The RAPIDS suite of open source software libraries gives you the freedom to execute end-to-end data science and analytics pipelines entirely on GPUs.
Learn about RAPIDS
Seamlessly scale from GPU workstations to multi-GPU servers and multi-node clusters with Dask.
Learn about Dask
Accelerate your Python data science toolchain with minimal code changes and no new tools to learn.
Learn about our libraries
Increase machine learning model accuracy by iterating on models faster and deploying them more frequently.
Learn about RAPIDS for model optimization
Drastically improve your productivity with more interactive data science tools like XGBoost.
Learn about XGBoost
Learn about accelerated ML with cuML
RAPIDS is an open source project. Supported by NVIDIA, it also relies on Numba, Apache Arrow, and many more open source projects.
Learn about our projects
The RAPIDS data science framework is designed to have a familiar look and feel to data scientist working in Python. Here’s a code snippet where we read in a CSV file and output some descriptive statistics:
import cudf
gdf = cudf.read_csv('path/to/file.csv')
for column in gdf.columns:
print(gdf[column].mean())
Find more details on our Get Started Page
Jump right into a GPU powered RAPIDS notebook, online, with either SageMaker Studio Lab or Colab (currently only supports RAPIDS v21.12):
Modeled after 10 Minutes to Pandas, this is a short introduction to cuDF that is geared mainly for new users.
Go to guide
A short introduction to multi-GPU solutions with a distributed DataFrame via Dask-cuDF.
Go to guide
A GitHub repository with our introductory examples of XGBoost, cuML demos, cuGraph demos, and more.
Go to repo
A second GitHub repository with our extended collection of community contributed notebook examples.
Go to repo
RAPIDS is committed to open source. We strive to release new features on a 2 month cadence with the generalized release schedule below. Learn more on our Release Blogs
RAPIDS is open source licensed under Apache 2.0, spanning multiple projects that range from GPU dataframes to GPU accelerated ML algorithms. It also provides native array_interface support, allowing Apache Arrow data to be pushed to deep learning frameworks.
Learn more
Whether you are new to RAPIDS, looking to help, or are part of the team, learn about our contributing guidelines on our contributing page.
Go to Docs
GitHub / Docs / Change Log
cuDF is a Python GPU DataFrame library (built on the Apache Arrow columnar memory format) for loading, joining, aggregating, filtering, and otherwise manipulating data all in a pandas-like API familiar to data scientists.
GitHub / Docs / Change Log
libcudf is a CUDA C++ library for implementing standard dataframe operations. It is part of the cuDF repository.
GitHub / Docs / Change Log
cuML is a suite of libraries that implement machine learning algorithms and mathematical primitives functions that are compatible with other RAPIDS projects, all in a scikit-learn-like API familiar to data scientists.
GitHub / Docs / Change Log
cuGraph is a GPU accelerated graph analytics library, with functionality like NetworkX, which is seamlessly integrated into the RAPIDS data science platform.
GitHub / Docs / Change Log
cuSignal is a GPU accelerated signal processing library built around a SciPy Signal-like API, CuPy, and custom Numba and CuPy CUDA kernels. cuSignal is written exclusively in Python and demonstrates GPU speeds without a C++ software layer.
GitHub / Docs / Change Log
cuSpatial is an efficient C++ library accelerated on GPUs with Python bindings to enable use by the data science community. cuSpatial provides significant GPU-acceleration to common spatial and spatiotemporal operations such as point-in-polygon tests, distances between trajectories, and trajectory clustering when compared to CPU-based implementations.
GitHub / Docs / Change Log
cuxfilter is a framework to connect web visualizations to GPU accelerated crossfiltering. Inspired by the JavaScript library crossfilter, it enables interactive and super fast multi-dimensional filtering of 100 million+ row tabular datasets via cuDF.
GitHub / Docs / Change Log
Cyber Log Accelerators (CLX), also pronounced “clicks”, provides a collection of RAPIDS examples for security analysts, data scientists, and engineers to quickly get started applying RAPIDS and GPU acceleration to real-world cybersecurity use cases.
GitHub / Docs / Change Log
RAPIDS Memory Manager (RMM) is a central place for all device memory allocations in cuDF (C++ and Python) and other RAPIDS libraries. In addition, it is a replacement allocator for CUDA Device Memory (and CUDA Managed Memory) and a pool allocator to make CUDA device memory allocation / deallocation faster and asynchronous.
GitHub / Docs / Change Log
cuCIM is a an extensible toolkit designed to provide GPU-accelerated I/O, computer vision and image processing primitives for N-Dimensional images with a focus on biomedical imaging. Our API mirrors scikit-image for image manipulation and OpenSlide for image loading.
RAPIDS integrates with Spark SQL and Dask-SQL to accelerate your SQL queries at scale and make GPU acceleration available to an even broader set of users.
Learn more about RAPIDS + Spark SQL
Learn more about RAPIDS + Dask SQL
Dask is an open source project providing advanced parallelism for analytics that enables performance at scale. RAPIDS is actively contributing to Dask, and it integrates with RAPIDS cuDF, XGBoost, and RAPIDS cuML for GPU-accelerated data analytics and machine learning.
Learn more on our Dask page
XGBoost is a well-known gradient boosted decision trees (GBDT) machine learning package used to tackle regression, classification, and ranking problems. The RAPIDS team works closely with the Distributed Machine Learning Common (DMLC) XGBoost organization to upstream code and ensure that all components of the GPU-accelerated analytics ecosystem work together.
Learn more on our XGBoost page
RAPIDS’ GPU accelerated data science tools can be deployed on all of the major clouds, allowing anyone to take advantage of the speed increases and TCO reductions that RAPIDS enables.
Learn more on our cloud page
Plotly’s Dash enables Data Science teams to focus on the data and models, while producing and sharing enterprise-ready analytic apps that sit on top of RAPIDS-accelerated Python dataframes.
Learn more on our Plotly page
Accelerate Hyperparameter Optimization (HPO) in the Cloud. The RAPIDS team works closely with major cloud providers and open source hyperparameter optimization solutions to ensure smooth integration and high performance, regardless of your deployment platform.
Learn more on our HPO page
NVIDIA Merlin is an open source library providing end-to-end GPU-accelerated recommender systems. Merlin leverages RAPIDS cuDF and Dask cuDF for dataframe transformation during ETL and inference, as well as for the optimized dataloaders in TensorFlow, PyTorch or HugeCTR to accelerate deep learning training.
Learn more on our Merlin page
RAPIDS works extremely well in traditional HPC environments where GPUs are often co-located with accelerated networking hardware such as InfiniBand.
Learn more on our HPC page
NVIDIA is bringing RAPIDS to Apache Spark to accelerate ETL workflows with GPUs.
Learn more on the RAPIDS for Apache Spark page
The Medical Open Network for AI (MONAI) has been named by some the PyTorch of healthcare.
RAPIDS cuCIM has been integrated into the MONAI Transforms component to accelerate the data pathology training pipeline on GPU.
Learn more on MONAI latest highlights