Fujitsu has developed a generative AI reconstruction technology that enhances the Takane LLM, achieving an impressive 89% accuracy retention rate and a 94% reduction in memory consumption through advanced 1-bit quantization and specialized AI distillation. This breakthrough enables the deployment of lightweight and energy-efficient AI models on edge devices, paving the way for improved real-time responsiveness and reduced power consumption in various applications.
This is an ainewsarticles.com news flash; the original news article can be found here: Read the Full Article…
