<h1>Canonical Releases Silicon-Optimized AI Model Snaps for Ubuntu</h1>
<p>Canonical has introduced beta snaps for the DeepSeek R1 and Qwen 2.5 VL AI models, now available on the Snap Store. These snaps are specifically optimized for Intel and ARM Ampere hardware, marking a significant milestone as the first large language models (LLMs) available as snaps. The primary aim of this release is to simplify the deployment of AI workloads on Ubuntu systems, providing hardware-specific optimizations without the need for cloud dependencies or subscription costs.</p>
<h2>Key Features</h2>
<ul>
<li>
<p><strong>Automatic Optimization</strong>: These snaps feature the ability to automatically select the most efficient engine, quantization, and architecture based on the device’s silicon. This innovative approach eliminates the need for manual research and deployment of quantized models, facilitating a seamless experience for developers and enterprises.</p>
</li>
<li>
<p><strong>Local Running Capability</strong>: Both DeepSeek R1 and Qwen 2.5 VL models are designed to run locally on users’ machines. This ensures privacy, reduces latency, and removes any additional costs associated with cloud services.</p>
</li>
<li>
<p><strong>Universal Compatibility</strong>: The beta release includes optimizations for both Intel and Ampere-based systems, making it a versatile solution applicable across various environments including desktops, servers, and edge devices.</p>
</li>
</ul>
<h2>Deployment Made Easy</h2>
<p>Canonical’s new system simplifies the deployment process through a single command that installs the best model configuration tailored for a user’s device. This development significantly lowers the barrier to entry for leveraging AI technologies, particularly for developers and enterprises eager for efficient, slam dunk AI solutions.</p>
<h2>Open Source Initiative</h2>
<p>As part of Canonical’s commitment to the open-source community, the framework has been made open to silicon partners. This allows partners to contribute their own optimizations, enabling a broader range of hardware-specific enhancements across different platforms.</p>
<p>This strategic release by Canonical not only enhances the capability of Ubuntu systems but also sets a new standard in the AI deployment space, making advanced AI technologies more accessible and efficient for a wider audience.</p>


Collections

Canonical launched beta snaps for DeepSeek R1 and Qwen 2.5 VL AI models on the Snap Store, optimized for Intel and ARM Ampere hardware. These are the first LLMs available as snaps, featuring automatic hardware-specific optimization, local execution for privacy and reduced latency, and one-command deployment. The open-source framework allows silicon partners to contribute their own optimizations, simplifying AI workload deployment on Ubuntu systems without cloud dependencies or subscription costs.