sdk
Universal Python SDK to run AI workloads on Kubernetes
File Explorer
Download Latest Version (.zip)- bug_report.yaml
- config.yml
- feature_request.yaml
- compare_versions.py
- extract_version.py
- README.md
- test_scripts.py
- update_overrides.py
- check-owners.yaml
- check-pr-title.yaml
- cleanup-overrides.yaml
- docs.yaml
- gh-workflow-approve.yaml
- github-stale.yaml
- osv-scanner.yaml
- release.yaml
- test-e2e.yaml
- test-python.yaml
- test-spark-examples.yaml
- validate-lockfile.yaml
- welcome-new-contributors.yaml
- copilot-instructions.md
- dependabot.yml
- PULL_REQUEST_TEMPLATE.md
- CHANGELOG-0.1.md
- CHANGELOG-0.2.md
- CHANGELOG-0.3.md
- CHANGELOG-0.4.md
- CHANGELOG-0.5.md
- kubeflow-sdk.drawio.svg
- custom.css
- kubeflow-icon.svg
- page.html
- index.rst
- installation.rst
- local-development.rst
- quickstart.rst
- api.rst
- index.rst
- configuration.rst
- index.rst
- tools.rst
- algorithms.rst
- api.rst
- index.rst
- search-space.rst
- api.rst
- index.rst
- api.rst
- batch-jobs.rst
- index.rst
- lifecycle.rst
- options.rst
- sessions.rst
- api.rst
- custom-training.rst
- distributed.rst
- index.rst
- initializers.rst
- options.rst
- runtimes.rst
- conf.py
- contributing.rst
- examples.rst
- index.rst
- batch_failed_job.py
- batch_func_job_lifecycle.py
- batch_job_lifecycle.py
- batch_job_options.py
- connect_existing_session.py
- demo_existing_sparkconnect.py
- README.md
- spark-connect-local-testing.md
- spark_advanced_options.py
- spark_connect_simple.py
- spark_job.py
- test_connect_url.py
- README.md
- sparkoperator.k8s.io_sparkconnects.yaml
- Dockerfile.spark-e2e-runner
- e2e-setup-cluster.sh
- kind-config.yaml
- __init__.py
- constants.py
- types.py
- utils.py
- utils_test.py
- __init__.py
- model_registry_client.py
- model_registry_client_test.py
- __init__.py
- types.py
- __init__.py
- OWNERS
- storage.py
- storage_test.py
- __init__.py
- optimizer_client.py
- __init__.py
- backend.py
- backend_test.py
- utils.py
- utils_test.py
- __init__.py
- base.py
- __init__.py
- constants.py
- __init__.py
- algorithm_types.py
- optimization_types.py
- search_types.py
- __init__.py
- __init__.py
- pipelines_client.py
- pipelines_client_test.py
- __init__.py
- __init__.py
- spark_client.py
- spark_client_test.py
- __init__.py
- backend.py
- backend_test.py
- constants.py
- utils.py
- utils_test.py
- __init__.py
- base.py
- __init__.py
- kubernetes.py
- kubernetes_test.py
- __init__.py
- common.py
- __init__.py
- __init__.py
- types.py
- types_test.py
- __init__.py
- __init__.py
- trainer_client.py
- trainer_client_test.py
- base.py
- docker.py
- podman.py
- backend.py
- backend_test.py
- runtime_loader.py
- runtime_loader_test.py
- types.py
- utils.py
- utils_test.py
- __init__.py
- backend.py
- backend_test.py
- constants.py
- utils.py
- utils_test.py
- __init__.py
- backend.py
- backend_test.py
- constants.py
- job.py
- types.py
- utils.py
- __init__.py
- base.py
- __init__.py
- constants.py
- __init__.py
- common.py
- kubernetes.py
- kubernetes_test.py
- localprocess.py
- common.py
- __init__.py
- types.py
- types_test.py
- __init__.py
- __init__.py
- py.typed
- kep.yaml
- README.md
- kep.yaml
- README.md
- detailed-workflow.svg
- high-level-arch.svg
- README.md
- optimization-job.drawio.svg
- README.md
- __init__.py
- cluster_watcher.py
- README.md
- run_in_cluster.py
- test_spark_examples.py
- __init__.py
- __init__.py
- .gitignore
- .pre-commit-config.yaml
- .readthedocs.yaml
- ADOPTERS.md
- AGENTS.md
- CLAUDE.md
- cliff.toml
- CONTRIBUTING.md
- LICENSE
- Makefile
- OWNERS
- pyproject.toml
- README.md
- RELEASE.md
- ROADMAP.md
- uv.lock
# Installation Guide
git clone https://github.com/kubeflow/sdk
Downloads the entire project code from GitHub to your computer.
cd sdk
Moves into the project folder you just downloaded.
2. Official Install Script
Easy Recommended- Python 3 Python is required to use pip.
pip install -U kubeflow
Installs the package published on PyPI directly โ no need to clone the source.
pip install 'kubeflow[spark]'
Installs the package published on PyPI directly โ no need to clone the source.
pip install 'kubeflow[hub]'
Installs the package published on PyPI directly โ no need to clone the source.
pip install 'kubeflow[pipelines]'
Installs the package published on PyPI directly โ no need to clone the source.
Pulled directly from this repo's README.
3. Docker
Easy- Git Needed to download the project code from GitHub.
- Docker Desktop Needed to build and run containers. Install it and keep it running in the background.
docker build -f hack/Dockerfile.spark-e2e-runner -t sdk .
Builds a runnable image based on the Dockerfile.
docker run -p 8080:80 sdk
Runs the built image as an actual container.
4. Python
Easypip install -U kubeflow
Installs the package published on PyPI directly โ no need to clone the source.
pip install 'kubeflow[spark]'
Installs the package published on PyPI directly โ no need to clone the source.
pip install 'kubeflow[hub]'
Installs the package published on PyPI directly โ no need to clone the source.
pip install 'kubeflow[pipelines]'
Installs the package published on PyPI directly โ no need to clone the source.
Pulled directly from this repo's README.
5. Make
Medium- Git Needed to download the project code from GitHub.
- Make Usually pre-installed on Linux/macOS. On Windows, install separately (e.g. via MSYS2 or WSL).
make
Compiles the code based on the generated build configuration to produce an executable.