tflite model inference speed is slow on limited hardware

Question

I have a text classification model that I converted to tflite. I am running the tflite model on a linux SBC running debian 10 on armv7 architecture with 512 mb of ram. Running a