ternaus/retinaface
A PyTorch implementation of RetinaFace for face detection with pre-trained models, distributed training, and mAP evaluation.
Not currently ranked — collecting fresh signals.
star history
This repository is a PyTorch implementation of RetinaFace, a face detection deep learning model. It uses PyTorch Lightning for the training loop, providing distributed training, fp16 support, and synchronized BatchNorm across GPUs. The project includes pre-trained models for inference, uses Albumentations for data augmentation, and computes mAP metrics during validation. Models can be installed via pip and used for face detection in images.
Frequently asked
- What is ternaus/retinaface?
- A PyTorch implementation of RetinaFace for face detection with pre-trained models, distributed training, and mAP evaluation.
- Is retinaface open source?
- Yes — ternaus/retinaface is open source, released under the MIT license.
- What language is retinaface written in?
- ternaus/retinaface is primarily written in Python.
- How popular is retinaface?
- ternaus/retinaface has 413 stars on GitHub.
- Where can I find retinaface?
- ternaus/retinaface is on GitHub at https://github.com/ternaus/retinaface.