This repository contains python ml model for recognising musical instruments from short instrument audio clips.
| Approach | Core Idea | Toolchain |
|---|---|---|
| Python / ML | Learn features end-to-end from log-Mel spectrograms using a lightweight CNN trained with PyTorch, accelerated on GPU + AMP. | Python 3.9 – 3.12, PyTorch 2.x |
The project is intentionally structured so you can swap datasets and replicate experiments quickly in either branch, then compare accuracy, training time, and interpretability.
If you'd like the datasets used for training, feel free to shoot me an email at [email protected]