Easiest method to create INT8 Calibration Table using TensorRT (trtexec preferrable)

AI & Data Science Deep Learning (Training & Inference) TensorRT
#1

Description

Hi NVIDIA Team,
Can you tell me the easiest method to create INT8 Calibration Table using TensorRT (trtexec preferrable) for a particular caffe/onnx/uff model

Environment

TensorRT Version: 7.0.0.11
GPU Type: T4
Nvidia Driver Version: 440+
CUDA Version: 10.2
CUDNN Version:
Operating System + Version: 18.04
Python Version (if applicable):
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag):

#2

Please refer to below link


Thanks

#3

Hi @SunilJB,
Can you share steps on how to create a dummy calibration table from trtexec as I know trtexec already makes a dummy one when a particular calibration table is not given. I just want to save it into a file.

#4

@GalibaSashi,
You can try to run the model in verbose mode to see the dynamic range setting used in each layer

Thanks