Hi, i’ve installed TensorFlow v2.5.0 on my Jetson Nano following these instructions Installing TensorFlow For Jetson Platform :: NVIDIA Deep Learning Frameworks Documentation . I use
tegrastats command to monitor GPU usage: from that i can see that TF model is running on GPU but the TFLite model keeps GPU to 0%. I load TFLite interpreter using
interpeter = tf.lite.Interpreter(model_path=‘lite_model.tflite’)
Did I skip some step? Is there a way to fix it? Or doesn’t Jetson Nano have GPU Support for TensorFlow Lite?