NGC

NGC Getting Started Collection

What's in this Collection?

Welcome to NGC!! If you're new to AI, deep learning, or NGC, you're in the right place. This Collection should get you up and running in no time. In our Getting Started collection, you'll find:

An introduction to NGC, what content we have, and why we are here
A walkthrough to show you how to get started with our containers
A tutorial to help you get the most from our models and resources

Collection Contents

TensorFlow Container - NVIDIA's accelerated TensorFlow runtime, designed to get the most out of our GPUs
MNIST Example - If you've never built an app with deep learning before, start here. This walkthrough for the popular MNIST dataset will get you to recognize handwritten numbers
BERT Jupyter Notebook - An end to end example that uses NGC's containers, models, and resources to deploy an app for question answering

Introduction

Welcome to NGC - NVIDIA's home of accelerated software.

If you're working in Artificial Intelligence, High Performance Computing or Visualisation, this is the place to get the latest GPU optimized software. Take a look at what we've got on offer:

Collections

NGC Collections provide all the assets you need to build cutting-edge AI software in one place. So whether you're looking for the latest conversational AI models or want to know what models we have that work with our Transfer Learning Toolkit, Collections provides the perfect starting point. We even have Collections that show you how you can get started with GPUs in the major CSPs. Whatever your experience level, this is a real boost to your workflow.

Containers

Containers package software applications, libraries, dependencies, and runtime compilers in a self-contained environment so they can be easily deployed across various compute environments. The deep learning frameworks and HPC containers from NGC are GPU-optimized and tested on NVIDIA GPUs for scale and performance. With a one-click operation, you can easily pull, scale, and run containers in your environment.

In NGC, we have everything from RAPIDS to DIGITS and all the major deep learning frameworks. Optimized by NVIDIA to ensure that you're getting the most from your GPU infrastructure.

We even host validated third-party applications such as H2O's DriverlessAI or MatLab from Mathworks.

Helm Charts

Helm provides a container orchestration tool that allows you to configure and manage your containerized application deployments. Helm charts allow DevOps engineers and system administrators to describe exactly what components an application needs to run, simplifying deployment pipelines and enabling the easy deployment and integration of new apps.

Models

While starting from scratch is awesome, we know that you want to start with a state-of-the-art model a lot of the time. Whether you're using these as a benchmark to measure performance or as a building block for your own applications based on your data. We've got hundreds of pretrained checkpoints that cover everything from speech generation with Waveglow to VehicleMakeNet that will tell you, well, what make a vehicle is in streaming video. These models can provide the perfect building blocks for you.

Resources

It's not just custom models, though. What about the code that you need to get up and running, retrain the asset, or integrate into an end-user application? We host code samples, example applications, videos, images, and instructions so that you can get up and running with NGC's content really quickly. AI is very much a collaborative effort, and our Resources provide structure to solving the challenges that matter to you. You can even view Jupyter Notebooks directly in NGC, there's no need to download anything. Simply select "view Jupyter" from the actions menu in the file browser, and you're good to go.

Working with NGC

Alright, so that's a rundown of the content we have in our Catalog, but what next? Check out some of the steps in this section to see how you can take your NGC usage to the next level.

Creating An Account

Whoa! That's right - while 99% of the content in our Catalog is free to use, there's even more available if you're a registered user. If you want access to everything in the Catalog, you need to create an account.

It's a really quick process. Simply follow these steps:

Visit https://ngc.nvidia.com/signup to get started
Enter your full name and email address
Accept the Terms and Conditions
Check the link in your email to verify your account
Tell us what content you're interested in so we can customize your experience

That's it! You've now got access to everything our Catalog has to offer.

Developer Forums

Need help with anything on NGC? Head over to our online community on the developer forums and ask away. You can get help with everything from our CSP integrations to submitting feature requests. Take a look at some of our favorite forums below:

Docker and NVIDIA Docker - Everything you ever wanted to know about Docker and NVIDIA Docker.
Feature Requests - Got something on your mind? Let us know here.
Hints and Tips - Get hints and tips from NVIDIANs and some of NGC's power users.

CLI Tools

We've got a really powerful set of CLI tools that let you get the most from NGC's content, whether you're working on your local GPU-powered machine or remote infrastructure. Visit https://ngc.nvidia.com/setup to get started!

Documentation

We've got detailed developer and end-user documentation here.

NGC Private Registry

IIf you think this is cool, wait until you see what you can do with an NGC Private Registry. Imagine all of the power of NGC, and the incredible features you see in our Catalog, but in your hands!

You can manage your own Collections, Containers, Helm Charts, Models, and Resources to improve how your team builds software. We've got a powerful CLI and REST API, so you can even integrate NGC into your workflow. Making the process of building and scaling AI software seamless.

To find out more about NGC private registries, take a look at this blog or book a tour here.

Fashion-MNIST Example with the NVIDIA TensorFlow Container

You can view this notebook in NGC here.

If you want to read through the notebook example without leaving our website, follow these steps:

Navigate to the File Browser tab of the asset in NGC
Select the version you'd like to see
Next to the .ipynb file select "View Jupyter"
There you have it! You can read a notebook for documentation and copy code samples without ever leaving NGC.

This Jupyter Notebook contains all the instructions you need to get started with NGC containers and the code for a beginner's image classification task using the NVIDIA TensorFlow container.

BERT Jupyter Notebook