Skip to content

Latest commit

 

History

History
31 lines (30 loc) · 1011 Bytes

README.md

File metadata and controls

31 lines (30 loc) · 1011 Bytes

classification-dashboard

A cool dashboard built to run inference on pretrained classifiers with streamlit.

Motivation

As a beginner to CNN architecture you might have difficulty in deciding the most appropriate and accurate architecture and this dashboard provides an ideal platform to do so.

Requirements

Python 3.8 or above with all requirements dependencies installed. To install run:

$ pip install -r requirements.txt

To Run

$ streamlit run dashboard.py

Available architectures

  • DenseNet121
  • DenseNet161
  • GoogLeNet
  • InceptionV3
  • ResNet34
  • ResNet50

Tutorial

Upload a picture, select an architecture and push "Classify"

Tutorial

Note

Weights for all of the architectures are downloaded and may take from 30 to 100 mb per model.

ToDo

  • Add more possible architectures.
  • Deploy on Herouku
  • Work on UI

By: Aryan Kargwal