English | 简体中文

YOLOv6-Face

Implementation based on YOLOv6 v3.0 code.

New Feature

• Face-landmarks localization
• Repulsion loss
• Same-channel Dehead

Performance on WIDERFACE

Model	Size	Easy	Medium	Hard	SpeedT4 trt fp16 b1 (fps)	SpeedT4 trt fp16 b32 (fps)	Params (M)	FLOPs (G)
YOLOv6-N	640	95.0	92.4	80.4	797	1313	4.63	11.35
YOLOv6-S	640	96.2	94.7	85.1	339	484	12.41	32.45
YOLOv6-M	640	97.0	95.3	86.3	188	240	24.85	70.59
YOLOv6-L	640	97.2	95.9	87.5	102	121	56.77	159.24
YOLOv6Lite-S	416	89.6	84.6	58.8	/	/	0.53	0.90
YOLOv6Lite-M	416	90.6	86.1	60.6	/	/	0.76	1.07
YOLOv6Lite-L	416	91.8	87.6	64.2	/	/	1.06	1.40

Table Notes

• All checkpoints are fine-tuned from COCO pretrained model for 300 epochs without distillation.
• Results of the mAP and speed are evaluated on WIDER FACE dataset with the input resolution of 640×640.
• Speed is tested with TensorRT 8.2 on T4.
• Refer to Test speed tutorial to reproduce the speed results of YOLOv6.
• Params and FLOPs of YOLOv6 are estimated on deployed models.

Quick Start

Install

git clone https://github.com/meituan/YOLOv6
cd YOLOv6
git checkout yolov6-face
pip install -r requirements.txt

Training

Single GPU

python tools/train.py --batch 8 --conf configs/yolov6s_finetune.py --data data/WIDER_FACE.yaml --fuse_ab --device 0

Multi GPUs (DDP mode recommended)

python -m torch.distributed.launch --nproc_per_node 8 tools/train.py --batch 64 --conf configs/yolov6s_finetune.py --data data/WIDER_FACE.yaml --fuse_ab --device 0,1,2,3,4,5,6,7

• fuse_ab: Anchor Aided Training Mode
• conf: select config file to specify network/optimizer/hyperparameters. We recommend to apply yolov6n/s/m/l_finetune.py when training on WIDER FACE or your custom dataset.
• data: prepare dataset and specify dataset paths in data.yaml ( WIDERFACE, YOLO format widerface labels )
• make sure your dataset structure as follows:

├── widerface
│   ├── images
│   │   ├── train
│   │   └── val
│   ├── labels
│   │   ├── train
│   │   ├── val

Inference

First, download a pretrained model from the YOLOv6 release or use your trained model to do inference.

Second, run inference with tools/infer.py

python tools/infer.py --weights yolov6s_face.pt --source ../widerface/images/val/ --yaml data/WIDER_FACE.yaml --conf 0.02 --not-save-img --save-txt-widerface --name widerface_yolov6s

Evaluation

cd widerface_evaluate
python evaluation.py --pred ../runs/inference/widerface_yolov6s/labels/

Deployment

• ONNX
• OpenCV Python/C++
• OpenVINO
• TensorRT

Tutorials

• User Guide(zh_CN)
• Train COCO Dataset
• Train custom data
• Test speed
• Tutorial of Quantization for YOLOv6

Third-party resources

• YOLOv6 NCNN Android app demo: ncnn-android-yolov6 from FeiGeChuanShu

• YOLOv6 ONNXRuntime/MNN/TNN C++: YOLOv6-ORT, YOLOv6-MNN and YOLOv6-TNN from DefTruth

• YOLOv6 TensorRT Python: yolov6-tensorrt-python from Linaom1214

• YOLOv6 TensorRT Windows C++: yolort from Wei Zeng

• YOLOv6 web demo on Huggingface Spaces with Gradio.

• Interactive demo on DagsHub with Streamlit

• Tutorial: How to train YOLOv6 on a custom dataset

• YouTube Tutorial: How to train YOLOv6 on a custom dataset

• Demo of YOLOv6 inference on Google Colab

• Blog post: YOLOv6 Object Detection – Paper Explanation and Inference

FAQ（Continuously updated）

If you have any questions, welcome to join our WeChat group to discuss and exchange.