Yolov4 int8 calibration steps required

Please provide complete information as applicable to your setup.

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) : NVIDIA Corporation GA106 [Geforce RTX 3050]
• DeepStream Version: 6.2
• JetPack Version (valid for Jetson only)
• TensorRT Version: with SDK
• NVIDIA GPU Driver Version (valid for GPU only): 525.85.12
*• Cuda Version: 11.8
• Issue Type( questions, new requirements, bugs)
• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)

I want to int8 calibrate yolov4 models with my GPU. Can someone please suggest step by step process for the same.

Have you generate your yolov4 model as ONNX model? You can download the tool TensorRT/tools/Polygraphy at main · NVIDIA/TensorRT (github.com), and follow the steps in this sample TensorRT/tools/Polygraphy/examples/cli/convert/01_int8_calibration_in_tensorrt at main · NVIDIA/TensorRT (github.com) to generate INT8 calibartion file.

Hi Fiona,

Thanks for prompt reply. I have not generated onnx model yet. I will follow the steps and update you about the same.

This is the way we generate yolov4 ONNX model. yolo_deepstream/tensorrt_yolov4 at main · NVIDIA-AI-IOT/yolo_deepstream (github.com)

Hi @Fiona.Chen ,

I got the onnx file. Can you tell me what is the next step as it is bit confusing and I am not sure if I am doing the right thing.

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

Please make sure you know about your model and set the gst-nvinfer configuration correctly. You can refer to the sample NVIDIA-AI-IOT/yolo_deepstream: yolo model qat and deploy with deepstream&tensorrt (github.com), since there may be some differences between your model and the model in our sample, please modify the corresponding customization parts according to your model.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.