Suresh Pasupuleti, Managing Director of Vedya Labs, presents the company’s work in bringing Stable Diffusion-based image generation to DSP-centric embedded platforms. The demonstration showcases a nearly 500-million-parameter model running on the Axera AX650N SoC, with the text encoder, U-Net, and VAE stages optimized for dual Cadence Tensilica Vision DSPs. Using INT8 quantization and a combination of proprietary and Cadence optimization tools, Vedya Labs achieves high DSP utilization while maintaining practical image generation performance. The demo highlights real-time prompt-to-image generation and illustrates how advanced generative AI models can be efficiently deployed on embedded hardware without relying solely on GPU or NPU accelerators.
