GitHub - microsoft/DeepSpeed-MII: MII makes low-latency and high-throughput inference possible, powered by DeepSpeed.

Open in new window