Tutorial 1
DataFrames at Scale
These notebooks introduce data acquisition, exploration, in- dexing, filtering, aggregation, and feature creation with cuDF. Participants compare familiar CPU-style patterns with GPU- executed alternatives, then extend to Dask-cuDF for distributed workflows, including scheduler setup, task graph inspection, and persist/compute patterns. .
- cuDF basics, data manipulation, filtering, aggregation, CPU-style comparisons
- Polars intro + cuDF Polars GPU engine patterns
- Dask-cuDF cluster setup, task graph, persist/compute, distributed dataframe practice
What you will do
- open and run a Colab-based RAPIDS workflow
- prepare data for GPU-accelerated experimentation
- train and compare models with a reusable notebook pipeline
- explore hyperparameter optimization in a familiar Python workflow