Dataset Viewer
Auto-converted to Parquet Duplicate
sample_id
stringclasses
3 values
speaker_id
stringclasses
3 values
dialect
stringclasses
3 values
source_language
stringclasses
2 values
target_language
stringclasses
3 values
duration_seconds
float64
3.8
5.1
transcript
stringclasses
3 values
translation
stringclasses
3 values
quality_score
float64
0.89
0.94
audio_path
stringclasses
3 values
v2v-001
spk-101
southern
ps
en
3.8
زه بازار ته ځم.
I am going to the market.
0.94
data/audio_clean/spk-101_001.wav
v2v-002
spk-102
central
ps
fa
4.2
ته څنګه يې؟
تو چطور هستی؟
0.91
data/audio_clean/spk-102_002.wav
v2v-003
spk-103
northern
fa
ps
5.1
امروز هوا خوب است.
نن هوا ښه ده.
0.89
data/audio_clean/spk-103_003.wav

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

ZamAI-Pashto Voice2Voice

This repository contains the dataset and scripts for the ZamAI-Pashto voice-to-voice project, focusing exclusively on the Pashto language.

Project Structure

  • data/: Contains raw and cleaned Pashto audio, transcripts, and metadata.
  • annotations/: Phoneme labels, dialect tags, and timing alignments.
  • scripts/: Preprocessing, alignment, and feature generation scripts.
  • configs/: Dataset configurations and dialect mappings.

License

This project is licensed under the Apache License 2.0.

Downloads last month
33

Space using ZamAI-Pashto/zamai-pashto-voice2voice 1