shufflenet

ShuffleNetV1

Model	Download	Download (with sample test data)	ONNX version	Opset version
ShuffleNetv1	5.3 MB	7 MB	1.1	3
ShuffleNetv1	5.3 MB	9 MB	1.1.2	6
ShuffleNetv1	5.3 MB	9 MB	1.2	7
ShuffleNetv1	5.3 MB	9 MB	1.3	8
ShuffleNetv1	5.3 MB	9 MB	1.4	9

Description

ShuffleNet is a deep convolutional networks for classification.

Paper

ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices

Dataset

ILSVRC2012

Source

Caffe2 ShuffleNet ==> ONNX ShuffleNet

Model input and output

Input

gpu_0/data_0: float[1, 3, 224, 224]

Output

gpu_0/softmax_1: float[1, 1000]

Pre-processing steps

Post-processing steps

Sample test data

random generated sampe test data:

• test_data_0.npz
• test_data_1.npz
• test_data_2.npz
• test_data_set_0
• test_data_set_1
• test_data_set_2

Results/accuracy on test set

License

MIT