Numenta partnered with Intel to deliver fast, cost-effective inference for NLP tasks like text classification and sentiment analysis. Traditionally run on GPUs, transformer models such as BERT and GPT can now achieve high performance on CPUs using Numenta’s neuroscience-inspired acceleration technology combined with Intel® Advanced Matrix Extensions in 4th Gen Intel® Xeon® Scalable processors. This solution reduces hardware costs and complexity while meeting the latency and throughput demands of real-time AI applications, making CPUs a powerful alternative for deploying performance-sensitive transformer workloads.

Join for free to read