-
Notifications
You must be signed in to change notification settings - Fork 17
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
Installation instructions #30
Comments
Fully agree with you on the Python version and about QSARtuna dependencies in general. |
@marco-chimfarm You can try the following. Create a conda environment with python 3.10, install QSARtuna and optuna with pip, install REINVENT into the same environment and then reinstall scikit-learn in the version QSARtuna requires (at the moment 1.5.1). I have not tested this much, so other packages, not in the right version, could cause problems: chemprop, protobuf, pandas (pip tells you which versions are "incompatible" with QSARtuna). EDIT: I have a successful run with REINVENT using the Qptuna scoring component (and some others). But I can't tell you if QSARtuna would still be fully functional. |
I have tried that as well, but when I install Reinvent with the requirements file it will uninstall and overwrite QSARtuna dependencies and everything is broken again. I also just need the Qptuna scoring component, I don't need the full QSARtuna functionalities (at least when associated with reinvent) |
I have installed as below which works with my setup, also on an AMD based machine. You would need to be more specific as to what is broken and how that manifests itself. conda create -n test python=3.10
conda activate test
cd QSARtuna
git pull
pip install . # after fixing the python version in pyproject.toml
pip install optuna
cd REINVENT4
git pull
pip install -r requirements-linux-64.lock # using torch 2.4.1
pip install --no-deps .
pip install scikit-learn==1.5.1 |
I think the installation instructions in the README should be presented earlier and tagged more clearly.
It is unclear to me why the Python version has been tied down to the exact patch level. That doesn't make sense.
Also, the instructions refer to a very specific QSARtuna version. I would expect this to be valid for at least the major version of the software.
I had to install optuna manually for some reason. Probably an issue on my side.
The text was updated successfully, but these errors were encountered: