Democratizing AI Access with Intel® Tiber™ Developer Cloud

This blog was updated on January 9, 2024.

Building, training, or inferring with intensive artificial intelligence models, such as large language models (LLMs), requires access to expensive resources, time, and expertise. Procuring hardware capable of running their workloads, or even knowing where to start, can be a barrier for developers innovating the latest advancements. Utilizing cloud platforms allows developers to access their projects from anywhere and get to market quickly by minimizing setup times and optimizing workflows. Developers can also have peace of mind in ensuring their data is secure by mitigating the risks of data loss due to hardware failures. Additionally, developers can distribute content efficiently from the cloud to the edge, as well as easily utilize containers and relevant ready-to-use software to optimize their projects further, building better solutions and innovations. Cloud platforms with competitive pricing help democratize AI access for developers looking to try new hardware and software toolkits to build their latest projects efficiently.

What is the Intel Developer Cloud?

The Intel® Developer Tiber™ Cloud is a cloud-based platform that empowers developers, data scientists, and AI enthusiasts by providing access to cutting-edge Intel hardware and software solutions so that AI and High-Performance Computing applications can be built, tested, run, and optimized at low cost and overhead. The Intel Developer Tiber Cloud provides developers an easy path to innovate with small or large workloads on Intel CPUs, GPUs, and AI accelerators with access to AI-optimized software such as oneAPI. General availability of Intel Developer Cloud was recently announced at Intel Innovation 2023.

In this post, we’ll explore how to get started on Intel Developer Tiber Cloud and unlock its potential in kickstarting your development projects.

Benefits of Intel® Tiber™ Developer Cloud

The Intel Developer Cloud offers an array of benefits to developers looking to build innovative AI and HPC applications. It provides access to the latest Intel hardware: Intel® Xeon® CPUs, Intel® Data Center GPUs, Intel® Gaudi® 2 AI Accelerators, and more. The platform includes flexible software and services to run large-scale AI training, inferencing, and high-performance computing. Based on their needs, Developers can also choose their machine type, between virtual machines or bare metal on a full system or a cluster. The latest AI tools, framework optimizations, and reference implementations are also included to help enhance workflows and speed up processes even further.

Getting Started

Visit cloud.intel.com to get started. Watch this video or follow the steps outlined below for detailed steps to begin with Intel Developer Tiber Cloud.

Making an account:

There are several different types of accounts, from standard free accounts to paid premium and enterprise accounts. Choose the account that best fits your needs, and you can always upgrade your account to fit your needs at any time.

Subscribe to your preferred service tier and follow the registration process by providing and verifying an email and some basic information.

Once your account is created and verified, you can access the platform console with quick access to tools and resources.

Launching a Free Jupyter Notebook for Training

On the left menu within the Intel Developer Tiber Cloud platform, you can find the Training page under the SOFTWARE header as shown below.

Within this Training and Workshops page, you can find several training workshops for building AI models, learning Gen AI essentials, and optimizing with SYCL C++. There is also an option on the top right of the page to launch your own JupyterLab where you can work with different notebooks and consoles to create your own machine learning models and will run on unspecified hardware on the Intel Developer Cloud – perfect for trying it out and learning how these tools can help you develop efficiently.

To easily get started on your own, launch a JupyterLab session on the top right of the Training and Workshops page with the button shown below. You can choose a Python 3 kernel to write your own Python code.

Check out the demo at the end of this article for a model you can build in JupyterLab on the Intel Developer Cloud!

Launching a Compute Instance

Steps to launch an instance:

From the Console Home of Intel Tiber Developer Cloud, you can access various hardware configurations through the Hardware Catalog. For this example, we’ll be working with the Intel® Max Series processors.

Select this option from the catalog, and note the matching instance family, the instance type (which defaults to a 4x GPU machine), and the machine image (which defaults to Ubuntu* OS). Name your instance and get ready to set up your SSH keys for access.

Setting up a key:

Intel Developer Tiber Cloud provides detailed instructions with command lines for setting up SSH keys and accessing your instance remotely. You will generate an SSH key on your local machine and provide Intel Developer Cloud with the public key.

Creating an SSH key depends on your operating system. Click Create a Key and follow the steps below per your OS.

Linux* OS and MacOS*:

Launch a Terminal on your system and copy and paste the following into the terminal:

ssh-keygen -t rsa -b 4096 -f ~/.ssh/id_rsa

If you are prompted to overwrite, select No. Create and repeat a passphrase if prompted.

Copy and paste the following to your terminal to open your SSH key:

vi ~/.ssh/id_rsa.pub

Copy the generated key from the terminal and paste it to the key contents field on the Intel Developer Cloud instance setup.

Microsoft Windows* OS:

Launch a PowerShell* terminal window on your system. If you haven’t generated a key before, copy and paste the following in the terminal:

mkdir $env:UserProfile\.ssh

Copy and paste the following to your terminal to generate SSH keys:

ssh-keygen -t rsa -b 4096 -f $env:UserProfile\.ssh\id_rsa

If you are prompted to overwrite, select No. Create and repeat a passphrase if prompted.

Copy and paste the following to your terminal to open your SSH key:

cat $env:UserProfile\.ssh\id_rsa.pub

Copy the generated key from the terminal and paste it to the key contents field on the Intel Developer Cloud instance setup.

To view more information about how to generate a key, follow Generate an SSH Key. The uploaded key will appear on the keys screen, allowing you to continue setting up your instance.

Running the machine:

Press launch and your instance will begin provisioning and will soon be ready! Clicking on the instance name will list the machine's specifications and a “How to Connect” button. Clicking on this button reveals instructions for connecting to this machine from your local device running Windows, macOS, or Linux. Follow these instructions in the command-line interface on your local device, using SSH to connect to the Intel Developer Cloud instance, and you’re in!

You can view and confirm the devices in your instance by running the following in the window:

sycl-ls

Setting up oneAPI:

After connecting to your Intel Developer Tiber Cloud instance, a quick first step is to initialize oneAPI on the machine. In the command line:

source /opt/intel/oneapi/setvars.sh --force

oneAPI provides your environment with libraries, compilers, and more for creating your applications.

Chatbot Sample Demo

Developing and running your own AI applications on the platform is as easy as it gets. Building an AI chatbot in just a few lines of code with Neural Chat, which utilizes Hugging Face Transformers and Intel AI software, is made possible with Intel Developer Tiber Cloud. Neural Chat provides single-line API calls that implement optimizations using Intel® Neural Compressor.

Getting started with building a Neural Chat chatbot requires just three lines of code:

from intel_extension_for_transformers.neural_chat import build_chatbot 
chatbot = build_chatbot() 
response = chatbot.predict("Tell me about Intel Xeon Scalable Processors.")

Here is an example showcasing the Neural Chat chatbot running on a 4th gen Intel Xeon processor instance on the Intel Developer Cloud using the code above, where the chatbot model responds to the textual prompt instruction given:

You can learn more about how to run and customize your own chatbot with Neural Chat by watching this video.

Check out Intel Tiber Developer Cloud to access the latest silicon hardware and optimized software to help develop and power your next innovative AI projects! We encourage you to check out Intel’s AI Tools and Framework optimizations and learn more about the unified, open, standards-based oneAPI programming model that forms the foundation of Intel’s AI Software Portfolio.