How to generate calibration table for Yolov2 with TensorRT 7.0


Deepstream 5.0 releases the int8 calib table for yolov3 but not yolov2.

With Deepstream 5.0, it is possible to generate a calib table for yolov2 and run it in int8 – below link was used.

But for DS5, which uses TRT 7 – there are some changes and the above generated calib table is not working.

I want to know how to generate the calib table for Yolov2 with TRT 7.


TensorRT Version: 7.0
GPU Type: Telsa T4
Nvidia Driver Version: 440.64.00
CUDA Version: 10.2.89
CUDNN Version:
Operating System + Version: Ubuntu 16.04.3 LTS
Python Version (if applicable):
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag):

Relevant Files

Please attach or include links to any models, data, files, or scripts necessary to reproduce your issue. (Github repo, Google Drive, Dropbox, etc.)

Steps To Reproduce

Please include:

  • Exact steps/commands to build your repro
  • Exact steps/commands to run your repro
  • Full traceback of errors encountered

Hi @andy.linluo,
Please refer to the below link to check for setting up the calibrator.

Also, please check if this link works for you


Thanks @AakankshaS. I need Yolo2 calibration table generated with TensorRT7. The calibration table you provided was generated by older version.

Hi ,
You can choose to provide the implementation for writeCalibrationCache() to write calibration table
Something like this
writeCalibrationCache(&(mCalibrationCache[0]), mCalibrationCache.size())