diff --git a/y2023/vision/yolov5.h b/y2023/vision/yolov5.h
index 23cefd5..ad04350 100644
--- a/y2023/vision/yolov5.h
+++ b/y2023/vision/yolov5.h
@@ -1,11 +1,6 @@
 #ifndef Y2023_VISION_YOLOV5_H_
 #define Y2023_VISION_YOLOV5_H_
 
-#include <tensorflow/lite/interpreter.h>
-#include <tensorflow/lite/kernels/register.h>
-#include <tensorflow/lite/model.h>
-#include <tflite/public/edgetpu_c.h>
-
 #include <chrono>
 #include <cmath>
 #include <cstdint>
@@ -29,55 +24,18 @@
 
 class YOLOV5 {
  public:
+  virtual ~YOLOV5();
+
   // Takes a model path as string and loads a pre-trained
   // YOLOv5 model from the specified path.
-  void LoadModel(const std::string path);
+  virtual void LoadModel(const std::string path);
 
   // Takes an image and returns a Detection.
-  std::vector<Detection> ProcessImage(cv::Mat image);
-
- private:
-  // Convert an OpenCV Mat object to a tensor input
-  // that can be fed to the TensorFlow Lite model.
-  void ConvertCVMatToTensor(const cv::Mat &src, uint8_t *in);
-
-  // Resizes, converts color space, and converts
-  // image data type before inference.
-  void Preprocess(cv::Mat image);
-
-  // Converts a TensorFlow Lite tensor to a 2D vector.
-  std::vector<std::vector<float>> TensorToVector2D(TfLiteTensor *src_tensor,
-                                                   const int rows,
-                                                   const int columns);
-
-  // Performs non-maximum suppression to remove overlapping bounding boxes.
-  void NonMaximumSupression(const std::vector<std::vector<float>> &orig_preds,
-                            const int rows, const int columns,
-                            std::vector<Detection> *detections,
-                            std::vector<int> *indices);
-  // Models
-  std::unique_ptr<tflite::FlatBufferModel> model_;
-  std::unique_ptr<tflite::Interpreter> interpreter_;
-  tflite::StderrReporter error_reporter_;
-
-  // Parameters of interpreter's input
-  int input_;
-  int in_height_;
-  int in_width_;
-  int in_channels_;
-  int in_type_;
-
-  // Parameters of original image
-  int img_height_;
-  int img_width_;
-
-  // Input of the interpreter
-  uint8_t *input_8_;
-
-  // Subtract this offset from class labels to get the actual label.
-  static constexpr int kClassIdOffset = 5;
+  virtual std::vector<Detection> ProcessImage(cv::Mat image);
 };
 
+std::unique_ptr<YOLOV5> MakeYOLOV5();
+
 }  // namespace vision
 }  // namespace y2023