Animal_10_Classification

Deep Learning Project
Explore the Repo »

View Streamlit app code · Model Building

About The Project

The major goal of this project is to create a system that will help animal researchers and photographers of wild animals deal with the numerous obstacles they face on a daily basis.

Method:

Convolutional neural networks is used to train images of animals from ten distinct species utilising thousands of labelled photos in the VGG16 transfer learning model.

Dataset:

Data came from Animals-10 dataset in kaggle.

CNN Model for Animal_10_classification:

This is the final model that yielded the highest accuracy:

Accuraccy and loss of above model:

• Training accuracy 97.08 % and loss of 11.22 %.
• Validation accuracy 93.99 % and loss of 24.05 %.

Drive model directory:

Link to Google Drive

Streamlit application for above app:

Link of streamlit app demo video