-
Notifications
You must be signed in to change notification settings - Fork 27
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
Add posibility to choose python version for module or make it independent from version #16
Comments
Can be fixed locally with one of:
This issue is a feature request for avoiding future issues with python versions. |
You are absolutely right! This should be done with priority! |
We did consider using the |
@amitsingh19975 Yeah, bindings and models as a module is very useful. I guess in the future we can cover your project under gRPC and some REST-like API to use it as a service. I will share my work with this approach in the next week or two, I guess. |
BTW, If you will find a way to create a standalone package of fastLLaMA that can be published to |
That sounds interesting, We can look into it! |
We added a |
We are going to be fixing this by identifying the python version that is used to run the setup.py file and passing flags accordingly to Cmake in the branch "feature/refactor" |
This has been added, will merge to main soon. @yevhen-kalyna could you kindly check the branch and let me know if it is working now! |
@PotatoSpudowski Also, the original text of this issue says: Is the realization of Limited API ready for testing? p.s. Maybe we should do testing and review using PRs? In this case, I can comment on parts of the code.. |
We now have no pybind11 dependency so this can be closed! |
I had an error after importing the module:
After investigation, I found that in the
build
folder, the fileCMakeCache.txt
contains next:And its interesting, because my activated venv using python3.10
Can we make this module independent of the version? I know that it is possible to ensure compatibility with any cpython3.x interpreter by only using the Limited API. This way, the package can be compiled against any interpreter version and used with any other, without the need to recompile. More here here
BTW, this issue can be fixed manually by adding arguments to Cmake with needed python versions (DPYTHON_LIBRARY, DPYTHON_EXECUTABLE etc.).
p.s. Thanks for your work. I wanted to create a subprocess to run the model (😂) but found your repository in time.
The text was updated successfully, but these errors were encountered: