TensorStack & AMD launch Amuse AI image tool in Beta

TensorStack and AMD have announced the launch of Amuse, a new AI image generation tool available in Beta. This collaboration between the New Zealand-based AI startup and the global semiconductor company enables Windows users to generate images locally on their computers.

Amuse allows users to generate AI images without the need for an internet connection, offering an alternative to cloud-based solutions. This local approach is designed to enhance privacy by reducing the risk of sensitive data being shared. Additionally, the absence of subscription fees means users maintain full ownership and control over their generated content.

AMD is utilising Amuse to showcase the capabilities of its latest hardware, specifically the AMD Ryzen AI 300 series processors. At the South by Southwest (SXSW) conference in March, AMD CEO Dr. Lisa Su demonstrated the use of stable diffusion models in Amuse to create images from text prompts, running on an AMD AI PC.

The release of Amuse aims to offer a seamless and private image generation experience, leveraging the high-performance capabilities of AMD's new laptops. "We are excited to collaborate with AMD and bring Amuse to the world," said Rita Langley, Co-Founder and CEO of TensorStack. "Our goal has always been to democratise access to the latest AI technology, and this partnership allows us to further that mission."

The Amuse tool features an easy installation process with no dependencies, and a user-friendly interface called EZ Mode that simplifies configuration. Among its key features are the AMD XDNA Super Resolution solution, which speeds up output and increases image size by twofold. Other functionalities include paint-to-image and drawing-to-image conversions, creating custom AI filters, and employing a multi-stage ONNX-based pipeline that incorporates stable diffusion, control nets, and feature extractors.

Adam Clark, CTO of TensorStack, played a pivotal role in the development of Amuse through the creation of OnnxStack, a .NET AI library for the Open Neural Network Exchange (ONNX). This library enhances the integration of AI capabilities into applications. Clark emphasised the importance of AMD's hardware in their development efforts, stating, "AMD's hardware has exponentially accelerated our AI development. This has been pivotal in the success of Amuse and OnnxStack."

While Amuse offers a range of advanced features, it is important to note that the current release is in Beta and considered experimental. Users may encounter bugs and instability during its use.

TensorStack is committed to making advanced AI technology accessible and safe for everyone, with a strong focus on open-source solutions. The partnership with AMD underscores this commitment, bringing cutting-edge AI tools directly into the hands of users, with an emphasis on privacy and control.