ML workload patterns with Ray Train

Pre-install all requirements

!pip install -r ../requirements.txt

These tutorials demonstrate common ML workload patterns with Ray Train. They're independent and can be taken in any order.

• Vision workloads - Computer vision with Food-101 dataset
• Tabular workloads - XGBoost with CoverType dataset
• Time series workloads - Transformer model for New York City taxi demand
• Generative computer vision workloads - Mini diffusion pipeline
• Policy learning workloads - Diffusion-policy on Pendulum-v1
• Recommendation system workloads - Matrix factorization with MovieLens