---
language:
- da
tags:
- Sorting
- Classification
- Trash
size_categories:
- 10K<n<100K
task_categories:
- audio-classification
pretty_name: Mini Project - Trash Sorting
---

The dataset can be used to train a model for trash sorting, which consists of five categories of labels/trash audio samples: Metal, Glass, Cardboard, Plastic and Noise. 
The model created from this data set uses MFE and classifier - with this the accuracy of the model is 94,31%.
It can be found on Edge Impulse: https://studio.edgeimpulse.com/public/574071/live

It consist of both training and testing data, more detail can be seen in the edge impulse.