Datasets:

African-Languages-Lab
/

fon-asr

---
license: cc-by-4.0
language:
- fon
pretty_name: Fon ASR Speech Dataset
task_categories:
- automatic-speech-recognition
size_categories:
- n<1K
tags:
- speech
- audio
- asr
- african-languages
- benin
configs:
- config_name: fon
  data_files:
  - split: train
    path: fon/train-*
dataset_info:
  config_name: fon
  features:
  - name: audio_id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_path
    dtype: string
  - name: transcription
    dtype: string
  - name: speaker_id
    dtype: string
  - name: speaker_gender
    dtype: string
  - name: speaker_age_range
    dtype: string
  - name: speaker_country
    dtype: string
  - name: duration
    dtype: float64
  - name: source_language
    dtype: string
  - name: dataset_tags
    dtype: string
  - name: prompt_type
    dtype: string
  - name: image
    dtype: string
  - name: image_path
    dtype: string
  - name: image_file
    dtype: image
  splits:
  - name: train
    num_bytes: 2029541505
    num_examples: 729
  download_size: 2029553234
  dataset_size: 2029541505
---

# Fon ASR Speech Dataset

A speech dataset in **Fon** (`fon`) for training and evaluating automatic speech recognition (ASR) models. It is part of an ASR project covering Fon, Gun and Baoulé. Each recording is a speaker describing an image prompt in their own words (semi-spontaneous speech), paired with its transcription and the image.

## Dataset summary

| | |
|---|---|
| Language | Fon (ISO 639-3 `fon`) |
| License | CC BY 4.0 |
| Task | Automatic speech recognition |
| Speech type | Semi-spontaneous (image description) |
| Utterances | 729 |
| Transcribed | 723 of 729 |
| Speakers | 7 |
| Total audio | 5.20 hours (5h 11m 43s) |
| Average utterance | 25.7 s |
| Recording country | Benin |

## Speakers

| | |
|---|---|
| Speakers | 7 |
| Gender | 5 male, 2 female |
| Age range | 3 25-34, 2 18-24, 2 35-44 |

Speakers are identified only by an opaque `speaker_id`. Use it to keep each speaker in a single split when making train/test splits, so evaluation is on unseen voices.

## Audio specification

| | |
|---|---|
| Format | WAV (PCM) |
| Sample rate | 48 kHz |
| Bit depth | 16-bit |
| Channels | Mono |
| Signal-to-noise ratio | ≥ 40 dB |
| Duration per file | 10–30 s (average 25.7 s) |

## Domain distribution

| Domain | Utterances | Hours | Share of total hours |
|---|---|---|---|
| Agriculture | 306 | 2.16 | 41.5% |
| Finance | 197 | 1.44 | 27.7% |
| Health | 8 | 0.06 | 1.1% |
| Government | 148 | 1.05 | 20.2% |
| Education | 70 | 0.49 | 9.4% |


## Data collection and quality

Recordings were collected through a recording app. Speakers were shown an image from one of the domains above and asked to describe it in their own words. Quality measures have been put in place at each step:

- Every recording was reviewed by validators; only recordings with status `approved` are included.
- Recordings shorter than 10 seconds are excluded.
- Recordings must have a signal-to-noise ratio of at least 40 dB.
- Audio must be 48 kHz, 16-bit, mono WAV.
- Every row was checked to be in Fon (`fon`).
- Metadata and media files were checked for consistency, so every row points to its own audio and image.

## Dataset structure

```text
audio/<audio_id>.wav
images/<image>.jpg
metadata.csv   (UTF-8 with BOM)
metadata.json
README.md
```

`metadata.csv` and `metadata.json` hold the same rows in the same order. All data is in a single `train` split.

### Data fields

| Field | Type | Description |
|---|---|---|
| `audio_id` | string | Unique ID of the recording |
| `audio` | string / audio | Audio file name; on the Hugging Face Hub, the playable audio |
| `audio_path` | string | Path of the WAV file inside the dataset folder |
| `transcription` | string | Transcription of the speech |
| `speaker_id` | string | Opaque speaker identifier (no names are included) |
| `speaker_gender` | string | Speaker gender |
| `speaker_age_range` | string | Speaker age range, e.g. `25-34` |
| `speaker_country` | string | Speaker country |
| `duration` | float | Length of the recording in seconds |
| `source_language` | string | ISO 639-3 language code |
| `dataset_tags` | string | Domain of the prompt (comma-separated if several) |
| `prompt_type` | string | Type of prompt; `semi_spontaneous` = describing an image in one's own words |
| `image` | string | ID of the image prompt |
| `image_path` | string | Path of the image prompt inside the dataset folder |

