Repository files navigation

A collection of face detection models pre-trained on theWiderfacedataset.

In the table below you can see each model detailed information including:

• meta architecture name
• model speed
• detector performance measured on theFDDB benchmark
• a download link to atar.gz file containing the model and configuration files
• a link for a live demo running on aGoogle Colaboratory notebook

Morghulis was used todownload and convert it to eitherDarknet orTensorflow Object Detection API format.

The remaining models were trained withTensorflow Object Detection APIonGoogle Cloud ML Engine.

There are 2 models trained withDarknet: one based on YOLOv2 and otheron Tiny YOLO. Both used convolutional weights that are pre-trained on Imagenet:darknet19_448.conv.23.