neuro - no-code TinyML

Parameters

Data source

Synthetic Your CSV

Feature CSV - see format below

Download sample CSV

CSV: 47 feature columns and label. Up to 1 MB; read only in your browser. Results remain illustrative.

Windows per class - synthetic dataset size

400

MLP hidden layers - network size

layer 1 16

layer 2 8

Training epochs

150

Learning rate

0.020

Batch size

32

Seed

Configure the parameters and click Train to explore the complete workflow: features, training, quantization, export and C-test reports. This demo does not run a model.

Resultsы

Training curve

Confusion matrix (int8, test set)

Per class

class	precision	recall	n

Mean features per class - how sensors distinguish events

Audio (filter-bank bands)

Vibration (filter-bank bands)

C-tests - illustrative logs