Fashion Object Detection

🚀 Fashion Object Detection Model

This is a fine-tuned object detection model based on microsoft/conditional-detr-resnet-50, specifically designed for fashion-related object detection tasks. It can identify various fashion items in images and is trained on a combination of fashion datasets.

🚀 Quick Start

Installation

To use this model, you need to have the transformers library installed. You can install it using the following command:

pip install transformers

Usage

Here is a basic example of how to use this model for object detection:

from PIL import Image
import torch
from transformers import  AutoImageProcessor, AutoModelForObjectDetection

device = 'cpu'
if torch.cuda.is_available():
    device = torch.device('cuda')
elif torch.backends.mps.is_available():
    device = torch.device('mps')

ckpt = 'yainage90/fashion-object-detection'
image_processor = AutoImageProcessor.from_pretrained(ckpt)
model = AutoModelForObjectDetection.from_pretrained(ckpt).to(device)

image = Image.open('<path/to/image>').convert('RGB')

with torch.no_grad():
    inputs = image_processor(images=[image], return_tensors="pt")
    outputs = model(**inputs.to(device))
    target_sizes = torch.tensor([[image.size[1], image.size[0]]])
    results = image_processor.post_process_object_detection(outputs, threshold=0.4, target_sizes=target_sizes)[0]

    items = []
    for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
        score = score.item()
        label = label.item()
        box = [i.item() for i in box]
        print(f"{model.config.id2label[label]}: {round(score, 3)} at {box}")
        items.append((score, label, box))

✨ Features

• Fine-tuned Model: It is a fine-tuned version of microsoft/conditional-detr-resnet-50, optimized for fashion object detection.
• Multi-dataset Training: Trained on a combination of modanet and fashionpedia datasets, providing a wide range of fashion item recognition capabilities.
• High Performance: Achieved a mAP of 0.7542 at the 96th epoch out of 100 epochs, showing good performance in object detection tasks.

📚 Documentation

• Model Details: You can find more details about the model in this github repo.
• Feature Extractor: A fashion image feature extractor model is available at yainage90/fashion-image-feature-extractor.

🔧 Technical Details

• Base Model: microsoft/conditional-detr-resnet-50
• Training Datasets:
  • modanet
  • fashionpedia
• Labels: ['bag', 'bottom', 'dress', 'hat', 'shoes', 'outer', 'top']
• Performance: The best score was achieved at the 96th epoch with a mAP of 0.7542, indicating there is still some room for performance improvement.

📄 License

This project is licensed under the MIT License.

📦 Additional Resources