This repository provides a PyTorch implementation of the EfficientDet object detection model, enabling users to train custom models. It targets researchers and developers working on object detection tasks who need a flexible and performant solution. The project offers pre-trained weights and clear instructions for training, prediction, and evaluation on custom datasets.

How It Works

The project implements the EfficientDet architecture, known for its scalability and efficiency, in PyTorch. It leverages a compound scaling method to balance network depth, width, and resolution for optimal performance across different model sizes (d0-d7). The implementation supports various training configurations, including different learning rate schedules, optimizers, and multi-GPU training.

Quick Start & Requirements

• Install: pip install -r requirements.txt (requires PyTorch 1.2.0)
• Prerequisites: PyTorch 1.2.0, Python. CUDA is supported but not strictly required.
• Data: Requires VOC format datasets. Pre-trained weights and VOC datasets are available via Baidu Netdisk links provided in the README.
• Setup: Estimated setup time depends on dataset download and preparation.
• Docs: Training, prediction, and evaluation steps are detailed in the README.

Highlighted Details

• Achieves competitive mAP scores on COCO, with d7 reaching 51.2 mAP (0.5:0.95) at 1536x1536 resolution.
• Supports training from scratch or fine-tuning with custom datasets in VOC format.
• Includes functionalities for FPS testing and video prediction.
• Offers configurable parameters for confidence threshold, NMS IoU, and image resizing.

Maintenance & Community

The repository has seen significant updates, including added comments, adjustable parameters, and new features like FPS and video prediction (as of Oct 2021). Further updates in April 2022 added more learning rate strategies, optimizers, and image cropping. No specific community links (Discord/Slack) or notable contributors are mentioned.

Licensing & Compatibility

The repository's license is not explicitly stated in the README. However, it references other repositories which may have different licenses. Users should verify licensing for commercial use.

Limitations & Caveats

The project requires PyTorch version 1.2.0, which is quite old and may pose compatibility issues with newer libraries or hardware. Data download links are via Baidu Netdisk, which might be inconvenient for some users. The README does not detail specific hardware requirements beyond CUDA support.