., 1., 1.]
Read more >Pytorch Model causing 500 server error in Django App
This is causing my app to return a 500 server error when hitting that URL; I know the model is causing it because...
Read more >Inference in TensorRT error due to F.pad : IShuffleLayer ...
Description I converted an Pytorch model in ONNX. ... Error Code 4: Internal Error (Reshape_172: IShuffleLayer applied to shape tensor must ...
Read more >MSVC 2022 Preview (4) crashes when compiling PyTorch ...
C:\Users\nickkreeger\workspace\pytorch\caffe2\operators\bbox_transform_op.h(37): fatal error C1001: Internal compiler error. ... To work around this problem, try ...
Read more >
Top Related Medium Post
No results found
Top Related StackOverflow Question
No results found
Troubleshoot Live Code
Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start Free
Top Related Reddit Thread
No results found
Top Related Hackernoon Post
No results found
Top Related Tweet
No results found
Top Related Dev.to Post
No results found
Top Related Hashnode Post
No results found
@dalthviz Thank you so much. The wiki helped a lot, and now everything is working perfectly. Thanks once again 😃
Awesome! Glad to hear that you got Spyder working! I will close this issue as duplicate then but if you have more questions let us know in a new issue