metadata

language: fa
license: cc-by-4.0
tags:
  - audio
  - speech-recognition
  - fleurs
  - persian
  - farsi
pretty_name: FLEURS Farsi (fa_ir) Processed
dataset_info:
  features:
    - name: audio
      dtype: audio
      sampling_rate: 16000
    - name: transcription
      dtype: string
  splits:
    dev:
      name: dev
      num_bytes: 352117826
      num_examples: 369
    test:
      name: test
      num_bytes: 852844223
      num_examples: 871
    train:
      name: train
      num_bytes: 2811193790
      num_examples: 3101
  download_size: 3934177703
  dataset_size: 4016155839
configs:
  - config_name: default
    data_files:
      - split: dev
        path: data/dev-*
      - split: test
        path: data/test-*
      - split: train
        path: data/train-*
task_categories:
  - automatic-speech-recognition

FLEURS Farsi (fa_ir) - Processed Dataset

Dataset Description

This dataset contains the Farsi (Persian, fa_ir) portion of the FLEURS (Few-shot Learning Evaluation of Universal Representations of Speech) dataset, processed into a Hugging Face datasets compatible format. FLEURS is a many-language speech dataset created by Google, designed for evaluating speech recognition systems, particularly in low-resource scenarios.

This version includes audio recordings and their corresponding transcriptions, split into train, development (dev), and test sets. The audio is sampled at 16kHz.

Languages: Farsi (Persian) - fa

How to Use

You can load this dataset using the Hugging Face datasets library:

from datasets import load_dataset

dataset_name = "msghol/fleurs-farsi" # Updated repository name
fleurs_farsi = load_dataset(dataset_name)

# Accessing a split
print(fleurs_farsi["train"])

# Accessing an example
example = fleurs_farsi["train"][0]
audio_sample = example["audio"]
transcription = example["transcription"]

print(f"Audio: {audio_sample}")
print(f"Transcription: {transcription}")