ML TUKE¶
Platform for machine learning and AI¶
Kubeflow on TUKE infrastructure. Train models, run experiments and deploy AI solutions without worrying about hardware.
-
Creating a notebook
Sign in to Kubeflow and start your environment.
-
Storage and files
Data volumes and the file browser.
-
Global shared dataset
The shared folder for lab datasets.
-
Create Image
Create your own Docker image for your projects.
-
FAQ
Frequently asked questions about ML TUKE.
Why ML TUKE?¶
- Kubeflow Platform
-
Complete ML ecosystem - from data preparation to model deployment.
- Persistent Storage
-
A 60 GB quota for storage volumes; data survives notebook shutdown.
- Ready Environments
-
Jupyter, RStudio, VS Code ready for immediate use.
- Free
-
For TUKE students and employees without any fees.
Development environments¶
-
Jupyter Notebook
- Python, data analysis
- Visualization and prototyping
- TensorFlow, PyTorch, scikit-learn
-
RStudio
- Statistical computing
- Working with data in R
- Analytical tasks
-
VS Code
- Universal development environment
- Git integration
- Advanced tools and extensions
What can you do with Kubeflow?¶
| Phase | Options |
|---|---|
| Data Preparation | ETL, transformations, dataset validation |
| Training | CPU-based model training, experiment tracking |
| Tuning | Hyperparameter tuning, experiments |
| Deployment | Inference services, production APIs |
| Monitoring | Performance tracking, model lifecycle |
Who is ML TUKE for?¶
-
Students
- Semester and thesis projects
- ML/AI experiments
- Machine learning courses
-
Researchers
- Scientific projects
- Publications and experiments
- Large dataset processing
-
Employees
- Research grants
- Project collaboration
- AI solution development
Get started now¶
-
First Steps
- Log in with TUKE ID
- Create a notebook server
- Choose environment (Jupyter/RStudio/VS Code)
- Start experimenting
-
Need Help?
Check out frequently asked questions.