FileNotFoundError - caffe2_detectron_ops.dll on Windows source build if Python 3.8 used #35803
Comments
|
Looking into it. |
|
I cannot reproduce the error locally. Could you please use Process Monitor to capture the logs of Windows API calls of |
|
Hi, I'm uploading logs for system python.exe and virtualenv python.exe below. |
|
I didn't build this with Cuda since I just wanted to test if importing is working or not. Should I build with Cuda as well and then post logs? |
You may resolve this by doing |
Yes my test environment didn't have intel-openmp and having intel-openmp has fixed the issue, thank you. |
|
I have installed PyTorch 1.6 on windows7 64 bits. Trying to train OpenNMT, getting this error from init.py |
|
@girishnjha Have you installed |
|
i installed intel-openmp. Still getting the same error |
|
@girishnjha Could you please provide the log using Process Monitor or the Debugging Tools for Windows? |
|
@girishnjha |
This worked for me. Thanks a lot! |
It's working. Thanks! |
|
If adding vcruntime related dlls does not fix this for you, consider deleting At first I was building without CUDA, then with CUDA. Because I was using the same conda environment, this lead me having both Deleting |
|
Thank you @skyline75489, Deleting caffe2_detectron_ops.dll worked for me. |
|
Thanks @skyline75489, this worked for me! |
|
@skyline75489, used it either, it made the problem disappear, but got another error (I'm trying to install Nvidia Apex packet and this problem appeared in that process): any clues how to fix it ? `(GAIN2) C:\Users\Student1\PycharmProjects\GAIN2\apex\apex>pip install -v --disable-pip-version-check --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" ./ Source in c:\users\student1\appdata\local\temp\2\pip-req-build-6rzo82d9 has version 0.1, which satisfies requirement apex==0.1 from file:///C:/Users/Student1/PycharmProjects/GAIN2/apex/apex ERROR: Command errored out with exit status 1: 'C:\Users\Student1\anaconda3\envs\GAIN2\python.exe' -u -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'C:\Users\Student1\AppData\Local\Temp\2\pip-req-build-6rzo82d9\setup.py'"'"'; file='"'"'C:\Users\Student1\AppData\Local\Temp\2\pip-req-build-6rzo82d9\setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(file);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, file, '"'"'exec'"'"'))' --cpp_ext --cuda_ext install --record 'C:\Users\Student1\AppData\Local\Temp\2\pip-record-k20gx55k\install-record.txt' --single-version-externally-managed --compile --install-headers 'C:\Users\Student1\anaconda3\envs\GAIN2\Include\apex' Check the logs for full command output. During handling of the above exception, another exception occurred: Traceback (most recent call last): |
|
thank you evey body |
|
Python version : 3.8.8 UPDATE: SOLVED: |
|
WOW!!! Thank you @skyline75489, it worked for me,new bee! |
|
|
That solves the loading problem thanks! |
|
I delete this file without downloads anything solved this problem! |
|
Just provide another possible workaround: I still bumped into the same error in the latest source code after building PyTorch with |
|
I was getting this error no matter which python version I used (3.7 or 3.8). The problem was that in my home directory I had a This was a bit nasty to debug, because the Anyway, after deleting .condarc and cleaning the package cache (with I hope it helps! P.S.
|
|
Thanks @skyline75489 , I deleted |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi, I'm getting this dll import errors when I'm trying to import torch since 1.4 if I use Python 3.8 for build. The error below is from source build in a virtualenv that I tried today. I can build and import it just fine if I use Python 3.7
Environment
The text was updated successfully, but these errors were encountered: