johnmartinsson/differentiable-mel-spectrogram

The official implementation of DMEL the method presented in the paper "DMEL: The differentiable log-Mel spectrogram as a trainable layer in neural networks".

22
/ 100
Experimental

This project helps audio machine learning engineers and researchers by providing a trainable, differentiable log-Mel spectrogram layer for neural networks. It takes raw audio data as input and produces optimized Mel spectrograms, which can then be used for tasks like audio classification or sound event detection. The primary users are researchers and practitioners working on deep learning models for audio processing.

No commits in the last 6 months.

Use this if you are a deep learning engineer or researcher looking to incorporate a learnable Mel spectrogram transformation directly into your neural network architectures to potentially improve audio classification or other audio analysis tasks.

Not ideal if you are an end-user simply needing to generate standard Mel spectrograms or if you are not working with neural networks for audio processing.

audio-deep-learning speech-processing-research sound-event-detection audio-classification neural-network-design
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 6 / 25
Maturity 16 / 25
Community 0 / 25

How are scores calculated?

Stars

22

Forks

Language

Python

License

Apache-2.0

Last pushed

Dec 21, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/johnmartinsson/differentiable-mel-spectrogram"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.