Nvidia LiteVLM cuts AI latency for self-driving cars

The new pipeline achieves a 2.5 reduction in end-to-end latency on the Nvidia Drive Thor platform.
Go to Source