Dfine Medium Obj365

🚀 D-FINE

D-FINE is a powerful real-time object detector. It redefines the bounding box regression task in DETR models, achieving outstanding localization precision. This repository provides the HF transformers implementation for D-FINE.

🚀 Quick Start

Prerequisites

• Python 3.x
• PyTorch
• transformers library

Installation

You can install the necessary libraries using the following command:

pip install torch transformers requests pillow

Usage Example

import torch
import requests

from PIL import Image
from transformers import DFineForObjectDetection, AutoImageProcessor

url = 'http://images.cocodataset.org/val2017/000000039769.jpg'
image = Image.open(requests.get(url, stream=True).raw)

image_processor = AutoImageProcessor.from_pretrained("ustc-community/dfine-medium-obj365")
model = DFineForObjectDetection.from_pretrained("ustc-community/dfine-medium-obj365")

inputs = image_processor(images=image, return_tensors="pt")

with torch.no_grad():
    outputs = model(**inputs)

results = image_processor.post_process_object_detection(outputs, target_sizes=torch.tensor([image.size[::-1]]), threshold=0.3)

for result in results:
    for score, label_id, box in zip(result["scores"], result["labels"], result["boxes"]):
        score, label = score.item(), label_id.item()
        box = [round(i, 2) for i in box.tolist()]
        print(f"{model.config.id2label[label]}: {score:.2f} {box}")

✨ Features

• Redefined Regression Task: D-FINE redefines the bounding box regression task in DETR models, achieving outstanding localization precision.
• Two Key Components: It comprises Fine-grained Distribution Refinement (FDR) and Global Optimal Localization Self-Distillation (GO-LSD).
• Multiple Datasets Support: Trained on COCO and Objects365 datasets, with different training configurations available (_coco, _obj365, _obj2coco).
• Wide Applications: Ideal for real-time object detection in diverse applications such as autonomous driving, surveillance systems, robotics, and retail analytics.

📚 Documentation

Overview

The D-FINE model was proposed in D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement by Yansong Peng, Hebei Li, Peixi Wu, Yueyi Zhang, Xiaoyan Sun, Feng Wu. This model was contributed by VladOS95-cyber with the help of @qubvel-hf.

Performance

D-FINE is a powerful real-time object detector that redefines the bounding box regression task in DETR models to achieve outstanding localization precision.

Training

D-FINE is trained on COCO and Objects365 (Lin et al. [2014]) train2017 and validated on COCO + Objects365 val2017 dataset. Standard AP metrics (averaged over uniformly sampled IoU thresholds ranging from 0.50 - 0.95 with a step size of 0.05) and APval5000 commonly used in real scenarios are reported.

Applications

D-FINE is suitable for real-time object detection in various applications, including:

• Autonomous Driving: Provide accurate object detection for self-driving vehicles.
• Surveillance Systems: Monitor and detect objects in real-time.
• Robotics: Enable robots to perceive and interact with the environment.
• Retail Analytics: Analyze customer behavior and object presence in retail stores.

📄 License

This project is licensed under the Apache-2.0 license.