Upload folder using huggingface_hub

NonSpeech7k/info.md

@@ -0,0 +1,82 @@
+# Nonspeech7k Dataset
+
+## Overview
+
+A collection of manually annotated audio files for classification and analysis of human non-speech sounds, designed for academic research on human vocal expressions.
+
+## Source
+
+- **URL**: https://zenodo.org/records/6967442
+- **DOI**: 10.5281/zenodo.6967442
+- **Publication Date**: 2022
+- **Paper**: IET Signal Processing, 2023
+
+## Description
+
+The Nonspeech7k dataset contains 7,014 manually annotated audio files featuring human non-speech sounds. Files range from 500 milliseconds to 4 seconds in duration and are strongly manually annotated with single ground-truth labels.
+
+### Sound Classes
+
+- Breathing
+- Coughing
+- Crying
+- Laughing
+- Screaming
+- Sneezing
+- Yawning
+
+### Original Dataset Statistics
+
+- **Total files**: 7,014
+- **Training set**: 6,289 files
+- **Test set**: 725 files
+- **Original format**: 32kHz, mono WAV
+- **File duration**: 0.5-4 seconds
+- **Total size**: 2.5 GB
+
+### Data Sources
+
+- Freesound.org
+- YouTube
+- Aigei
+
+## Authors
+
+- Muhammad Mamunur Rashid
+- Guiqing Li
+- Chengrui Du
+
+**Affiliation**: South China University of Technology
+
+## License
+
+Creative Commons Attribution-NonCommercial-ShareAlike 4.0 (CC BY-NC-SA 4.0)
+
+**Usage**: Non-commercial and academic research purposes only
+
+## WebDataset Format
+
+| Property | Value |
+|----------|-------|
+| Total Samples | 7,014 |
+| Train | 6,289 samples |
+| Test | 725 samples |
+| Audio Format | FLAC, 48kHz, 16-bit, mono |
+| Metadata Format | JSON with "text" key |
+| Tar Files (train) | 7 |
+| Tar Files (test) | 1 |
+
+## Research Results
+
+- Best model (M12): 79% test accuracy, 0.81 precision, 0.80 F1-score
+- More recent research using CatBoost: 96.21% accuracy
+
+## Citation
+
+If you use this dataset, please cite:
+
+```
+Rashid, M. M., Li, G., & Du, C. (2023).
+Nonspeech7k dataset: Classification and analysis of human non-speech sound.
+IET Signal Processing.
+```

NonSpeech7k/test/0000.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1ab5ec51059e4179a46c853d323514ae0abd6a62eb5bd7ca997b02ee3b9a5b5
+size 167731200

NonSpeech7k/train/0000.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e1a82a7a7e113edd4e7f8d0f15c16afdcd6dd9a10084c871ebf8e5ea1b5ad89
+size 274053120

NonSpeech7k/train/0001.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5310b0a96a862d4c6c8235dfbc4464edaf6a1ed5cc1018f4ec6b59e688c18c7
+size 277022720

NonSpeech7k/train/0002.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2939c1c0bda2689b9e4bc9ec0ff8f60bc8c5d598a564d646db493239aad7fa7
+size 270632960

NonSpeech7k/train/0003.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a46fb16344c5be5936174faedba35c7406e52ad0daac008544df7f57e86c3d4c
+size 274411520

NonSpeech7k/train/0004.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06e7b81e3df90e4a03c82f5a103ec06d9542017fdad8ae2ca0925de400b74328
+size 276654080

NonSpeech7k/train/0005.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c8ed578e69a26d4d31eab015df011370810e691f9bf8af120ff42ce550b97369
+size 276203520

NonSpeech7k/train/0006.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e5d27c671c9a274932cf559a53ae9ecbac724efb278410c30755852f97400f0
+size 76523520