Conference Workshops & Training

Discover how deep learning works through hands-on exercises in computer vision and natural language processing. You’ll train deep learning models from scratch, learning tools and tricks to achieve highly accurate results. You’ll also learn to leverage freely available, state-of-the-art pre-trained models to save time and get your deep learning application up and running quickly.

By participating in this is workshop, you’ll:

> Learn the fundamental techniques and tools required to train a deep learning model
> Gain experience with common deep learning data types and model architectures
> Enhance datasets through data augmentation to improve model accuracy
> Leverage transfer learning between models to achieve efficient results with less data and computation
> Build confidence to take on your own project with a modern deep learning framework

Learn how to identify anomalies and failures in time-series data, estimate the remaining useful life of the corresponding parts, and use this information to map anomalies to failure conditions.

You'll learn how to:

> Leverage predictive maintenance to manage failures and avoid costly unplanned downtimes
> Identify key challenges around identifying anomalies that can lead to costly breakdowns
> Use time-series data to predict outcomes using machine learning classification models with XGBoost
> Apply predictive maintenance procedures by using a long short-term memory ( LSTM)-based model to predict device failure
> Experiment with autoencoders to detect anomalies by using the time-series sequences from the previous steps

Upon completion, you’ll understand how to use AI to predict the condition of equipment and estimate when maintenance should be performed.

The CUDA computing platform enables the acceleration of CPU-only applications to run on the world’s fastest massively parallel GPUs. Experience C/C++ application acceleration by:

> Accelerating CPU-only applications to run their latent parallelism on GPUs
> Utilizing essential CUDA memory management techniques to optimize accelerated applications
> Exposing accelerated application potential for concurrency and exploiting it with CUDA streams
> Leveraging Nsight Systems to guide and check your work

Upon completion, you’ll be able to accelerate and optimize existing C/C++ CPU-only applications using the most essential CUDA techniques and Nsight Systems. You’ll understand an iterative style of CUDA development that will allow you to ship accelerated applications fast.

You’ll learn how to use Transformer-based natural language processing models for text classification tasks, such as categorizing documents. You will also learn how to leverage Transformer-based models for named-entity recognition (NER) tasks and how to analyze various model features, constraints, and characteristics to determine which model is best suited for a particular use case based on metrics, domain specificity, and available resources.

By participating in this workshop, you’ll be able to:

> Understand how text embeddings have rapidly evolved in NLP tasks such as Word2Vec, recurrent neural network (RNN)-based embeddings, and Transformers
> See how Transformer architecture features, especially self-attention, are used to create language models without RNNs
> Use self-supervision to improve the Transformer architecture in BERT, Megatron, and other variants for superior NLP results
> Leverage pre-trained, modern NLP models to solve multiple tasks such as text classification, NER, and question answering
> Manage inference challenges and deploy refined models for live applications

Learn how GPUs change the game for Data Engineering on the hardware level and how to optimize pipelines with CUDA Python Libraries.

Practice GPU based Data Engineering by:

> Surveying different Data Formats and their strengths and weaknesses
> Using statistics to get a "feel" for the data and visualizing it in a dashboard
> Scaling an ETL pipeline to multiple GPUs using NVTabular

Upon completion, learners will be able to make appropriate hardware and software choices to optimize the speed of their data pipelines.

Writing CUDA C++ applications that efficiently and correctly utilize all available GPUs on a node drastically improves performance over single-GPU code, and makes the most cost-effective use out of compute nodes with multiple GPUs. In this workshop you will learn to utilize multiple GPUs on a single node by:

- Learning how to launch kernels on multiple GPUs, each working on a subsection of the required work
- Learning how to use concurrent CUDA Streams to overlap memory copy with computation on multiple GPUs

Upon completion, you will be able to build robust and efficient CUDA C++ applications that can leverage all available GPUs on a single node.

RAPIDS is a collection of data science libraries that allows end-to-end GPU acceleration for data science workflows. In this training, you'll:

> Use cuDF and Dask to ingest and manipulate massive datasets directly on the GPU
> Apply a wide variety of GPU-accelerated machine learning algorithms, including XGBoost, cuGRAPH, and cuML, to perform data analysis at massive scale
> Perform multiple analysis tasks on massive datasets in an effort to stave off a simulated epidemic outbreak affecting the UK

Upon completion, you'll be able to load, manipulate, and analyze data orders of magnitude faster than before, enabling more iteration cycles and drastically improving productivity.

Deep learning-based recommender systems are the secret ingredient behind personalized online experiences and powerful decision support tools in retail, entertainment, healthcare, finance, and other industries. This workshop covers the fundamental tools and techniques for building highly effective recommender systems, as well as how to deploy GPU-accelerated solutions for real-time recommendations.

By participating in this is workshop, you’ll learn how to:

> Build a content-based recommender system using the open-source cuDF library and Apache Arrow
> Construct a collaborative filtering recommender system using alternating least squares (ALS) and CuPy
> Design a wide and deep neural network using TensorFlow 2 to create a hybrid recommender system
> Optimize performance for both training and inference using large, sparse datasets
> Deploy a recommender model as a high-performance web service

You’ll learn how to use Transformer-based natural language processing models for text classification tasks, such as categorizing documents. You will also learn how to leverage Transformer-based models for named-entity recognition (NER) tasks and how to analyze various model features, constraints, and characteristics to determine which model is best suited for a particular use case based on metrics, domain specificity, and available resources.

By participating in this workshop, you’ll be able to:

> Understand how text embeddings have rapidly evolved in NLP tasks such as Word2Vec, recurrent neural network (RNN)-based embeddings, and Transformers
> See how Transformer architecture features, especially self-attention, are used to create language models without RNNs
> Use self-supervision to improve the Transformer architecture in BERT, Megatron, and other variants for superior NLP results
> Leverage pre-trained, modern NLP models to solve multiple tasks such as text classification, NER, and question answering
> Manage inference challenges and deploy refined models for live applications

Deep learning-based recommender systems are the secret ingredient behind personalized online experiences and powerful decision support tools in retail, entertainment, healthcare, finance, and other industries. This workshop covers the fundamental tools and techniques for building highly effective recommender systems, as well as how to deploy GPU-accelerated solutions for real-time recommendations.

By participating in this is workshop, you’ll learn how to:

> Build a content-based recommender system using the open-source cuDF library and Apache Arrow
> Construct a collaborative filtering recommender system using alternating least squares (ALS) and CuPy
> Design a wide and deep neural network using TensorFlow 2 to create a hybrid recommender system
> Optimize performance for both training and inference using large, sparse datasets
> Deploy a recommender model as a high-performance web service

With the increase in traffic cameras, growing prospect of autonomous vehicles, and promising outlook of smart cities, there's a rise in demand for faster and more efficient object detection and tracking models. This involves identification, tracking, segmentation and prediction of different types of objects within video frames.

In this workshop, you’ll learn how to:

- Efficiently process and prepare video feeds using hardware accelerated decoding methods
- Train and evaluate deep learning models and leverage ""transfer learning"" techniques to elevate efficiency and accuracy of these models and mitigate data sparsity issues
- Explore the strategies and trade-offs involved in developing high-quality neural network models to track moving objects in large-scale video datasets
- Optimize and deploy video analytics inference engines by acquiring the DeepStream SDK

Upon completion, you'll be able to design, train, test and deploy building blocks of a hardware-accelerated traffic management system based on parking lot camera feeds.

Learn how to identify anomalies and failures in time-series data, estimate the remaining useful life of the corresponding parts, and use this information to map anomalies to failure conditions.

You'll learn how to:

> Leverage predictive maintenance to manage failures and avoid costly unplanned downtimes
> Identify key challenges around identifying anomalies that can lead to costly breakdowns
> Use time-series data to predict outcomes using machine learning classification models with XGBoost
> Apply predictive maintenance procedures by using a long short-term memory ( LSTM)-based model to predict device failure
> Experiment with autoencoders to detect anomalies by using the time-series sequences from the previous steps

Upon completion, you’ll understand how to use AI to predict the condition of equipment and estimate when maintenance should be performed.

Modern deep learning challenges leverage increasingly larger datasets and more complex models. As a result, significant computational power is required to train models effectively and efficiently. In this course, you will learn how to scale deep learning training to multiple GPUs. Using multiple GPUs for deep learning can significantly shorten the time required to train lots of data, making solving complex problems with deep learning feasible.

This course will teach you how to use multiple GPUs to train neural networks. You'll learn:

- Approaches to multi-GPU training
- Algorithmic and engineering challenges to large-scale training
- Key techniques used to overcome the challenges mentioned above

Upon completion, you'll be able to effectively parallelize training of deep neural networks using Horovod.

You’ll learn how to use Transformer-based natural language processing models for text classification tasks, such as categorizing documents. You will also learn how to leverage Transformer-based models for named-entity recognition (NER) tasks and how to analyze various model features, constraints, and characteristics to determine which model is best suited for a particular use case based on metrics, domain specificity, and available resources.

