Hi there, Our startup needs to be able to record video while also performing inference on each frame using TensorRT. Our use case needs the frame rate to be as high as possible, and at least above 10FPS. I am facing and issue where when I try to run a python script which records video while also p…

FYI, I found a work-around to this bug in jetson-io.py that makes it complain about a missing “APP_b” partition. Check out my last comment here: [image] SPI_IOC_MESSAGE ioctl for SPI extremely slow Jetson Orin Nano I found the bug in jetson-io.py and implemente…

Hi, We have not replicated the issue yet. From your experience, how long does it take to observe the issue?

It should be practically instantly when used with the x264 encoder script above. When saving images using a jog encoder it can take 5 to 60 mins

Hi, It looks like we also need jetson_orin_dynamic_output.trt. Could you zip one package with all files and share the full steps? So that we can simply execute the steps one-by-one to replicate the issue.

I already attached it above in the Google drive link @DaneLLL . Look at May 22 message

Hi is there any update? @DaveYYY I got an email that you were confused about how to import files in Python which looks like you might have deleted. If you’re still confused, let me know; in this specific example above, I should have written import augmentations, but I trust that you guys are able t…

How frequently did you hit the issue? We ran the code for more than 30 minutes without crashing, but the output trensor is all zero. Is that expected? We cannot measure FPS because apparrently you didn’t include that part of the code, but it worked well otherwise.

Hi @DaveYYY , Thanks for running the experiment! I’ll be back in the office tomorrow morning Irish time and will re-run the script on my two local cameras to confirm, however the issue was hit on all at all eight of our cameras, where it would usually happen instantly with this specific script if …

I tried on Orin Nano with JetPack 5.1.3 + an IMX 219 camera, instead of your 5.1.1 + IMX 477. If you are using DevKit, upgrading from 5.1.1 to 5.1.3 via APT should be easy, if it is really the JetPack version that makes a difference.

Sorry would you excuse my ignorance and tell me how I can upgrade to 5.1.3 via apt from 5.1.1? I thought that you had to manually take out and flash the SD card, but sounds like there’s a much easier way?

X264 and TensorRT sudden reboot (MJPG encoder not affected, but not fast enough) on Jetson Orin Nano

Robotics & Edge Computing Jetson Systems Jetson Orin Nano

timf34 May 30, 2024, 3:08pm 12

Hi @DaneLLL,

Wondering if there’s any update?

Especially regarding “No GPU found” issue? This is quite a critical issue for us which needs to be fixed by itself so will reopen an issue if not - will help future people with the same issue find it directly through the title too when search for the same issue.

Thanks
Tim

Topic		Replies	Views
GPU not detected on Nvidia Jetson Orin Nano after sudden reboot (I need to run TensorRT engine inference) Jetson Orin Nano cuda	2	467	May 23, 2024
JetPack 4.3 - L4T R32.3.1 released Jetson Nano opencv	98	23201	June 24, 2020
JetPack 4.6 Production Release with L4T 32.6.1 Jetson Nano	47	12559	March 10, 2022
Yolov6 Slow inference speed on the Nvidia Jetson NX board Jetson Xavier NX yolo	8	1756	August 24, 2022
Jetson Orin Nano Developer Kit, Jetpack, Cuda, Tensorflow with GPU and TensorRT Jetson Orin Nano tensorflow	16	4067	September 28, 2023
JetPack 4.4 Developer Preview - L4T R32.4.2 released Jetson Nano	58	8953	July 9, 2020
JetPack 3.1 — L4T R28.1 released for Jetson TX1/TX2 Jetson TX2	42	8708	October 20, 2017
Failed to flash Jetson Orin Nano Jetson Orin Nano reflash	6	1609	September 3, 2023
Jetson AGX Orin FAQ Jetson AGX Orin kb	4	26695	December 21, 2022
Orin Nano Black Screen Jetson Orin Nano boot	9	436	July 31, 2024

X264 and TensorRT sudden reboot (MJPG encoder not affected, but not fast enough) on Jetson Orin Nano

Related topics