How to install scikit-learn on Fedora, CentOS Stream, or Red Hat Enterprise Linux

Scikit-learn depends on compiled scientific Python packages, so Red Hat-family systems need both a suitable interpreter and an isolated environment before model code can import it. DNF provides Python and pip, while pip installs the current scikit-learn wheel and its Python dependencies inside the project environment.

The upstream installation path recommends a virtual environment on Linux to avoid mixing pip packages with distribution-managed Python files. On Red Hat Enterprise Linux 9 and compatible builds, the default python3 command may be too old for recent scikit-learn releases, so installing the versioned python3.12 packages avoids silently pinning the library to an older release.

A completed environment should resolve python from /home/user/sklearn-env, show scikit-learn in that environment, and run a small estimator without import errors. Fedora systems whose default python3 already meets the Python requirement for scikit-learn can substitute python3 for python3.12 consistently.

1. Open a terminal with sudo privileges.
2. Install Python 3.12 and pip from the DNF repositories.

   $ sudo dnf install python3.12 python3.12-pip

   Red Hat Enterprise Linux 9.8 resolves these packages from AppStream. On Fedora, use sudo dnf install python3 python3-pip when python3 --version already returns Python 3.10 or newer. If DNF cannot find the requested interpreter, enable the supported repositories for the host before continuing.

3. Confirm that the versioned interpreter is available.

   $ python3.12 --version
   Python 3.12.13

   Recent scikit-learn releases require Python 3.10 or newer. Fedora users may use python3 when this check returns a supported version.

4. Create a virtual environment for scikit-learn.

   $ python3.12 -m venv ~/sklearn-env

5. Activate the virtual environment.

   $ source ~/sklearn-env/bin/activate

   After activation, python and pip resolve inside /home/user/sklearn-env. Reactivate this environment in each new terminal before running project code.

6. Upgrade pip inside the virtual environment.

   $ python -m pip install --upgrade pip
   Requirement already satisfied: pip in /home/user/sklearn-env/lib64/python3.12/site-packages (23.2.1)
   Collecting pip
   ##### snipped #####
   Successfully installed pip-26.1.2

7. Install scikit-learn into the activated environment.

   $ python -m pip install --upgrade scikit-learn
   Collecting scikit-learn
   ##### snipped #####
   Successfully installed joblib-1.5.3 narwhals-2.22.1 numpy-2.5.0 scikit-learn-1.9.0 scipy-1.18.0 threadpoolctl-3.6.0

   The exact dependency versions change as PyPI publishes new wheels. Keep the install inside the activated environment instead of using sudo pip.

8. Check the installed package location.

   $ python -m pip show scikit-learn
   Name: scikit-learn
   Version: 1.9.0
   Location: /home/user/sklearn-env/lib64/python3.12/site-packages
   Requires: joblib, narwhals, numpy, scipy, threadpoolctl

9. Run a small estimator smoke test.

   $ python - <<'PY'
   from sklearn.datasets import load_iris
   from sklearn.tree import DecisionTreeClassifier
   
   X, y = load_iris(return_X_y=True)
   model = DecisionTreeClassifier(random_state=0).fit(X, y)
   print("classes:", model.classes_.tolist())
   print("training accuracy:", round(model.score(X, y), 3))
   PY
   classes: [0, 1, 2]
   training accuracy: 1.0