esc-50

[ICLR 2024] SemiReward: A General Reward Model for Semi-supervised Learning

Building a Sound Classification iOS Application using AI

This repository contains an end-to-end pipeline to train a convolutional neural network (CNN) for environmental sound classification on ESC-50. Serve the model for inference using Modal + a FastAPI endpoint. - Visualize model outputs (top predictions, input mel-spectrogram, waveform, and CNN feature maps).

Classification of Environment Sound using CNN and ImageDataGenerator

audio classification by pytorch.

Classification of Environmental Sound using Deep Learning.

Audio Diffusion

A deep learning classifier for urban sounds using the EfficientNet network

This project builds a system to split spoken sentences into words (speaker-independent) and calculates the speaker's average pitch, comparing different methods to identify word boundaries.

Residual CNN for Environmental Sound Classification — 84.5% accuracy on ESC-50, outperforming SVM baseline by 22.5%

REST API based on PyTorch (ResNet18) for classifying 50 categories of natural and household sounds (rain, chainsaw, glass breaking, etc.) from audio files. Mel spectrograms + FastAPI. Val accuracy 86%. Trained in Google Colab on ESC-50.

Reconnaissance de sons du quotidien (ESC-50) avec des descripteurs audio + XGBoost, sans deep learning

Exploration of various machine learning techniques to classify audio data using the ESC-50 dataset.

End to End ML Classification with Streamlit Frontend

Environmental sound classification using MFCC features, Random Forest, and a Streamlit web app.

ESC-50 environmental sound classification: supervised CNN (~62%) vs. COLA-style contrastive self-supervision — with a measured writeup of why SSL needs scale.

The aim of this project was to design and implement a Flask web application for classifying environmental sounds which uses convolutional neural network architecture.