Please provide complete information as applicable to your setup.
• Hardware Platform (Jetson / GPU) Jetson AGX Xavier
• DeepStream Version 5.0
• JetPack Version (valid for Jetson only) 5.0.2
• TensorRT Version 7.1.3.0+cuda10.2
• NVIDIA GPU Driver Version (valid for GPU only)
• Issue Type( questions, new requirements, bugs) question
• 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)
Hi,
I am running a system with an old Jetson(AGX Xavier) to detect vehicle.
I have been using yolov5 and it worked fine. But now, I want to use yolov7 instead so I found some repositories like this. Although its requirement is TensorRT 8.0+, I first converted yolov7-tiny.pt(official model) to yolov7-tiny.wts then to yolov7-tiny.engine successfully by using the methods they provide. I also tested the engine model that was no problem.
Unfortunately, when I configured a yolov7 model customized by myself(also tiny) in the same way, an exception occurred. The bbox output by the model is not drawn correctly as shown in the figure1.
The command I used to train yolov7 is python train.py --device 0 --epochs 100 --batch-size 8 --data dataset/dataset/train4v7.yaml --img 416 416 --cfg cfg/training/yolov7-tiny.yaml --weights ./yolov7-tiny.pt
I thought the model itself was not trained well before I checked its pt model. But as shown in the figure2, the bbox of the pt model is obviously drawn correctly.
I guess it’s because my TensorRT version is too old. But it is hard to upgrade it due to it has been deployed to the field.
I’m not sure if my guess is correct. Do you know the real reason? Please tell me.
Thanks