ankandrew/open-image-models
ankandrew/open-image-models has 126 GitHub stars and 27 forks. GitHot measured 0 stars in 24 hours. Last checked 2026-11-16 19:10 UTC. Last push 2026-07-27.
What you can do
- 🚀 Pre-trained: Models are ready for immediate use, no additional training required.
- 🌟 ONNX: Cross-platform support for fast inference on both CPU and GPU environments.
- ⚡ Performance: Optimized for both speed and accuracy, ensuring efficient real-time applications.
- 💻 Simple API: Power up your applications with robust model inference in just one line of code.
Measured star growth
- in 24h
- 0
- in 7 days
- Collecting history
- in 15 days
- Collecting history
- in 1 month
- Collecting history
Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.
About
Pre-trained image models using ONNX for fast, out-of-the-box inference.
Developer ToolsREADME.md
OriginalOpen Image Models
Introduction
Ready-to-use models for a range of computer vision tasks like detection, classification, and more. With ONNX support, you get fast and accurate results right out of the box.
Easily integrate these models into your apps for real-time processing—ideal for edge devices, cloud setups, or production environments. In one line of code, you can have powerful model inference running!
from open_image_models import create_detector
detector = create_detector("yolo-v9-t-256-license-plate-end2end")
detector.predict("path/to/license_plate_image.jpg")
✨ That's it! Powerful license plate detection with just a few lines of code.
Features
- 🚀 Pre-trained: Models are ready for immediate use, no additional training required.
- 🌟 ONNX: Cross-platform support for fast inference on both CPU and GPU environments.
- ⚡ Performance: Optimized for both speed and accuracy, ensuring efficient real-time applications.
- 💻 Simple API: Power up your applications with robust model inference in just one line of code.
Installation
To install open-image-models via pip, use the following command:
pip install open-image-models[onnx]
[!NOTE] For hardware acceleration, you can use one of the following extras instead:
onnx-gpu,onnx-openvino,onnx-directml, oronnx-qnn. Example:pip install open-image-models[onnx-gpu]
Available Models
Object Detection
COCO Detection
| Model | Image Size | Classes | Dynamic Batch |
|---|---|---|---|
rf-detr-nano-384-coco |
384 | 80 | Yes |
rf-detr-small-512-coco |
512 | 80 | Yes |
rf-detr-medium-576-coco |
576 | 80 | Yes |
rf-detr-large-704-coco |
704 | 80 | Yes |
CoreML is excluded from automatic RF-DETR provider selection.
Plate Detection

| Model | Image Size | Precision (P) | Recall (R) | mAP50 | mAP50-95 |
|---|---|---|---|---|---|
yolo-v9-s-608-license-plate-end2end |
608 | 0.957 | 0.917 | 0.966 | 0.772 |
yolo-v9-t-640-license-plate-end2end |
640 | 0.966 | 0.896 | 0.958 | 0.758 |
yolo-v9-t-512-license-plate-end2end |
512 | 0.955 | 0.901 | 0.948 | 0.724 |
yolo-v9-t-416-license-plate-end2end |
416 | 0.94 | 0.894 | 0.94 | 0.702 |
yolo-v9-t-384-license-plate-end2end |
384 | 0.942 | 0.863 | 0.92 | 0.687 |
yolo-v9-t-256-license-plate-end2end |
256 | 0.937 | 0.797 | 0.858 | 0.606 |
Usage
import cv2
from rich import print
from open_image_models import create_detector
# Create a detector from a pretrained model
detector = create_detector("yolo-v9-t-384-license-plate-end2end")
# Load an image
image_path = "path/to/license_plate_image.jpg"
image = cv2.imread(image_path)
# Perform license plate detection
detections = detector.predict(image)
print(detections)
# Display predictions on the image
annotated_image = detector.display_predictions(image)
# Show the annotated image
cv2.imshow("Annotated Image", annotated_image)
cv2.waitKey(0)
cv2.destroyAllWindows()
[!TIP] Checkout the docs!
Contributing
Contributions to the repo are greatly appreciated. Whether it's bug fixes, feature enhancements, or new models, your contributions are warmly welcomed.
To start contributing or to begin development, you can follow these steps:
- Clone repo
git clone https://github.com/ankandrew/open-image-models.git - Install all dependencies using Poetry:
make install - To ensure your changes pass linting and tests before submitting a PR:
make checks
Citation
@article{wang2024yolov9,
title={{YOLOv9}: Learning What You Want to Learn Using Programmable Gradient Information},
author={Wang, Chien-Yao and Liao, Hong-Yuan Mark},
booktitle={arXiv preprint arXiv:2402.13616},
year={2024}
}