YOLOv11 Large achieves an inference time of 20 ms on an NVIDIA 3080 Ti.

[suhyoung89-pro]

help me

[justincdavis]

Hi @suhyoung89-pro

Do you have the commands you used to export to ONNX and generate the engine file? The most immediate solutions you could try are ensuring you simplified the ONNX model and that you used fp16 (or half) during the engine generation.