Skip to content

Latest commit

 

History

History
69 lines (42 loc) · 2.23 KB

octave-se-resnet-50-0.125.md

File metadata and controls

69 lines (42 loc) · 2.23 KB

octave-se-resnet-50-0.125

Use Case and High-Level Description

The octave-se-resnet-50-0.125 model is a modification of se-resnet-50 from this paper with octave convolutions from Drop an Octave: Reducing Spatial Redundancy in Convolutional Neural Networks with Octave Convolution with alpha=0.125. As origin, it's designed to perform image classification. For details about family of octave convolution models, check out the repository.

The model input is a blob that consists of a single image of 1x3x224x224 in RGB order. The RGB mean values need to be subtracted as follows: [124,117,104] before passing the image blob into the network. In addition, values must be divided by 0.0167.

The model output for octave-se-resnet-50-0.125 is the typical object classifier output for the 1000 different classifications matching those in the ImageNet database.

Example

Specification

Metric Value
Type Classification
GFLOPs 7.246
MParams 28.082
Source framework MXNet*

Accuracy

Performance

Input

Original model

Image, name - data, shape - 1,3,224,224, format is B,C,H,W where:

  • B - batch size
  • C - channel
  • H - height
  • W - width

Channel order is RGB. Mean values - [124,117,104], scale value - 59.880239521

Converted model

Image, name - data, shape - 1,3,224,224, format is B,C,H,W where:

  • B - batch size
  • C - channel
  • H - height
  • W - width

Channel order is BGR

Output

Original model

Object classifier according to ImageNet classes, name - prob, shape - 1,1000, output data format is B,C where:

  • B - batch size
  • C - Predicted probabilities for each class in [0, 1] range

Converted model

Object classifier according to ImageNet classes, name - prob, shape - 1,1000, output data format is B,C where:

  • B - batch size
  • C - Predicted probabilities for each class in [0, 1] range

Legal Information

LICENSE