Qualcomm has made a significant breakthrough in the world of generative AI models by developing a way to run Stable Diffusion on Android smartphones. Previously, generative AI models like Stable Diffusion had to run in the cloud, but Qualcomm has shrunk the model using a combination of software techniques and hardware optimization to run inference models on common Android smartphones. Stable Diffusion, created by startup Stability AI, is one of the most popular generative AI models for image creation, frequently competing with OpenAI's DALL-E.
Users can now generate Stable Diffusion generative AI images directly on an Android smartphone, eliminating the need to call out to the cloud to do the heavy lifting. Because users' information or ideas remain solely on the device, this has significant implications for privacy and security. This advancement is even more important for enterprise use of generative AI, as company confidential information can now be protected.
Qualcomm's achievement is a critical step toward efficient edge AI processing. It protects user privacy while running Stable Diffusion and other generative AI models because the input text and generated image are never sent outside of the device. This is significant for both consumer and enterprise application adoption. Lessons learned from this project can be applied to bring other large generative AI models from the cloud to the device, benefiting future iterations and Qualcomm AI Stack users.
Source: VentureBeat (link)
Images: MidJourney, Prompts by Lewis Farrell
Interesting story? Please click on the ? button below!