Skip to content

An example of performing inference using the MobileNetV3 model with the Tract library

License

Notifications You must be signed in to change notification settings

banitalebi/MobileNetV3-Inference

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MobileNetV3 Inference using Tract Library

This repository provides a simple example of performing inference using the MobileNetV3 model with the Tract library. The model is evaluated using the mobilenetv3_model.onnx file on a test dataset, with performance metrics calculated, including true positives, false positives, true negatives, false negatives, and overall accuracy. The results are compared against a Python implementation.

Dataset

The dataset used for fine-tuning the model is sourced from Kaggle. This dataset can be organized into three subsets:

  • Training Set: 70% of the total dataset
  • Validation Set: 20% of the total dataset
  • Test Set: 10% of the total dataset

Note that the test dataset is used exclusively for evaluation. and it can be found here.

Results

The results are listed as follows:

  • True Positives (Violence): 491
  • True Negatives (Non-Violence): 404
  • False Positives (Predicted Violence, Actual Non-Violence): 120
  • False Negatives (Predicted Non-Violence, Actual Violence): 94
  • Total Images: 1109
  • Accuracy: 80.70%

About

An example of performing inference using the MobileNetV3 model with the Tract library

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Languages