Optimizing Large CV models using TensorRT and Triton Inference Server - Quantiphi - Quantiphi