hamezksm/machine_noise_dataset

Machine Sound Doctor Dataset

Labeled audio clips of running machinery, used to train the classifier behind Machine Sound Doctor — a phone-based predictive maintenance tool: call in, hold your phone near a running machine, get an SMS back diagnosing the sound.

Classes

Class	Clips	Description
normal	33	Machine running normally
bearing_fault	33	Bearing fault sound signature
belt_slip	33	Belt slip sound signature

99 clips total. Most are 16-bit PCM mono WAV at 16 kHz; a small number are synthetic augmentations (filenames prefixed synth_).

Layout

normal/
bearing_fault/
belt_slip/

Each class is a folder of .wav clips.

Use

pip install huggingface_hub
hf download hamezksm/machine_noise_dataset --repo-type dataset --local-dir ./dataset

or with git:

git clone https://huggingface.co/datasets/hamezksm/machine_noise_dataset dataset

See the Machine Sound Doctor repo for the training pipeline (src/train.py) that consumes this data.