Rapid AI image creation is now possible on an Android phone. In a recent demo video, Qualcomm was able to create a 512 x 512-pixel image of a cat wearing armor using AI-image-generator Stable Diffusion version 1.5 on an Android phone. And the kicker is that it’s done in less than 14.42 seconds, which makes it close to what PCs can do with the technology.
To achieve this, the company’s engineers used a process called “quantization“. Without getting too deep into the weeds here, they’ve taken the desktop version of Stable Diffusion, shortened it, and optimized it for mobile devices. According to Qualcomm, the process “increases performance [while] ALSO [saving] power by allowing the model” to run more efficiently thus”[consuming] less memory bandwidth.”
It is not known exactly what mobile device was used in the demonstration. However, we do know that it is powered by the Snapdragon 8 Gen 2 Mobile Platform, the exact same chipset found in Samsung Galaxy S23. It is worth noting that some developers managed to install Stable Diffusion on Sony Xperia 5 II, but generating images on that phone takes an hour due to weaker hardware. The main takeaway from Qualcomm’s success is its speed and what it could mean for users down the line.
Future of generative AI
The thing about generative AI models is that they require a fair amount of computing power to work. The developers of Stable Diffusion, for example, recommend using a computer with NVIDIA hardware and, at least, 6.9 GB of VRAM specifically (video RAM) to create images. Smartphones don’t come with VRAM because manufacturers opt for regular RAM to run the device’s software. But now this difference in power is starting to disappear, so we can see this technology growing in many mobile devices.
Jilei Hou, vice president of engineering at Qualcomm, told VentureBeat that the lessons learned in optimizing Stable Diffusion for smartphones can be used for other devices running on the company’s hardware. Hou specifically pointed to laptops and XR headsets as potential future homes for generative AI. As for what they can do, there are many applications such as “image editing [plus] in-painting”, a process where AI fills in the missing parts of a created artwork. The company also hopes that all this can be done without an internet connection in the Stable Diffusion cloud. Although it is cool , it is not known if or when this technology will see a widespread launch in phones running on Qualcomm hardware.
Besides creating images, developers are creating writing AI to help write content. Be sure to check out our recent list of the best AI writers for 2023 – Jasper ranks among our favorites.