Hi Team,
How to build tensorrt engine for ssd_mobilenet_v2_coco architecture on jetson tx2 nx ?
We are using jetpack version 4.5.1 .
We are facing following error while building the engine file ,
import error : libcublas.so.10.0 no such file or directory
Hi,
It depends on which model format you use.
For TensorFlow 1.15.x model, please convert it via .pb -> .uff -> .plan
.
https://github.com/NVIDIA/TensorRT/tree/main/samples/sampleUffMaskRCNN
For other format, please convert it via .xxx -> .onnx -> .plan
.
https://github.com/NVIDIA/TensorRT/tree/main/samples/sampleOnnxMNIST
Based on your error log, the sample has some dependency on CUDA 10.0.
But JetPack 4.5.1 includes CUDA 10.2 so you can find it on the TX2.
Thanks.
system
Closed
5
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.