구현: 라즈베리파이 tflite mobilenet inference 구현

deep_learning/pi/install.sh

@@ -0,0 +1,15 @@
+#!/bin/bash
+
+# TFLite runtime
+echo "deb https://packages.cloud.google.com/apt coral-edgetpu-stable main" | sudo tee /etc/apt/sources.list.d/coral-edgetpu.list
+curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo apt-key add -
+sudo apt-get update
+sudo apt-get install python3-tflite-runtime
+
+# OpenCV
+sudo apt-get -y install libjpeg-dev libtiff5-dev libjasper-dev libpng12-dev
+sudo apt-get -y install libavcodec-dev libavformat-dev libswscale-dev libv4l-dev
+sudo apt-get -y install libxvidcore-dev libx264-dev
+sudo apt-get -y install qt4-dev-tools libatlas-base-dev
+sudo apt-get -y install libhdf5-dev libqtgui4  libqt4-test
+pip3 install opencv-python

deep_learning/pi/run_inference.py

@@ -0,0 +1,52 @@
+import argparse
+import os
+import time
+from pathlib import Path
+
+import tflite_runtime.interpreter as tflite
+import cv2
+import numpy as np
+
+
+def read_input_images(data_path):
+  img_files = data_path.rglob('*.jpg')
+  imgs = [cv2.imread(str(img_file)) for img_file in img_files]
+  imgs = [cv2.cvtColor(img, cv2.COLOR_BGR2RGB) for img in imgs]
+
+  return imgs
+
+
+def main():
+  parser = argparse.ArgumentParser()
+  parser.add_argument('-m', '--model_file', help="path to tflite model file")
+  parser.add_argument('-d',
+                      '--data_path',
+                      help="path to dataset to run inference")
+  args = parser.parse_args()
+
+  interpreter = tflite.Interpreter(model_path=args.model_file)
+  interpreter.allocate_tensors()
+
+  input_details = interpreter.get_input_details()
+  output_details = interpreter.get_output_details()
+  input_shape = input_details[0]['shape']
+
+  images = read_input_images(Path(args.data_path))
+
+  for image in images:
+    input_data = np.expand_dims(image, axis=0)
+    input_data = np.float32(input_data)
+
+    interpreter.set_tensor(input_details[0]['index'], input_data)
+    start_time = time.time()
+    interpreter.invoke()
+    stop_time = time.time()
+
+    print('Inference time: {:.3f}'.format((stop_time - start_time) * 1000))
+
+    prediction = interpreter.get_tensor(output_details[0]['index'])
+    print(f'Prediction: {prediction[0]}')
+
+
+if __name__ == "__main__":
+  main()