Bing Search gets faster, more accurate and efficient through SLM models and TensorRT-LLM
The Bing Search crew shared the best way it helped make Bing Search and Bing’s Deep Search sooner, further appropriate and cheaper by transitioning to SLM fashions and the blending of TensorRT-LLM.
Bing wrote“to reinforce effectivity, we expert SLM fashions (~100x throughput enchancment over LLM), which course of and understand search queries further precisely.”
The benefits. Bing talked about it has helped make search larger complete by bringing these three core benefits to its searchers:
- Sooner Search Outcomes: With optimized inference, Bing clients can benefit from sooner response events, making the Bing search experience further seamless and setting pleasant.
- Improved Accuracy: The improved capabilities of SLM fashions allow Microsoft to ship further appropriate and contextualized search outcomes, serving to bing searchers uncover the info they need further efficiently.
- Worth Effectivity: By reducing the value of web internet hosting and working big fashions, Microsoft talked about it may truly proceed to spend cash on further enhancements and enhancements, guaranteeing that Bing stays on the forefront of search experience.
Why we care. A sooner searcher experience and a further appropriate search experience might also assist Bing flip into further trusted and useful to searchers. This can possible lead to further searchers adopting Bing Search eventually, taking search market share away from bigger avid gamers like Google.
New on Search Engine Land