### Example

```json
{
  "audio_id": "5675",
  "audio": "5675.wav",
  "audio_path": "audio/5675.wav",
  "transcription": "Ɖiɖe ɔ ɖé xlɛ́ távo ɖokpó é jí wɛ e zé hlɔ̌ba amamú vovo ɛnɛ ɖé jí é. Bɔ hlɔ̌ba e nɛ ɔ lɛ́ɛ mɛ ɔ xumɛ hwevi hunmɛhunmɛ lɛ́ɛ ɖé mɛ. Ɖokpó ɖo hlɔ̌ba lɛ́ɛ mɛ bó nyí degɔn, bɔ degɔn ɔ jí ɔ é sɔ́ wema é mɛ e o wema ɔ ɖò ɖò haglan mɛ bɔ é ɖo degɔn ɔ jí.",
  "speaker_id": "b49e6f9d-781e-4bf7-9eaf-884291b5c710",
  "speaker_gender": "male",
  "speaker_age_range": "25-34",
  "speaker_country": "Benin",
  "duration": 27.0,
  "source_language": "fon",
  "dataset_tags": "Agriculture",
  "prompt_type": "semi_spontaneous",
  "image": "5675",
  "image_path": "images/5675.jpg"
}
```

## Usage

From the Hugging Face Hub:

```python
from datasets import load_dataset

ds = load_dataset("all-lab/fon-asr", split="train")
print(ds[0]["transcription"])
```

From the delivered folder:

```python
import pandas as pd

df = pd.read_csv("metadata.csv", encoding="utf-8-sig")
```

## License

This dataset is released under the [Creative Commons Attribution 4.0 International (CC BY 4.0)](https://creativecommons.org/licenses/by/4.0/) license.

	---
	license: cc-by-4.0
	language:
	- fon
	pretty_name: Fon ASR Speech Dataset
	task_categories:
	- automatic-speech-recognition
	size_categories:
	- n<1K
	tags:
	- speech
	- audio
	- asr
	- african-languages
	- benin
	configs:
	- config_name: fon
	data_files:
	- split: train
	path: fon/train-*
	dataset_info:
	config_name: fon
	features:
	- name: audio_id
	dtype: string
	- name: audio
	dtype: audio
	- name: audio_path
	dtype: string
	- name: transcription
	dtype: string
	- name: speaker_id
	dtype: string
	- name: speaker_gender
	dtype: string
	- name: speaker_age_range
	dtype: string
	- name: speaker_country
	dtype: string
	- name: duration
	dtype: float64
	- name: source_language
	dtype: string
	- name: dataset_tags
	dtype: string
	- name: prompt_type
	dtype: string
	- name: image
	dtype: string
	- name: image_path
	dtype: string
	- name: image_file
	dtype: image
	splits:
	- name: train
	num_bytes: 2029541505
	num_examples: 729
	download_size: 2029553234
	dataset_size: 2029541505
	---

	# Fon ASR Speech Dataset

	A speech dataset in Fon (`fon`) for training and evaluating automatic speech recognition (ASR) models. It is part of an ASR project covering Fon, Gun and Baoulé. Each recording is a speaker describing an image prompt in their own words (semi-spontaneous speech), paired with its transcription and the image.

	## Dataset summary

	\| \| \|
	\|---\|---\|
	\| Language \| Fon (ISO 639-3 `fon`) \|
	\| License \| CC BY 4.0 \|
	\| Task \| Automatic speech recognition \|
	\| Speech type \| Semi-spontaneous (image description) \|
	\| Utterances \| 729 \|
	\| Transcribed \| 723 of 729 \|
	\| Speakers \| 7 \|
	\| Total audio \| 5.20 hours (5h 11m 43s) \|
	\| Average utterance \| 25.7 s \|
	\| Recording country \| Benin \|

	## Speakers

	\| \| \|
	\|---\|---\|
	\| Speakers \| 7 \|
	\| Gender \| 5 male, 2 female \|
	\| Age range \| 3 25-34, 2 18-24, 2 35-44 \|

	Speakers are identified only by an opaque `speaker_id`. Use it to keep each speaker in a single split when making train/test splits, so evaluation is on unseen voices.

