October 5-16, 2020
Data science (DS), machine learning (ML), and deep learning (DL) yield exciting new capabilities for data scientists. But using multiple tools that don’t always work together can add layers of complexity. With RAPIDS, a software accelerator from NVIDIA that speeds up DS workflows with GPUs, scientists can use the same tools for DS, ML, and DL—and deliver results faster.
The NVIDIA Deep Learning Institute (DLI) offers hands-on training in accelerated data science and AI. As a special gift to celebrate JupyterCon 2020, NVIDIA is offering 250 people who register at the following link a free DLI course: Register Now
Jacob Tomlinson NVIDIA Sr. Software Engineer
What Is My GPU Doing? Using PyNVML and the NVDashboard Jupyter Lab Extension to Access GPU Metrics
Brad Miro Google Machine Learning Engineer
Apache Spark 3.0: Big Data Analytics with GPUs and Jupyter Notebooks
Nanthini Balasubramanian NVIDIA Data Scientist
Optimizing Model Performance with Feature Engineering and Hyperparameter Optimisation
Josh Patterson NVIDIA Sr. Director of Engineering
Live interview with Anthony Scopatz. Details to follow
Anima Anandkumar NVIDIA Director of MLResearch
Next-generation frameworks for Large-scale AI
As datasets become larger and more complex, visualization tools that can handle this new scale are key in maintaining manageable workflows. Walk through a generalized data visualization-focused workflow that takes advantage of RAPIDS’s performance and integrated libraries like datashader, bokeh, holoviews, and plotly dash.
The open-source RAPIDS project allows data scientists to GPU-accelerate end-to-end data science workflows. Learn how to use RAPIDS cuDF (GPU-enabled pandas dataframes) and Dask-cuDF to ingest and manipulate large datasets directly on the GPU. Then run machine learning at scale on RAPIDS cuML.
Explore a multi-GPU, RAPIDS/BlazingSQL-based interactive analysis tool with a Jupyter Notebook interface that provides a major speedup for querying and analyzing datasets compared to conventional, persistent databases. This pipeline is the result of an intense collaboration between Oak Ridge National Laboratory, Scripps Research, NVIDIA, and BlazingSQL and represents a new open-science capability for massively parallel drug discovery in response to a public health crisis.
Learn how to leverage well-defined APIs to build custom machine learning estimators that scale using Dask and RAPIDS. Create drop-in replacements for ensembles and pipelines that work with existing estimators from scikit-learn, Dask-ML, XGBoost, and RAPIDS cuML. Learn from Capital One use cases on what’s working.
Learn about our journey of building real-time streaming analytics pipelines for NVIDIA® GeForce NOW™ Cloud Gaming using cuStreamz. Deep dive into the cuStreamz architecture on how we built production-grade streaming features like checkpointing, state management, and accelerated source and sink connectors.
Sign up to receive the latest news from NVIDIA.