Skip to content

Latest commit

 

History

History
59 lines (39 loc) · 1.45 KB

hbonet-1.0.md

File metadata and controls

59 lines (39 loc) · 1.45 KB
Raw

hbonet-1.0

Use Case and High-Level Description

The hbonet-1.0 model is one of the classification models from repository with width_mult=1.0

Specification

Metric Value
Type Classification
GFLOPs 0.305
MParams 4.5447
Source framework PyTorch*

Accuracy

Metric Original model
Top 1 73.10%
Top 5 91.00%

Input

Original Model

Image, name: input, shape: [1x3x224x224], format: [BxCxHxW], where:

  • B - batch size
  • C - number of channels
  • H - image height
  • W - image width

Expected color order: RGB. Mean values: [123.675, 116.28, 103.53], scale factor for each channel: [58.395, 57.12, 57.375]

Converted Model

Image, name: input, shape: [1x3x224x224], format: [BxCxHxW], where:

  • B - batch size
  • C - number of channels
  • H - image height
  • W - image width

Expected color order: BGR.

Output

Object classifier according to ImageNet classes, shape: [1,1000] in [BxC] format, where:

  • B - batch size
  • C - vector of probabilities for all dataset classes.

Legal Information

The original model is distributed under the Apache License, Version 2.0. A copy of the license is provided in APACHE-2.0.txt.