Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Error when installing: TypeError: Class advice impossible in Python3. Use the @implementer class decorator instead. #16

Open
jimfangx opened this issue Feb 3, 2020 · 0 comments

Comments

@jimfangx
Copy link

jimfangx commented Feb 3, 2020

Hi, I am getting TypeError: Class advice impossible in Python3. Use the @implementer class decorator instead. when installing CX_DB8.

After pasting in the install script from the readme, I got a number of errors:

  1. ERROR: Could not find a version that satisfies the requirement tensorflow_gpu==1.14.0 (from -r requirements.txt (line 6)) (from versions: none) ERROR: No matching distribution found for tensorflow_gpu==1.14.0 (from -r requirements.txt (line 6)) -> I fixed this by updating the tensorflow_gpu in the requirements.txt to 1.1.0 then reinstalled all the requirements using pip3
  2. flair 0.4.2 has requirement urllib3<1.25,>=1.20, but you'll have urllib3 1.25.8 which is incompatible. -> Updated flair in requirements.txt to 0.4.5. Then reinstalled requirements
  3. Command "/Library/Frameworks/Python.framework/Versions/3.6/bin/python3.6 -u -c "import setuptools, tokenize;__file__='/private/var/folders/sj/dh5wbtx51wvfns1mclspk8d80000gp/T/pip-install-hfqih3yb/cryptacular/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" install --record /private/var/folders/sj/dh5wbtx51wvfns1mclspk8d80000gp/T/pip-record-cws__7h9/install-record.txt --single-version-externally-managed --compile" failed with error code 2 in /private/var/folders/sj/dh5wbtx51wvfns1mclspk8d80000gp/T/pip-install-hfqih3yb/cryptacular/ got this weird error. I ignored this and this never came back
  4. While trying to downgrade urllib3 to be compatiable with flair, I got: requests 2.22.0 has requirement urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1, but you'll have urllib3 1.20 which is incompatible. -> I replaced it with 1.25.8 using pip3 install urllib3==1.25.8
  5. The requirements install came back without errors but as soon as I tried to run the script using python3 -i cx_db8_flair.py , I got ImportError: dlopen(/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/torch/_C.cpython-36m-darwin.so, 9): Library not loaded: /usr/local/opt/libomp/lib/libomp.dylib Referenced from: /Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/torch/lib/libshm.dylib Reason: image not found -> I did some searching online and installed brew install libomp
  6. Re ran the script and got Class advice impossible in Python3. Use the @implementer class decorator instead.

Seeing this has not worked, I deleted everything and started over. This instance I got flair 0.4.5 has requirement pytest>=5.3.2, but you'll have pytest 4.6.1 which is incompatible. while updating flair. -> I updated it to "5.3.2" in the requirements.txt

Later it turns out that pytorch_transformers was outdated too: so I updated it to 1.2.0

Then I realized apex was extremely out dated: the website is down and the github was last updated like 6 years ago. I think this is the main cause of the series of errors that I am getting.


Below is the full install log:
CX_DB8 Install Log

Below is my requirements.txt with the updated version:
requirements.txt

Issues ref.: #11 pytorch/pytorch#20030
https://github.com/cd34/apex
flairNLP/flair#1236
huggingface/transformers#163

Let me know if you need any other information.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant