Natural language processing (NLP) and speech-based models have become core to the majority of AI development in the enterprise today. NLP-based workloads are powering conversational AI and recommendation use cases in the consumer internet space. In this session, we’ll explore key trends …

Compute resources required to continue advancing state-of-the-art results are increasing due to larger neural network models and datasets. To keep up, NVIDIA GPUs, available on AWS, have been adding features designed to accelerate neural network operations. Tensor Cores are one of these features, …

Utilizing GPU accelerators in Apache Spark 3.x presents opportunities for significant speedup of ETL, ML and DL applications. In this session we’ll dive into the Apache Spark 3.x RAPIDS plugin, which enables applications to take advantage of GPU acceleration with no code change and then show how to …

Learning from graph and relational data plays a major role in many applications. In the last few years, Graph Neural Networks (GNNs) have emerged as a promising new machine learning framework, capable of bringing the power of deep representation learning to graph and relational data. This ever-…

Learn about the features, improvements, and performance capabilities of MXNet on NVIDIA hardware. We’ll explore the ecosystem and the vision that makes the framework scale efficiently on large GPU clusters, both in terms of time and cost savings for cloud-based inference. We’ll also demonstrate …

AI is a driving force in Fintech innovation and enables startups to compete with the large banks across the entire spectrum of financial services. From autonomous finance to automated insurance claims and virtual assistants, Fintechs are creating new experiences for consumers while creating competitive …

In the midst of a global pandemic, startups are requiring more support than ever before. Join this session to learn how NVIDIA Inception is nurturing cutting edge AI and data science startups through access to technical expertise, preferred pricing for on-prem NVIDIA GPUs, technical training through …

Learn how developers can work with multiple components of the NVIDIA stack available on AWS to effectively accelerate training models, inference applications, and HPC workloads using AWS infrastructure. In this 201-level session you’ll be able to identify multiple NVIDIA stack modules …

AI is a driving force in Fintech innovation and enables startups to compete with large banks across the entire spectrum of financial services. From autonomous finance to automated insurance claims and virtual assistants, Fintechs are creating new experiences for consumers, while creating a competitive …

Triton Inference Server is a model serving software that simplifies the deployment of AI models at-scale in production. It allows teams to deploy trained AI models from any framework (NVDIA TensorFlow, NVIDIA TensorRT, PyTorch, ONNX Runtime, or a custom framework) on any GPU- or CPU-based …

Game designers have traditionally been left out of the remote work equation, based solely on the location of the equipment they need to do their job. 2020 has forced a rapid evolution. By moving their production equipment into the cloud, game studios are increasing their agility, accelerating …

In this session, we’ll discuss MatchboxNet, an end-to-end neural network for speech command recognition. MatchboxNet is composed from blocks of 1D time-channel separable convolution, batch-normalization, ReLU, and dropout layers. It reaches state-of-the-art accuracy on the Google Speech …

Deep Neural Network (DNN) model complexity has increased in the last decade, from Alexnet with 61 million parameters to GPT-3 with 175 billion parameters. Running real-time inference on such large models is difficult without a graph “compression” technology that optimizes a pre-trained neural …

Conversational AI is comprised of three exciting areas of AI research: automatic speech recognition (ASR), natural language processing (NLP), and speech synthesis (or text-to-speech, TTS). NVIDIA NeMo, an open source toolkit, democratizes and accelerates progress in these areas. In this session, we’ll …

Data Scientists spend a significant amount of time on ETL and ML model training. For example, processing a dataset on the scale of 1TB in a popular TPC-xBB benchmark takes a few hours. In this talk we will demonstrate a significant speedup in data processing and training using RAPIDS : a set of …

CloudXR, NVIDIA’s solution for streaming VR, AR, and MR content allows you to experience high-fidelity models in immersive environments. Recently, NVIDIA announced that CloudXR will become available on the Amazon Web Services Marketplace in early 2021. In this presentation, you will see how CloudXR …

Attention based models like BERT have revolutionized natural language processing (NLP) due to its ability to outperform traditional models on language tasks as shown by their high scores on various NLP benchmarks However, even smaller BERT models have more than 100 million parameters, …

In this session, we’ll demonstrate how to speedup DL training with Automatic Mixed Precision and leverage Tensor Cores on the new NVIDIA GPUs on AWS. You’ll learn how to make small modifications in your training code to achieve an ~8X speedup in DL training and how to run inference on this trained…

Deep learning inference is a compute-intensive workload that affects user-experience. Real-time applications need low latency and data center efficiency requires high throughput. In this session, we’ll demonstrate how developers can use NVIDIA TensorRT to optimize neural network models, trained in …

Transformer networks have revolutionized the field of machine translation. They’ve been shown to produce better translations, especially for long input sentences, than the traditional recurrent neural networks. However, Transformer models have been growing with the latest GPT-3 having 175 billion …

Automatic Speech Recognition (ASR) is the first step to conversational AI and Kaldi ASR has become a de-facto speech recognition toolkit. In this session, learn how to deploy GPU-optimized Kaldi ASR at-scale with NVIDIA Triton Inference Server from the NGC Catalog in AWS marketplace, on Amazon …

In this session, learn about NVIDIA cutting-edge technologies that help accelerate end-to-end recommender system applications. From data loading, experimentation, model training, and real-time inference to production deployment, we’ll discuss a real use case and use it to demonstrate how you …

Find out how to train and deploy your own chatbot capable of both closed and open domain conversation using NVIDIA Jarvis. Discover how to support multiple closed domains, which can answer queries ranging from weather forecasts to movie recommendations (the recommender system build …

Join us to learn about Multi Instance GPU (MIG), one of the new features of the NVIDIA A100 GPU model, which plays a special role in DL-based applications. MIG makes it possible to use a single GPU as if it were multiple smaller GPUs, maximizing utilization of the GPU for inference, providing dynamic …