junbinwang / pretrained-models

Some Pretrained Models for TensorLayer

Feel free to add more.

CNN for ImageNet

The tl.models API description here, and the discussion for network architecture that can be easily use here.

Model	Code	Parameter	Top-1 Accuracy	Top-5 Accuracy
VGG 16	code1 code2	model	71.5	89.8
VGG 19	code	model (from machrisaa/tensorflow-vgg)	71.1	89.8
ResNet V1 50			75.2	92.2
ResNet V1 101		resnet_v1_101_2016_08_28.tar.gz	76.4	92.9
ResNet V1 152		resnet_v1_152_2016_08_28.tar.gz	76.8	93.2
ResNet V2 50		resnet_v2_50_2017_04_14.tar.gz	75.6	92.8
ResNet V2 101		resnet_v2_101_2017_04_14.tar.gz	77.0	93.7
ResNet V2 152		resnet_v2_152_2017_04_14.tar.gz	77.8	94.1
ResNet V2 200		TBA	79.9*	95.2*
Inception V1		inception_v1_2016_08_28.tar.gz	69.8	89.6
Inception V2		inception_v2_2016_08_28.tar.gz	73.9	91.8
Inception V3	code	inception_v3_2016_08_28.tar.gz	78.0	93.9
Inception V4			80.2	95.2
Xception
Inception-ResNet-v2			80.4	95.3
SqueezeNet V1	code1 code2	model
SqueezeNet V2
MobileNet V1	code1 code2	model
MobileNet V2_1.4_224			74.9	92.5
MobileNet V2_1.0_224			71.9	91.0
NASNet-A_Mobile_224		nasnet-a_mobile_04_10_2017.tar.gz	74.0	91.6
NASNet-A_Large_331		nasnet-a_large_04_10_2017.tar.gz	82.7	96.2
PNASNet-5_Large_331		pnasnet-5_large_2017_12_13.tar.gz	82.9	96.2
DenseNet
NASNet

Others

Model	Code	Parameter	Description
SRGAN	code	model	Super Resolution for Image
YOLO3 for VOC
YOLO3 for COCO
OpenPose for MPII
OpenPose for COCO

More Examples in github and docs.

References

• TF-Slim
• Keras