facenet is a TensorFlow implementation of FaceNet for face recognition and clustering using deep learning embeddings. It provides pretrained models and tools for training face recognition classifiers.
Project overview
facenet combines pretrained FaceNet models with classifier-training and LFW-validation scripts, while its documented legacy dependency stack makes environment compatibility a central adoption consideration.
Project type
Model Development · Model Runtime · Data Processing · Image & Vision
Deployment
Refer to project documentation
License
MIT
Best for
Developers and researchers who need a TensorFlow FaceNet implementation for face recognition, clustering, or custom classifier training.
Key capabilities
Provides code to train a face recognition classifier on user-supplied images.
Includes scripts to validate model accuracy on the Labeled Faces in the Wild (LFW) benchmark.
Limitations and risks
The project was tested only on legacy environments including Ubuntu 14.04, Python 2.7 or 3.5, and TensorFlow r1.7.
Its Python/TensorFlow MTCNN implementation does not produce identical results to the Matlab/Caffe implementation.
Dependency on Python 2.7 and TensorFlow r1.7 may create compatibility issues on modern systems.
Getting started
Clone the repository, install TensorFlow r1.7 under Python 3.5, download a pretrained model manually from Google Drive, then run the validate_on_lfw.py validation script.
Evidence and sources
README: This is a TensorFlow implementation of the face recognizer described in the paper ["FaceNet: A Unified Embedding for Face Recognition and Clustering"](http://arxiv.org/abs/1503.03…
GitHub project description: Face recognition using Tensorflow
README: Added code to [train a classifier on your own images](https://github.com/davidsandberg/facenet/wiki/Train-a-classifier-on-own-images). Renamed facenet_train.py to train_tripletlos…
README: The code is tested using Tensorflow r1.7 under Ubuntu 14.04 with Python 2.7 and Python 3.5.
README: | Model name | LFW accuracy | Training dataset | Architecture | |-----------------|--------------|------------------|-------------| | [20180408-102900](https://drive.google.com/op…