This project is related to the Kaggle competition 100-bird-species. The objective of the competition is to build a CNN (Convolutional Neural Network) for accurately classifying 15 selected bird species. In this project, we leverage transfer learning, specifically using the MobileNetV2 model as the base model, to construct the CNN.
The CNN achieved an average accuracy of 95% on the test data, demonstrating the effectiveness of transfer learning in image classification. Additionally, the model achieved a perfect accuracy of 100% on a small validation set. This project serves as a showcase of how transfer learning can enhance the performance of a CNN in the task of bird species classification.
To use the code in this repository, please follow these steps:
-
Clone the repository to your local machine using the following command:
git clone https://github.com/CMonnin/kaggle_bird_classifier.git
-
Change directory
cd kaggle_bird_classifier -
Install the required dependencies using:
pip install -r requirements.txt
-
Download the data from 100-bird-species (approx. 2 GB). Extract the data and store it in a directory kaggle_bird_classifier/kaggle/archive/
Please follow the notebook found here to see how the model was trained, tested, and validated
The following people contributed to this project:
https://github.com/Franck816
https://github.com/sqossain
https://github.com/Anshuboom
https://github.com/CMonnin