2019.10.11 How to train YOLO with my own dataset? – 14 : Success to detect carrots in various distance, angle
Today, I tested with the model that I still am training with my all dataset including carrot pictures of various distances and angle. The loss of the train becomes around 0.45 to 0.5. Since the training dataset is large – 538 images and xml files, it took time to train the model. The model only could be trained around 50 epochs each day. However, the test to check the distance, angle variance for carrots succeeded as following video. Yolo could detect carrots independent of distance, and angle. 2019.10.11 How to train YOLO with my own dataset? – 14 : Success to detect carrots in various distance, angle 더보기