I was using tensorrt to convert a yolov8s model, and I got this warning, and the detection result was so bad.
[W] [TRT] Half2 support requested on hardware without native FP16 support, performance will be negatively affected.
and I found this
my question is
my GPU’s compute capability is 6.1 and it’s support FP16, why there is still the warning?
and is this warning actually the reason that affect the performance?
TensorRT Version: 126.96.36.199-1+cuda11.3
GPU Type: NVIDIA TITAN Xp
Nvidia Driver Version: 510.85.02
CUDA Version: 11.3
CUDNN Version: 188.8.131.52-1+cuda11.3
Operating System + Version: ubuntu 20.04
Python Version (if applicable):
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag):