This repository contains code for training a YOLO model to detect a bullseye.
A custom dataset was used to train the model. The dataset consists of images containing bullseyes.
The dataset must follow YOLOv11 standards and must be placed inside a directory named dataset at the project root.
Two primary models were trained:
- best: Trained on my PC
- best-colab: Trained on Google Colab
train.py: Contains all parameters to train the model using Ultralytics library.comp_ts_onnx.py: Compares the inference time of ONNX and TorchScript formats for both best and best-colab models on a Raspberry Pi 4B.
benchmark_model.py: Runs benchmarks on the trained models.benchmarks.log: Contains benchmark results for bothbestandbest-colabmodels.
opencv_test.py: Tests the model on a set of pre-saved bullseye images.picam_test.py: Tests the Raspberry Pi camera feed and evaluates OpenCV’s HoughCircles method for detecting bullseyes (which was found inefficient).yolo_test.py: Runs live inference using the trained YOLO model on a camera stream.
data-linux.yaml&data-win.yaml: YAML files specifying dataset paths (different absolute paths for Linux and Windows).
- The results.csv has training results over 130 epochs.
| Epoch | Train Box Loss | Train Cls Loss | Train DFL Loss | Precision (B) | Recall (B) | mAP50 (B) | mAP50-95 (B) | Val Box Loss | Val Cls Loss | Val DFL Loss | LR PG0 | LR PG1 | LR PG2 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 3.8923 | 3.30073 | 4.08417 | 0.01545 | 0.46829 | 0.0866 | 0.06706 | 3.62925 | 3.66435 | 4.14368 | 0.0006 | 0.0006 | 0.0006 |
| 2 | 2.87993 | 1.98997 | 3.08428 | 0.00706 | 0.51951 | 0.18509 | 0.12929 | 3.6334 | 3.48378 | 4.14384 | 0.00121 | 0.00121 | 0.00121 |
| 3 | 1.96915 | 1.49711 | 2.19448 | 0.23748 | 0.34634 | 0.15484 | 0.06874 | 2.9761 | 2.70784 | 3.03805 | 0.00182 | 0.00182 | 0.00182 |
| ..... | .............. | .............. | .............. | ............. | .......... | ........... | .......... | ............ | ............ | ............ | ....... | ....... | ....... |
| 129 | 0.80191 | 0.58796 | 1.18512 | 0.99647 | 0.97805 | 0.97559 | 0.79891 | 0.74167 | 0.3649 | 1.15386 | 0.00073 | 0.00073 | 0.00073 |
| 130 | 0.79106 | 0.58632 | 1.17945 | 0.99751 | 0.97783 | 0.97597 | 0.78646 | 0.77879 | 0.3609 | 1.16972 | 0.00072 | 0.00072 | 0.00072 |
| 131 | 0.79467 | 0.58256 | 1.17882 | 0.99669 | 0.97805 | 0.97628 | 0.79426 | 0.75459 | 0.34918 | 1.15876 | 0.00071 | 0.00071 | 0.00071 |
