You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
These notebooks provide examples of how to use RAPIDS libraries, like cuML and cuGraph. These notebooks are designed to be self contained with the RAPIDS Docker Container and RAPIDS Nightly Docker Containers and can run on airgapped systems. You can quickly get this continaer using the install guide from the RAPIDS.ai Getting Started page
For additional, community driven notebooks, which will include our blogs, tutorials, workflows, and more intricate examples, please see the Notebooks Contrib Repo. These notebooks use generally use real world data sets.
This notebook shows the acceleration one can gain by using GPUs with XGBoost in RAPIDS.
CuML Notebooks
The cuML notebooks showcase how to use the machine learning algorithms implemented in cuML along with the advantages of using cuML over scikit-learn. These notebooks compare the time required and the performance of the algorithms. Below are a list of such algorithms:
This notebook includes code examples of lasso and elastic net models. These models are placed together so a comparison between the two can also be made in addition to their sklearn equivalent.
This notebook shows how to use the forest inference library to load saved models and perform prediction using them. In addition, it also shows how to perform training and prediction using xgboost and lightgbm models.
This notebook showcases principal component analysis (PCA) algorithm where the model can be used for prediction (using fit_transform) as well as converting the transformed data into the original dataset (using inverse_transform).
In this notebook, T-Distributed Stochastic Neighborhood Embedding is demonstrated applying the Barnes Hut method on the Fashion MNIST dataset using our fit_transform function
This notebook showcases truncated singular value decomposition (tsvd) algorithm which like PCA performs both prediction and transformation of the converted dataset into the original data using fit_transform and inverse_transform functions respectively
The uniform manifold approximation & projection algorithm is compared with the original author's equivalent non-GPU Python implementation using fit and transform functions
Demonstration of cuML uniform manifold approximation & projection algorithm's supervised approach against mortgage dataset and comparison of results against the original author's equivalent non-GPU \Python implementation.
Demostration of UMAP supervised training. Uses a set of labels to perform supervised dimensionality reduction. UMAP can also be trained on datasets with incomplete labels, by using a label of "-1" for unlabeled samples.
This notebook showcases two special methods where cuDF goes beyond the Pandas library: apply_rows and apply_chunk functions. They utilized the Numba library to accelerate the data transformation via GPU in parallel.
This notebook showcases how to use Numba CUDA to accelerate cuDF data transformation and how to step by step accelerate it using CUDA programming tricks
Demonstrate of using the renumbering features to assigned new vertex IDs to the test graph. This is useful for when the data sets is non-contiguous or not integer values
Demonstration of using cuGraph to identify clusters in a test graph using Spectral Clustering using both the (A) Balance Cut and (B) the Modularity Maximization quality metrics