TOKYO -- Tokyo-based startup Sakana AI says it has developed a system capable of accelerating artificial intelligence development and deployment by up to 100 times, as the race to improve AI efficiency heats up.
The AI CUDA Engineer system, unveiled on Thursday, can speed up the training and inference phase of AI by 10 to 100 times by automating the creation of code used to control Nvidia's graphic processing units, the most widely used GPUs in AI, according to Sakana.
.jpg?width=700&fit=cover&gravity=faces&dpr=2&quality=medium&source=nar-cms&format=auto&height=394)
