Ceva’s BX DSP cores and Whispro speech recognition software for conversational AI now also support Tensorflow Lite for microcontrollers, a framework for deploying tiny machine learning on  processors in edge devices.

Ceva’s BX DSP cores and Whispro speech recognition software for conversational AI now also support Tensorflow Lite for microcontrollers, a framework for deploying tiny machine learning on  processors in edge devices.

Ceva says its approach to AI at the edge means customers using TensorFlow Lite for can use one processor architecture to run both the framework and the associated neural network workloads.

The firm’s Whispro speech recognition software and custom command models are integrated with the framework, further accelerating the development of small footprint voice assistants and other voice controlled IoT devices.

The Ceva-BX is a programmable hybrid DSP/controller designed for signal processing and controlling workloads in real-time applications.

Using an 11-stage pipeline and 5-way VLIW micro-architecture, it offers parallel processing with dual scalar compute engines, load/store and program control that reaches a Coremark per MHz score of 5.5, making is suitable for real time signal control.

Its support for SIMD instructions makes it suitable for signal processing applications and the double precision floating point units are capable of contextual awareness and sensor fusion algorithms.

It facilitates simultaneous processing of front-end voice, sensor fusion, audio processing, and general DSP workloads in addition to AI runtime inferencing, the firm claims. 

Erez Bar-Niv, CTO, Ceva, stated: “With full optimization of this framework for our Ceva-BX DSPs and our Whispro speech recognition models, we are lowering the entry barrier for SoC companies and OEMs to add intelligent sensing to their devices.”