Datasets:

aranemini
/

kmr-tatoeba

Tasks:

Automatic Speech Recognition

Audio-Text-to-Text

Modalities:

Formats:

Languages:

Northern Kurdish

Size:

Libraries:

License:

Dataset card Data Studio Files Files and versions

aranemini commited on Aug 13, 2025

Commit

2dbe946

·

verified ·

1 Parent(s): 6ec8a27

Update README.md

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -49,7 +49,7 @@ It contains **only the test split** and is intended for **automatic speech recog
 - **Split:** Test only
 - **Speakers:** ~30 native speakers
 - **Content:**
-  - `file_name`: recorded sentence (`.wav`, 48 kHz mono)
   - `speaker`: anonymized speaker ID (SHA-256 hash)
   - `kurdish`: transcription in Northern Kurdish
   - `english`: English translation
@@ -76,7 +76,7 @@ ds = load_dataset("aranemini/kmr-tatoeba", split="test")
 # Example access
 print(ds[0]['kurdish'])
 print(ds[0]['english'])
-audio_data = ds[0]['file_name']  # dict with audio array & sampling rate
 ```
 ---

 - **Split:** Test only
 - **Speakers:** ~30 native speakers
 - **Content:**
+  - `audio`: recorded sentence (`.wav`, 48 kHz mono)
   - `speaker`: anonymized speaker ID (SHA-256 hash)
   - `kurdish`: transcription in Northern Kurdish
   - `english`: English translation
 # Example access
 print(ds[0]['kurdish'])
 print(ds[0]['english'])
+audio_data = ds[0]['audio']  # dict with audio array & sampling rate
 ```
 ---