import tensorflow as tf

interpreter = tf.lite.Interpreter(model_path="yolo_training/NAO_detector/weights/best_saved_model/best_float16.tflite")
interpreter.allocate_tensors()