YoloV3 model deployment via Deepstream SDK -Deserialize engine failed because file path::Engine open error

[sriharish.kalidass1]

Please provide complete information as applicable to your setup.

**• Hardware Platform (Jetson / GPU) - Jetson Xavier
**• DeepStream Version - 5.0
**• JetPack Version - 4.4
**• TensorRT Version -7.1.3
• NVIDIA GPU Driver Version (valid for GPU only)
**• Issue Type - I am trying to run YoloV3 model based on DeepStream sample application based on NVIDIA DeepStream SDK.
Please see the link below for more information;

GitHub - NVIDIA-AI-IOT/deepstream_tao_apps: Sample apps to demonstrate how to...

I have also sucessfully manage to build the TRT-OSS application which I believe is required to run YoloV3 model.

When I tried running this command;
./apps/tao_detection/ds-tao-detection -c configs/yolov3_tao/pgie_yolov3_tao_config.txt -i sample_outdoor_car_1080p_10fps.h264 -d

I had the following error;

[AastaLLL]

Hi,

We have new Deepstream 6.0 + JetPack 4.6 release, would you mind updating your software to the latest first?
The GitHub sample is also verified with Deepstream 6.0.

Prerequisites

• Deepstream SDK 6.x GA
  …
• TensorRT OSS (release/8.x branch)
  …

Thanks.

[sriharish.kalidass1]

Thank you so much for your response.

I am planning to upgrade to Deepstream 6.0 and I just found out that this will require Jetpack 4.6?

I believe currently the Jetson Xavier NX is configured with Jet pack 4.4 and hosts Deepstream 5.0 where I am able to run a few Deepstream test applications. Does, upgrading/flashing new Jetpack affect existing files and applications on Jetson? Do I need to do a back up before upgrading/flashing Jetpack 4.6?

I am concerned that I may lose existing files and Deepstream applications that are already hosted on Jetson.

Looking forward to hearing from you.

Many thanks, Sri

[AastaLLL]

Hi,

Yes. Deepstream 6.0 requires a JetPack4.6 environment.
And it’s recommended to backup the data first.

Thanks.