By participating in this workshop, you’ll be able to:

> Understand how text embeddings have rapidly evolved in NLP tasks such as Word2Vec, recurrent neural network (RNN)-based embeddings, and Transformers
> See how Transformer architecture features, especially self-attention, are used to create language models without RNNs
> Use self-supervision to improve the Transformer architecture in BERT, Megatron, and other variants for superior NLP results
> Leverage pre-trained, modern NLP models to solve multiple tasks such as text classification, NER, and question answering
> Manage inference challenges and deploy refined models for live applications

Modern deep learning challenges leverage increasingly larger datasets and more complex models. As a result, significant computational power is required to train models effectively and efficiently. In this course, you will learn how to scale deep learning training to multiple GPUs. Using multiple GPUs for deep learning can significantly shorten the time required to train lots of data, making solving complex problems with deep learning feasible.

This course will teach you how to use multiple GPUs to train neural networks. You'll learn:

> Approaches to multi-GPU training
> Algorithmic and engineering challenges to large-scale training
> Key techniques used to overcome the challenges mentioned above

Upon completion, you'll be able to effectively parallelize training of deep neural networks using Horovod.

Discover how deep learning works through hands-on exercises in computer vision and natural language processing. You’ll train deep learning models from scratch, learning tools and tricks to achieve highly accurate results. You’ll also learn to leverage freely available, state-of-the-art pre-trained models to save time and get your deep learning application up and running quickly.

By participating in this is workshop, you’ll:

> Learn the fundamental techniques and tools required to train a deep learning model
> Gain experience with common deep learning data types and model architectures
> Enhance datasets through data augmentation to improve model accuracy
> Leverage transfer learning between models to achieve efficient results with less data and computation
> Build confidence to take on your own project with a modern deep learning framework

You’ll learn how to use Transformer-based natural language processing models for text classification tasks, such as categorizing documents. You will also learn how to leverage Transformer-based models for named-entity recognition (NER) tasks and how to analyze various model features, constraints, and characteristics to determine which model is best suited for a particular use case based on metrics, domain specificity, and available resources.

By participating in this workshop, you’ll be able to:
> Understand how text embeddings have rapidly evolved in NLP tasks such as Word2Vec, recurrent neural network (RNN)-based embeddings, and Transformers
> See how Transformer architecture features, especially self-attention, are used to create language models without RNNs
> Use self-supervision to improve the Transformer architecture in BERT, Megatron, and other variants for superior NLP results
> Leverage pre-trained, modern NLP models to solve multiple tasks such as text classification, NER, and question answering
> Manage inference challenges and deploy refined models for live applications

In this workshop you will learn the tools and techniques required to write CUDA C++ applications that can scale efficiently to clusters of NVIDIA GPUs. You will do this by working on code from several CUDA C++ applications in an interactive cloud environment backed by several NVIDIA GPUs. You will gain exposure to a handful of multi-GPU programming methods including CUDA-Aware MPI before proceeding to the main focus of this course, NVSHMEM.

By participating in the workshop you will:

> Learn several methods for writing correct multi-GPU CUDA C++ applications.
> Use a variety of multi-GPU communication patterns and learn to reason about their tradeoffs.
> Write portable, scalable CUDA code with the Single-Process Multiple-Data (SPMD) paradigm using CUDA-aware MPI and NVSHMEM.
> Improve multi-GPU SPMD code with NVSHMEM’s symmetric memory model, and its ability to perform GPU-initiated data transfers.
> Get practice with common multi-GPU coding paradigms like domain decomposition and halo exchanges.
> Receive guidance for how to reason about scaling considerations for a variety of GPU cluster configurations.

Would you like to write and run a script to have your data-center configured in 5 minutes? In this session we will discuss, build and deploy step-by-step Ansible-Playbooks to create a high-performing data-center infrastructure. The outcome will be a state-of-the-art Layer 3 data-center fabric, which will include compute nodes (Ubuntu), simulated hypervisors and provides multi-tenant isolation providing IPv4 and IPv6 connectivity. The infrastructure runs on individual “simulation”/”digital-twins” provided by “air.nvidia.com”.

