Model Zoo

InsightFace Model Zoo

1. Face Detection models.

SSH model_v1 on BaiduDrive and Dropbox

Evaluation on WIDER FACE:

Impelmentation	Easy-Set	Medium-Set	Hard-Set	Speed
MTCNN	0.851	0.820	0.607	CPU realtime
Original Caffe SSH	0.93123	0.92106	0.84582	GPU realtime
Our SSH Model_v1	0.93394	0.92187	0.83682	GPU realtime

2. Face Alignment models.

2.1 2D Face Alignment

2.2 3D Face Alignment

2.3 Dense Face Alignment

3. Face Recognition models.

3.1 LResNet100E-IR,ArcFace@ms1m-refine-v2

BaiduDrive and Dropbox

Performance:

Method	LFW(%)	CFP-FP(%)	AgeDB-30(%)	MegaFace(%)
Ours	99.77	98.27	98.28	98.47

3.2 LResNet50E-IR,ArcFace@ms1m-refine-v1

BaiduDrive and Dropbox

Performance:

Method	LFW(%)	CFP-FP(%)	AgeDB-30(%)	MegaFace(%)
Ours	99.80	92.74	97.76	97.64

3.3 LResNet34E-IR,ArcFace@ms1m-refine-v1

BaiduDrive and Dropbox

Performance:

Method	LFW(%)	CFP-FP(%)	AgeDB-30(%)	MegaFace(%)
Ours	99.65	92.12	97.70	96.70

3.4 MobileFaceNet,ArcFace@ms1m-refine-v1

BaiduDrive and Dropbox

Performance:

Method	LFW(%)	CFP-FP(%)	AgeDB-30(%)	MegaFace(%)
Ours	99.50	88.94	95.91	-----

4. Face Attribute models.

4.1 Gender&Age

ResNet50x2 trained on Asian dataset. BaiduDrive Dropbox

4.2 Expression