Good news!
"... research team from Cornell University performed small-scale optical experiments with a prototype accelerator to demonstrate that Transformer operations can run on optical hardware despite noise and errors. They presented the research results in a paper investigating the energy-efficiency advantage that could be achieved in executing state-of-the-art Transformer models on optical hardware. They conclude that for large enough models, an energy-efficiency advantage of >8,000x versus current electronic hardware (GPUs) should be possible."
Credits: Marktechpost AI Newsletter
No comments:
Post a Comment