A project that optimizes Whisper for low latency inference using NVIDIA TensorRT - View it on GitHub