search

YonghaoHe / LFD-A-Light-and-Fast-Detector

LFD is a big update upon LFFD. Generally, LFD is a multi-class object detector characterized by lightweight, low inference latency and superior precision. It is for real-world appilcations.
416 stars 83 forks source link

Error: onnx2tensorRT in NVIDIA TX2 #28

Closed HT-Yuan closed 3 years ago

HT-Yuan commented 3 years ago

Assertion failed: !isDynamic(tensor_ptr->getDimensions()) && "InstanceNormalization does not support dynamic inputs!".

do u have a solution ?

YonghaoHe commented 3 years ago

@HT-Yuan you can check the readme for more details.

HT-Yuan commented 3 years ago

@HT-Yuan you can check the readme for more details.

ok,I found that this problem is caused by the difference of tensorRT version, thank you for your reply.

Manideep08 commented 3 years ago

Did you find any solution for this? I do use tx2 and tensorrt version is 7.1.3. Or any suggestion on deploying this in Jetson devices with jetpack 4.5 (tensorrt version is 7.1.x.x)? @YonghaoHe do you suggest any?

HT-Yuan commented 3 years ago

Did you find any solution for this? I do use tx2 and tensorrt version is 7.1.3. Or any suggestion on deploying this in Jetson devices with jetpack 4.5 (tensorrt version is 7.1.x.x)? @YonghaoHe do you suggest any?

Hello,because of the TensorRT version, I did not run the code on TX2.(I tried to upgrade TensorRT to 7.2, but failed ) 。 In order to achieve fast face detection on TX2 , I used the Ultra face slim algorithm in my project Link: https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB maybe it can help you

Manideep08 commented 3 years ago

Oh okay. My requirement is on a different use case and I have trained the model accordingly. Did you try deploying in onnx or any, which are supported in deepstream?

Then will have to figure a way for this!