4 ms·Benchmarking Triton (TensorRT) Inference Server for Transformer Models3 points by julien_c 6y ago