	## Audio specification

	\| \| \|
	\|---\|---\|
	\| Format \| WAV (PCM) \|
	\| Sample rate \| 48 kHz \|
	\| Bit depth \| 16-bit \|
	\| Channels \| Mono \|
	\| Signal-to-noise ratio \| ≥ 40 dB \|
	\| Duration per file \| 10–30 s (average 25.7 s) \|

	## Domain distribution

	\| Domain \| Utterances \| Hours \| Share of total hours \|
	\|---\|---\|---\|---\|
	\| Agriculture \| 306 \| 2.16 \| 41.5% \|
	\| Finance \| 197 \| 1.44 \| 27.7% \|
	\| Health \| 8 \| 0.06 \| 1.1% \|
	\| Government \| 148 \| 1.05 \| 20.2% \|
	\| Education \| 70 \| 0.49 \| 9.4% \|


	## Data collection and quality

	Recordings were collected through a recording app. Speakers were shown an image from one of the domains above and asked to describe it in their own words. Quality measures have been put in place at each step:

	- Every recording was reviewed by validators; only recordings with status `approved` are included.
	- Recordings shorter than 10 seconds are excluded.
	- Recordings must have a signal-to-noise ratio of at least 40 dB.
	- Audio must be 48 kHz, 16-bit, mono WAV.
	- Every row was checked to be in Fon (`fon`).
	- Metadata and media files were checked for consistency, so every row points to its own audio and image.

	## Dataset structure

	```text
	audio/<audio_id>.wav
	images/<image>.jpg
	metadata.csv (UTF-8 with BOM)
	metadata.json
	README.md
	```

	`metadata.csv` and `metadata.json` hold the same rows in the same order. All data is in a single `train` split.

	### Data fields

	\| Field \| Type \| Description \|
	\|---\|---\|---\|
	\| `audio_id` \| string \| Unique ID of the recording \|
	\| `audio` \| string / audio \| Audio file name; on the Hugging Face Hub, the playable audio \|
	\| `audio_path` \| string \| Path of the WAV file inside the dataset folder \|
	\| `transcription` \| string \| Transcription of the speech \|
	\| `speaker_id` \| string \| Opaque speaker identifier (no names are included) \|
	\| `speaker_gender` \| string \| Speaker gender \|
	\| `speaker_age_range` \| string \| Speaker age range, e.g. `25-34` \|
	\| `speaker_country` \| string \| Speaker country \|
	\| `duration` \| float \| Length of the recording in seconds \|
	\| `source_language` \| string \| ISO 639-3 language code \|
	\| `dataset_tags` \| string \| Domain of the prompt (comma-separated if several) \|
	\| `prompt_type` \| string \| Type of prompt; `semi_spontaneous` = describing an image in one's own words \|
	\| `image` \| string \| ID of the image prompt \|
	\| `image_path` \| string \| Path of the image prompt inside the dataset folder \|

	### Example

	```json
	{
	"audio_id": "5675",
	"audio": "5675.wav",
	"audio_path": "audio/5675.wav",
	"transcription": "Ɖiɖe ɔ ɖé xlɛ́ távo ɖokpó é jí wɛ e zé hlɔ̌ba amamú vovo ɛnɛ ɖé jí é. Bɔ hlɔ̌ba e nɛ ɔ lɛ́ɛ mɛ ɔ xumɛ hwevi hunmɛhunmɛ lɛ́ɛ ɖé mɛ. Ɖokpó ɖo hlɔ̌ba lɛ́ɛ mɛ bó nyí degɔn, bɔ degɔn ɔ jí ɔ é sɔ́ wema é mɛ e o wema ɔ ɖò ɖò haglan mɛ bɔ é ɖo degɔn ɔ jí.",
	"speaker_id": "b49e6f9d-781e-4bf7-9eaf-884291b5c710",
	"speaker_gender": "male",
	"speaker_age_range": "25-34",
	"speaker_country": "Benin",
	"duration": 27.0,
	"source_language": "fon",
	"dataset_tags": "Agriculture",
	"prompt_type": "semi_spontaneous",
	"image": "5675",
	"image_path": "images/5675.jpg"
	}
	```

	## Usage

	From the Hugging Face Hub:

	```python
	from datasets import load_dataset

	ds = load_dataset("all-lab/fon-asr", split="train")
	print(ds[0]["transcription"])
	```

	From the delivered folder:

	```python
	import pandas as pd

	df = pd.read_csv("metadata.csv", encoding="utf-8-sig")
	```

	## License

	This dataset is released under the [Creative Commons Attribution 4.0 International (CC BY 4.0)](https://creativecommons.org/licenses/by/4.0/) license.