Int8 callibration with point cloud

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

Description

A clear and concise description of the bug or issue.

Environment

TensorRT Version: 8.2.1
GPU Type: Jetson Xavier
Nvidia Driver Version:
CUDA Version: 10.2
CUDNN Version:
Operating System + Version:
Python Version (if applicable):
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag):

I want to convert my model from onnx to trt with int8 precision. When one want’s to convert model that precesses images he/she should made calibration on images. But I’m working with point clouds the dimensions of input tensor is (batch_size, 3, 1024) 1024 points with distance in x, y, z directions.

Is there some guide or suggestions how to calibrate on point clouds?

2

Hi, Please refer to the below links to perform inference in INT8

Thanks!