Apple embraces Nvidia GPUs to accelerate LLM inference via its open source ReDrafter tech
ReDrafter delivers 2.7x more tokens per second compared to traditional auto-regression ReDrafter could reduce latency for ...