This training introduces the book Learning Deep Learning (LDL) from NVIDIA Deep Learning Institute (DLI). LDL was published by Pearson this year and teaches Deep Learning from scratch to people with a programming or computer science/engineering background. Without requiring any prior machine learning skills, LDL starts the first chapter with describing an individual artificial neuron. The remaining chapters build upon this single neuron, while alternating theory and programming. LDL includes programming examples of networks for natural language translation and image captioning, and detailed descriptions of important image and language processing networks such as VGGNet, Inception, ResNet, Mask R-CNN, BERT, and GPT. The talk begins with a high-level description of the topics covered by LDL. It then moves on to a more detailed description of some of the content in the initial chapters, including a couple of programming examples. It concludes with an overview of a more complex application, an image captioning network, that generates textual descriptions of images. This image captioning network is the focus of one of the final programming examples in LDL, and gives a flavor of the type of skills the reader will master after reading the book.

With the arrival of NVIDIA RTX and realtime Ray Tracing APIs like DXR and Vulkan Ray Tracing, it's now easier than ever to create stunning visuals at interactive frame rates.

In this workshop, you'll learn how to utilize NVIDIA® Nsight™ Graphics to profile and optimize 3D Applications that are using Ray Tracing. Using an example application, you'll learn how to:

> Understand how modern GPUs function and how to properly feed the graphics pipeline
> Identify GPU bottlenecks that degrade performance by inspecting low level metrics
> Analyze profiler data and understand what actions you should ta
> Optimize some example workloads to achieve peak performance

The content covers how to deploy PyTorch, ONNX, TensorFlow, and TensorRT models to Triton. This includes how to use functionality like perf_analyzer, model-analyzer, Prometheus and Grafana, Python and C++ clients, ensembling models, and custom Python and C++ backends. Additionally, this content is motivated by real-world challenges that data scientists face as they explore Triton (NLP and RecSys use cases).

Upon completion of this lab, students will be able to:

> Create an end-to-end inferencing pipeline
> Create models across all frameworks (PyTorch, ONNX, TensorFlow, TensorRT) for deployment in Triton
> Define model configurations for these models, with different settings
> Send requests to Triton for each of these different models using Python synchronous and asynchronous calls as well as dynamic batching
> Analyze and understand opportunities for further optimization
> Use model-analyzer and perf_analyzer tools to assess model performance
> Use Prometheus and Grafana to build monitoring dashboards to visualize performance metrics (latency, throughput, dynamic batch size, etc.)
> Leverage gRPC, asynchronous inferencing, dynamic batching, shared memory, etc. to optimize latency and throughput
> Create Python and C++ backends for custom operations
> Create ensembling modules (Python and C++)
> Create Python and C++ clients to interact with the Triton server

This lab teaches how to use NVIDIA's Nsight tools for analyzing and optimizing CUDA applications. Attendees will be using Nsight Systems to analyze the overall application structure and explore parallelization opportunities. Nsight Compute will be used to analyze and optimize CUDA kernels, using an online machine learning code for 5G.

Many resources, such as university courses, (online) tutorials or academic literature, focus mainly on different models and model types and rarely discuss the steps for preprocessing or feature engineering. Feature engineering is an important component in (tabular) machine learning problems, which can be easily integrated into an existing model. In this tutorial, we teach best practices for feature engineering techniques specific to tabular data building off our teams’ collective experience competing in data science competitions, such as Kaggle and RecSys. The tabular data structure limits the models capabilities to learn the relationships between features and adding hand crafted features can significantly boost their performance. For example, our team won the RecSys2020 challenge by developing hand crafted features, which outperformed complex models without them. We were able to reduce the calculation time from multiple days to less than an hour by adding GPU acceleration. The speed up was crucial for our high score as we could run more experiments.

In this hands-on tutorial, we will teach you how to create features for tabular data and accelerate your code with GPUs:

> Feature Engineering: Best practices on how to create features from categorical, numerical and time series data
> GPU acceleration: Accelerate your data frame operations with RAPIDS.AI cuDF using its pandas-like API
> Examples: Get a lot of example functions for your machine learning problem

Session-based recommendation, a sub-area of sequential recommendation, has been an important task in online services like e-commerce and news portals. Session-based recommenders provide relevant and personalized recommendations even when prior user history is not available or their tastes change over time. They recently gained popularity due to their ability to capture short-term or contextual user preferences towards items.

In this hands-on workshop, participants will learn:

> the main concepts and algorithms for session-based recommendation
> implementation of preprocessing and feature engineering techniques for session-based recommendation model on GPU
> how to build, train and evaluate a session-based recommendation model based on RNN and Transformer architectures with NVIDIA Merlin Transformers4Rec library
> how to deploy a trained model to the NVIDIA Triton Inference Server