|
Title: Apple collaborates with NVIDIA to research faster LLM performance Post by: HCK on December 26, 2024, 04:05:05 pm Apple collaborates with NVIDIA to research faster LLM performance
<div class="feat-image">(https://9to5mac.com/wp-content/uploads/sites/6/2024/12/tensor-rt-llm-graphic.png?w=1600)</div><p>In a blog post today (https://machinelearning.apple.com/research/redrafter-nvidia-tensorrt-llm), Apple engineers have shared new details on a collaboration with NVIDIA to implement faster text generation performance with large language models. </p> <p>Apple published (https://machinelearning.apple.com/research/recurrent-drafter) and open sourced (https://github.com/apple/ml-recurrent-drafter) its Recurrent Drafter (ReDrafter) technique earlier this year. It represents a new method for generating text with LLMs that is significantly faster and “achieves state of the art performance.” It combines two techniques: beam search (to explore multiple possibilities) and dynamic tree attention (to efficiently handle choices).</p> <a data-layer-pagetype="post" data-layer-postcategory="aapl,nvidia" data-layer-viewtype="unknown" data-post-id="983169" href="https://9to5mac.com/2024/12/18/apple-collaborates-with-nvidia-to-research-faster-llm-performance/#more-983169" class="more-link">more…[/url] Source: Apple collaborates with NVIDIA to research faster LLM performance (https://9to5mac.com/2024/12/18/apple-collaborates-with-nvidia-to-research-faster-llm-performance/) |