This is a repo for the Multiclass gun classification using gunshot audio data's It is implemented by using YAMNet transfer learning approach.
This GitHub repository hosts the project "Gunshot Detection using YAMNet," a deep learning-based solution designed to classify gunshot sounds in audio data(trained with 12 different gun sounds). YAMNet, a pre-trained neural network developed by Google, serves as the project's backbone. The repository contains code, data, and documentation that showcase the development and evaluation of the gunshot detection model.
YAMNet, with its extensive audio expertise, enables the extraction of powerful audio features. By leveraging transfer learning, the model is fine-tuned on a custom dataset comprising various gunshot sounds, ensuring its adaptability to real-world scenarios. The dataset encompasses a wide range of acoustic environments and gunshot variations, collectted from kaggle and other resources.