Find file History
Permalink
Type Name Latest commit message Commit time
..
Failed to load latest commit information.
README.md bump models to opset 9 (#98) Sep 20, 2018

README.md

ShuffleNet

Download:

Model size: 5.3 MB

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