search

Qengineering / TensorFlow_Lite_Segmentation_Jetson-Nano

TensorFlow Lite segmentation on a Jetson Nano at 11 FPS
https://qengineering.eu/install-tensorflow-2-lite-on-raspberry-pi-4.html
BSD 3-Clause "New" or "Revised" License
14 stars 1 forks source link

GPU DELEGATE #2

Open LuChengYu1119 opened 2 years ago

LuChengYu1119 commented 2 years ago

Hi, If I want to use the version without GPU, where or which code should I uncommented #define GPU_DELEGATE ?

Qengineering commented 2 years ago

Delegates are enabled during the building of the TF-lite library used by the software. If you don't want to use them, link with a TF-lite lib without delegates. See the website.

LuChengYu1119 commented 2 years ago

@Qengineering Thanks for your reply, In fact, I did exactly what the website said, except after the GPU delegate. But when I want to run the program through code::block, it pops out that I don't have libtensorflowlite_gpu_delegate.so error.

Qengineering commented 2 years ago

Did you remove the reference in the Code::Blocks project file to the libtensorflowlite_gpu_delegate.so? As you probably know, can use the menu option Project->Build options, or edit the TestTensorFlow_Lite.cbp manually. At line 46 you find the reference. <Add library="/home/jetson/tensorflow/bazel-bin/tensorflow/lite/delegates/gpu/libtensorflowlite_gpu_delegate.so" />