Session: How Transformers are Changing the Nature of Deep Learning Models

Tom Michiels, System Architect, ARC Processors, Synopsys

The neural network models used in embedded real-time applications are evolving quickly. Transformer networks are a deep learning approach that has become dominant for natural language processing and other time-dependent, series data applications. Now, transformer-based deep learning network architectures are also being applied to vision applications with state-of-the-art results compared to CNN-based solutions. In this presentation, we will introduce transformers and contrast them with the CNNs commonly used for vision tasks today. We will examine the key features of transformer model architectures and show performance comparisons between transformers and CNNs. We will conclude with insights on why we think transformers will become increasingly important for visual perception tasks.

Booth #309 Demos:

Tuesday, May 23 | 12:30 PM – 7:30 PM PDT
Wednesday, May 24 | 10:30 AM – 6:00 PM PDT

Time to meet with Synopsys executives and technical experts.

Visit the Synopsys booth to see demonstrations of emerging neural network architectures like transformers, neural network based ISP, and vision/object detection for safety-critical automotive SoCs. 

Visionary.ai True Night Vision SW ISP

Come see this live hardware demonstration of Visionary.ai's AI software image signal processor (ISP) as it delivers drastically better results in low light, and superior color accuracy in high light compared to a camera using a standard ISP.