nanmi / YOLOX-deepstream

deploy yolox algorithm use deepstream
89 stars 27 forks source link

Results not shown #8

Closed JustdoITcom closed 2 years ago

JustdoITcom commented 2 years ago

Thanks for your job

I am currently using xavier nx. I want to use deepstream there, but the detection result is not displayed.

I got the model_trt.engine pth file using trt.py provided at https://github.com/Megvii-BaseDetection/YOLOX address.

Do I have to get the engine file using the method you mentioned at https://zhuanlan.zhihu.com/p/391693130?

Thank you

lantudou commented 2 years ago

Hi, Please refer to this issue

5

JustdoITcom commented 2 years ago

I solved this problem by following your instructions.

Thank you so much for your help.

tulbureandreit commented 2 years ago

@JustdoITcom could you offer a bit more detail about your problem?

JustdoITcom commented 2 years ago

@JustdoITcom could you offer a bit more detail about your problem?

I made the engine file properly and turned on deepstream detection, but I didn't get any result box on the screen.

However, thanks to lantudou, I solved it with net-scale-factor=1 instead of net-scale-factor=0.0039215697906911373.

tulbureandreit commented 2 years ago

@JustdoITcom thanks. which kind of model did you use? I struggle using yolox-nano or tiny. Inference won`t run unless I use yolox-s

JustdoITcom commented 2 years ago

@JustdoITcom thanks. which kind of model did you use? I struggle using yolox-nano or tiny. Inference won`t run unless I use yolox-s

I also used yolox-s . I haven't tried tiny and nano.

lantudou commented 2 years ago

please check if your model input setting is correct. because the yolox-tiny and nano input size is 416 and 416. not 640*640

---Original--- From: @.> Date: Mon, Feb 14, 2022 10:49 AM To: @.>; Cc: @.**@.>; Subject: Re: [nanmi/YOLOX-deepstream] Results not shown (#8)

@JustdoITcom thanks. which kind of model did you use? I struggle using yolox-nano or tiny. Inference won`t run unless I use yolox-s

I also used yolox-s . I haven't tried tiny and nano.

— Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. You are receiving this because you commented.Message ID: @.***>

tulbureandreit commented 2 years ago

@lantudou yep. Can jetson xavier support depthwise convolution operations?

lantudou commented 2 years ago

you can check the supported operators for your installed tensorrt version. i think the answer is yes

---Original--- From: @.> Date: Mon, Feb 14, 2022 20:24 PM To: @.>; Cc: @.**@.>; Subject: Re: [nanmi/YOLOX-deepstream] Results not shown (#8)

@lantudou yep. Can jetson xavier support depthwise convolution operations?

— Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. You are receiving this because you were mentioned.Message ID: @.***>