---
dataset_info:
  features:
  - name: src_file
    dtype: string
  - name: fold
    dtype: int64
  - name: label
    dtype:
      class_label:
        names:
          '0': dog
          '1': rooster
          '2': pig
          '3': cow
          '4': frog
          '5': cat
          '6': hen
          '7': insects
          '8': sheep
          '9': crow
          '10': rain
          '11': sea_waves
          '12': crackling_fire
          '13': crickets
          '14': chirping_birds
          '15': water_drops
          '16': wind
          '17': pouring_water
          '18': toilet_flush
          '19': thunderstorm
          '20': crying_baby
          '21': sneezing
          '22': clapping
          '23': breathing
          '24': coughing
          '25': footsteps
          '26': laughing
          '27': brushing_teeth
          '28': snoring
          '29': drinking_sipping
          '30': door_wood_knock
          '31': mouse_click
          '32': keyboard_typing
          '33': door_wood_creaks
          '34': can_opening
          '35': washing_machine
          '36': vacuum_cleaner
          '37': clock_alarm
          '38': clock_tick
          '39': glass_breaking
          '40': helicopter
          '41': chainsaw
          '42': siren
          '43': car_horn
          '44': engine
          '45': train
          '46': church_bells
          '47': airplane
          '48': fireworks
          '49': hand_saw
  - name: esc10
    dtype: bool
  - name: take
    dtype: string
  - name: audio
    dtype: audio
  splits:
  - name: train
    num_bytes: 882179256
    num_examples: 2000
  download_size: 773038488
  dataset_size: 882179256
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: cc-by-nc-2.0
task_categories:
- audio-classification
size_categories:
- 1K<n<10K
---
# Dataset Card for "esc50"

This is a mirror for the ESC-50 dataset. Original sources:

https://github.com/karolpiczak/ESC-50
K. J. Piczak. ESC: Dataset for Environmental Sound Classification. Proceedings of the 23rd Annual ACM Conference on Multimedia, Brisbane, Australia, 2015.
[DOI: http://dx.doi.org/10.1145/2733373.2806390]

The dataset is available under the terms of the Creative Commons Attribution Non-Commercial license.

## Exploring the dataset

You can visualize the dataset using Renumics Spotlight:

```python
import datasets
from renumics import spotlight

ds = datasets.load_dataset('renumics/esc50', split='train')

spotlight.show(ds)
```

## Explore enriched dataset

To fully understand the dataset, you can leverage model results such as embeddings or predictions.

Here is an example how to use zero-shot classification with MS CLAP for this purpose:

```python
ds_results = datasets.load_dataset("renumics/esc50-clap2023-results",split='train')
ds = datasets.concatenate_datasets([ds, ds_results], axis=1)

spotlight.show(ds, dtype={'text_embedding': spotlight.Embedding, 'audio_embedding': spotlight.Embedding})
```


![image/png](https://cdn-uploads.huggingface.co/production/uploads/63dd29ffaf221a78fa4ec8d1/G13QHhxaVbSV7k5M4X5BN.png)