Date: 16/10/2023

Achronix unveil improved speech recognition based on FPGA

FPGA expert Achronix Semiconductor said in its latest collaboration with Myrtle.ai, compant has introduced speedster7t FPGA-Accelerated automatic speech recognition solution.

This solution enables the conversion of speech into text across more than 1000 concurrent streams, with high accuracy and rapid response times. With an increase of over 20 times in performance when compared to rival companies .This solution will be demonstrated by Achronix at the upcoming SC23 conference in Denver on November 12-17, 2023 at booth 2019.

This solution is made using VectorPath accelerated card activated by Speedster7t FPGA running Myrtle.ai’s Achronix-FPGA-optimized ASR IP, provides real-time, low-latency speech-to-text capabilities. Where a single Vectorpath card can replace up to 20 CPU-only based servers of 15 GPU cards. With good word-error rate and 99th percentile latency of 54 ms end-to-end, it is set to disrupt the latency. The AI model can be customized for accuracy versus performance when support for 1,000 concurrent streams is not required. The solution can also be customized or retrained with vertical-specific or custom data sets in standard ML frameworks.
The Myrtle.ai ASR accelerator IP simplifies the adoption process by hiding the fact that it is powered by an FPGA. This solution is set to revolutionize the way industries handle speech data, as it has proven to be much better than GPU-based solutions. It offers a performance-to-cost ratio that is 16 times higher.

"One of the key aspects of the Myrtle.ai accelerated ASR solution built on Achronix Speedster7t FPGAs is its ability to reduce both OpEx and CapEx while maintaining top-tier performance significantly," noted Bill Jenkins, the Director of AI Product Marketing at Achronix. "This solution, powered by a Speedster7t FPGA, can reduce costs by up to 90% compared to traditional CPU/GPU-based server solutions, whether enterprise or in the cloud. This capability translates to tangible business savings while providing exceptional real-time speech-to-text capabilities."

“The architecture of the Achronix Speedster7t FPGA with its 2D network on chip (NoC) and ML processor (MLP) arrays gave us the building blocks required to create an ASR product that is significantly more optimized than anything available on the market today,” said Peter Baldwin, CEO of Myrtle.ai, a company known for its expertise in optimizing low-latency ML inference for real-time applications. “The extremely low latency inherent in these FPGAs makes them ideal for real-time workloads. We are excited to provide users the ability to scale up their ASR services at lower cost and faster than ever before.”

The Achronix-Myrtle.ai accelerated ASR solution will have a big impact on fast and accurate speech-to-text conversion related industries. Some of its features includes support to deep learning frameworks such as Pytorch with support for re-trainability in multiple languages or specialities.

The solution is presently being implemented with Achronix’ early-access clientele and is now accessible to the wider market.

News Source: Achronix

Tweet Follow @ecewire