Achronix Semiconductor Corporation announced its latest innovation in partnership with Myrtle.ai — an accelerated automatic speech recognition (ASR) solution based on the Speedster7t FPGA. This transformative solution converts spoken language to text in over 1,000 concurrent real-time streams with high accuracy and fast response times while delivering up to a 20 times improvement in performance over competing solutions. Achronix will demonstrate this solution at the upcoming SC23 conference in Denver on November 12-17, 2023, at booth 2019.
The solution is powered by a VectorPath accelerator card featuring a Speedster7t FPGA running Myrtle.ai’s Achronix-FPGA-optimized ASR IP — delivering industry-leading, real-time, ultra-low latency speech-to-text capabilities. A single card in a server can replace up to 20 CPU-only-based servers or 15 GPU cards. The AI model can be easily customized to trade off accuracy versus performance when support for 1,000 concurrent streams is unnecessary. It is poised to disrupt the ASR landscape with its exceptional word-error rate and 99th percentile latency of 54 ms end-to-end, including pre- and post-processing plus data movement back and forth to the CPU. In addition, the solution can be customized or retrained with vertical-specific or custom data sets in standard machine learning (ML) frameworks.
“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 Myrtle.ai state-of-the-art ASR accelerator IP running on a Speedster7t FPGA and software stack abstracts away the fact that the FPGA powers the solution, making for easy adoption. This solution is poised to redefine how industries process speech data with its demonstrated superiority over competitive GPU-based solutions, delivering a 16 times better performance-to-cost ratio.
“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 revolutionary impact on industries that depend on rapid and accurate speech-to-text conversion. Its features include compatibility with major deep learning frameworks such as PyTorch, plus re-trainability for multiple languages or specialties. The solution is currently being deployed with our early-access customers and is now available to the general market.