TensorRT 6.0.1: Recommended sample to optimize resnet50 models to INT8 with data calibration

Hi all, I was using TRT5.1.5 trtexec sample to optimize models to INT8; however, with the latest version of TRT 6.0.1 the script does not provide calibration capability.

Could you please recommend a sample script in the new TensorRT 6.0.1 to optimize resnet50 models (TensorFlow or ONNX) to INT8 using Imagenet calibration dataset?.

Hi,
Can you provide the following information so we can better help?

Provide details on the platforms you are using:

  • Linux distro and version
  • GPU type
  • Nvidia driver version
  • CUDA version
  • CUDNN version
     - Python version [if using python]
  • Tensorflow version
  • TensorRT version
  • If Jetson, OS, hw versions

Also, could you please share more details about the issues you are getting while using trtexec on latest build?

Please refer to below sample for INT8 optimization:
https://github.com/NVIDIA/TensorRT/tree/release/6.0/samples/opensource/sampleINT8