How to use YoloV4

[deveso]

I am currently using YoloV3 with nvdsinfer_custom_impl_Yolo. I would now like to use YoloV4 and convert my .cfg and .weights files within deepstream. However, I get the error “Yolo type is not defined from config file name”. How can I use yolov4? There is also no yolov4 in the samples.

config_infer_primary_yoloV4.txt (3.6 KB)
deepstream_app_config_yoloV4.txt (5.5 KB)

[yingliu]

Please refer the sample here: yolo_deepstream/deepstream_yolo at main · NVIDIA-AI-IOT/yolo_deepstream · GitHub.

[deveso]

Thank you that works. However, I only get 32 ​​FPS with my Jetson Orin NX. With YoloV3 I get 180FPS in comparison

[yuweiw]

Are your pipelines for yolov3 and yolov4 the same?

Please provide complete information as applicable to your setup.

Hardware Platform (Jetson / GPU)
• DeepStream Version
• JetPack Version (valid for Jetson only)
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)
• 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)
• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description) GLib, Gst, GstRtspServer, GObject

[deveso]

This is my configuration file for yoloV3:

config_infer_primary_yoloV3.txt (3.7 KB)
deepstream_app_config_yoloV3.txt (5.5 KB)

Here is a similar post from me. However, this time it is a different bug with yolov4. I think it might have something to do with nvdsinfer_custom_impl_yolo though.

Orin NX performance issue

Yes, the Jetson is running at maximum power. Jetson_clocks is active and NVP model 0 at MAXN. The Jetson consumes almost 25W when I run deepstream. So that shouldn’t be the problem. @adventuredaisy I’ve been following your work for a while. You already have experience with the performance of Xavier and Orin. Can you maybe share some information? With your Xavier NX you get up to 55fps with Yolo. What performance can be expected with your Orin dev kit with the same configuration? Did you get you…

I tested your suggestion with this github repository:

config_infer_primary.txt (3.6 KB)
deepstream_app_config.txt (4.7 KB)

config_infer_primary_yoloV4.txt (3.5 KB)
deepstream_app_config_yolo.txt (5.0 KB)

I also tested this Github repository:

GitHub - marcoslucianops/DeepStream-Yolo: NVIDIA DeepStream SDK 6.3 / 6.2 /...

config_infer_primary_yoloV4.txt (3.6 KB)
deepstream_app_config_yoloV4.txt (5.5 KB)

I had a comparable low fps with yolov4 with both implementations. I also tried different network modes.

[yuweiw]

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

1.Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
• DeepStream Version
• JetPack Version (valid for Jetson only)
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)

2.How do you get the fps?

3.Could you use both our model to test it? deepstream_tao_apps This project have both our yolov3, yolov4 model.