README.md

---
dataset_info:
  features:
  - name: audio
    dtype: audio
  - name: speaker
    dtype: string
  - name: transcription
    dtype: string
  - name: translation
    dtype: string
  - name: Class
    dtype: string
  - name: Class_index
    dtype: float64
  splits:
  - name: train
    num_bytes: 515094059.0
    num_examples: 750
  - name: test
    num_bytes: 171698025.0
    num_examples: 250
  download_size: 590736573
  dataset_size: 686792084.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---

concatentation of "thomasavare/waste-classification-audio-deepl" and "thomasavare/waste-classification-audio-deepl2", so its easier to use.

there are 3 duplicates, idk why but it's too long to remake the datasets (maybe later) and not necessary.