DOCSEARCHbyAlgolia
Submit Docs

Roboflow Notebooks

repository·main·Indexed 27 days ago

https://github.com/roboflow/notebooks

A collection of 61 Jupyter notebooks providing practical tutorials for computer vision tasks, including object detection, segmentation, pose estimation, data extraction, and OCR. The repository features implementation and fine-tuning guides for state-of-the-art models such as YOLO (v5 through v12), SAM 2, Florence-2, PaliGemma 2, Qwen2.5-VL, and RT-DETR. Notebooks are designed for execution in cloud environments like Google Colab, Kaggle, and SageMaker Studio Lab.

Tokens
16.7K
Snippets
93
Records
112
Agent score
94%

What's inside roboflow-notebooks

  1. Explore Roboflow Computer Vision Tutorials

    main

    This repository provides a collection of computer vision tutorials using State-of-the-Art (SOTA) models. You can use these notebooks to learn implementation details for various tasks including:

    • Object Detection
    • Segmentation
    • Pose Estimation
    • Data Extraction
    • OCR

    Supported models include YOLOv11, SAM 2, Florence-2, PaliGemma 2, and Qwen2.5-VL.

  2. Explore Computer Vision Skills Notebooks

    main
    The repository contains a collection of 23 specialized notebooks focused on advanced computer vision skills. These notebooks cover topics such as object detection, tracking, segmentation, and automated labeling. Most notebooks can be opened directly in Google Colab, Kaggle, or SageMaker Studio Lab via the links provided in the repository.

  3. Fine-Tune YOLOv7 for various tasks

    main

    The repository provides specialized notebooks for fine-tuning YOLOv7 on different computer vision tasks:

    • Object Detection: train-yolov7-object-detection-on-custom-data.ipynb
    • Instance Segmentation: train-yolov7-instance-segmentation-on-custom-data.ipynb

  4. Object Detection and Tracking with YOLO and Supervision

    main

    Several notebooks demonstrate how to combine YOLO models (v5, v7, v8) with the supervision library for advanced tasks:

    • Vehicle Speed Estimation: Estimate vehicle speed using YOLOv8.
    • Polygon Zone Counting: Detect and count objects within a specific polygon zone using YOLOv5/v8 or Detectron2.
    • Vehicle Tracking and Counting: Track and count vehicles using YOLOv8 combined with ByteTRACK.
    • Football Player Tracking: Track football players using YOLOv5 and ByteTRACK.

  5. Run Roboflow Notebooks locally

    main

    To run the notebooks on your own machine, clone the repository, set up a Python virtual environment to avoid global dependency conflicts, and install Jupyter Notebook.

    Prerequisites:

    • Git
    • Python 3
    • venv module
    # clone repository and navigate to root directory
git clone git@github.com:roboflow-ai/notebooks.git
cd notebooks

# setup python environment and activate it
python3 -m venv venv
source venv/bin/activate

# install and run jupyter notebook
pip install notebook
jupyter notebook

  6. Fine-Tune YOLOv4-tiny Darknet on Object Detection Dataset

    main
    This notebook provides a tutorial for fine-tuning a YOLOv4-tiny Darknet model on a custom object detection dataset. It includes links to run the notebook on Google Colab or Kaggle, and provides supplementary resources via Roboflow blog posts and YouTube tutorials.

  7. Track objects using RF-DETR and ByteTrack

    main

    Learn how to implement object tracking by combining the RF-DETR model with the ByteTrack tracker. This tutorial is available via interactive notebooks in Colab, Kaggle, or SageMaker Studio Lab and includes complementary video content.

    https://github.com/roboflow-ai/notebooks/blob/main/notebooks/how-to-track-objects-with-bytetrack-tracker.ipynb

  9. Train a YOLOv8 Classification Model with No Labeling

    main
    This notebook demonstrates how to train a YOLOv8 classification model without manual labeling. It is compatible with Google Colab and Kaggle, and includes a Roboflow blog post guide for the workflow.

  10. Update the Roboflow Notebooks list in README.md

    main

    To add a new notebook to the repository's main table, do not edit README.md manually. Instead, follow these steps to ensure the table is updated automatically via the automation script:

    1. Add your new .ipynb notebook file to the notebooks directory.
    2. Add a corresponding new entry to the automation/notebooks-table-data.csv file.
    3. Run the autogeneration script to update the README.md table:
      python3 automation/autogenerate_notebooks_table.py
    4. Commit the changes and create a Pull Request.
    python3 automation/autogenerate_notebooks_table.py

  11. Automated Labeling with Grounding DINO and SAM

    main
    Learn how to perform automated dataset annotation and evaluation using Grounding DINO and Segment Anything (SAM). These notebooks provide workflows for enhancing image annotation efficiency by leveraging these foundation models.

  12. Track objects using RF-DETR and OC-SORT

    main

    Learn how to implement object tracking by combining the RF-DETR model with the OC-SORT (Observation-Centric SORT) tracker. This tutorial is available via interactive notebooks in Colab, Kaggle, or SageMaker Studio Lab and includes complementary video content.

    https://github.com/roboflow-ai/notebooks/blob/main/notebooks/how-to-track-objects-with-ocsort-tracker.ipynb