Run Hugging Face Models on AMD GPUs with One Click

Testing popular large language models often means waiting for GPU access and managing environment setup. The Hugging Face One-Click Notebook experience, powered by AMD Developer Cloud, eliminates this friction. Pick a model on Hugging face and try it directly on a notebook running on AMD GPUs.

With access to AMD GPUs featuring up to 192 GB of GPU high bandwidth memory , developers can experiment with larger models and demanding AI workloads. The integrated notebook experience makes it easy to evaluate, prototype, and iterate without worrying about infrastructure setup or memory constraints.

Check this collection to see which models are supported!

How it works: Step by Step

Step 1: Start with Hugging Face Models

To try this workflow, start from the AMD Model Collection Page. In this example, search for and open the Qwen3.6-35B-A3B model page.

Step 2: Launch with AMD Developer Cloud

On the right side of the model page, open the Use this model drop-down menu and select AMD Developer Cloud. Your browser will redirect you to the AMD AI Notebooks sign-in page.

Step 3: Sign into AMD Developer Program

Sign in using your AMD AI Developer Program credentials, just as you would with any other cloud-based Jypyter notebook service.  If you aren’t already a member,  create a free account here, to recieve 100$ in free GPU credits and access to training content through the AMD AI Academy.

After signing in with your credentials, you will be directed to the Notebooks page.

Once authenticated, your notebook environment is automatically provisioned. The First time Authentication might involve a code verification via your email.

Step 4: Notebook and GPU Ready

After signing in, you'll be presented with a familiar cloud-based notebook environment which will load a powerful AMD GPU enabled in the backend. Step through the Python notebook to run the model example directly in the hosted environment.

Get Started Today

The Hugging Face One-Click Notebook experience makes it easy to evaluate the latest AI models directly from a Hugging Face model page using AMD GPU-powered notebook instances. With a growing catalog of supported models and new additions being enabled regularly, developers can quickly access cloud-based AMD GPU environments for AI development and evaluation.

Whether you're exploring the latest LLMs, benchmarking model performance, or prototyping AI applications, you can get from model discovery to experimentation in just a few clicks.

Ready to get started? Visit Hugging Face, launch a supported model with AMD Developer Cloud, and experience how quickly you can start building and evaluating AI workloads on AMD GPUs.