yolo-object-detection

Custom YOLO model trained on my own dataset for object detection practice

YOLOv11 — Detection & Segmentation

This repository contains my assignment: training and inference for object detection/segmentation using Ultralytics YOLO (custom dataset).

Files included

• notebooks/YOLO_Detection_and_Segmentation.ipynb — main Colab notebook (how I trained / ran inference).
• models/best.pt — trained model checkpoint (best).
• data/data.yaml — dataset config (train/val/test paths and class names).
• dataset/Task11.v1i.yolov11.zip — Roboflow dataset export (optional).
• videos/input_video.mp4 — sample input video.
• videos/segmented_output_video1.mp4 — sample output from inference.
• results/results.png — training loss/metrics plot.

How to run (quick)

1. Open the notebook in Colab:
   https://colab.research.google.com/github/Arbab0097/yolo-object-detection/blob/main/notebooks/YOLO_Detection_and_Segmentation.ipynb
2. Install requirements: pip install -r requirements.txt (or run the notebook cells that install dependencies).
3. Load the model: from ultralytics import YOLO; model = YOLO('models/best.pt')
4. Run inference on a video: call the notebook cell that runs infer_video('videos/input_video.mp4','videos/segmented_output_video1.mp4').

Requirements

• ultralytics
• opencv-python
• torch
• torchvision
• numpy

(See requirements.txt for exact versions.)

Notes

• I have included a small dataset export (Roboflow) and a sample video. If any file is missing, please contact me.