name
stringlengths 16
21
| transcript
stringlengths 3
590
| audio
dict |
|---|---|---|
1159431_chunk000.wav
|
menimcha notebook bilan uy kompyuterining anaqa
| {"array":[-1.8683495000004768e-6,-5.044508725404739e-6,-4.705623723566532e-6,-9.16019780561328e-6,-8(...TRUNCATED)
|
1159431_chunk001.wav
|
put menyusi oʻzgacha boʻladi
| {"array":[-0.004753559827804565,-0.0063455854542553425,-0.0007537417113780975,0.0034691840410232544,(...TRUNCATED)
|
1159994_chunk000.wav
| "xo'p qarang aka tredingda tushunchangiz bo'lsa demo akkaunt ishlatganmisiz demo akkaunt ishlatmagan(...TRUNCATED)
| {"array":[3.6187702789902687e-6,-3.326509613543749e-6,-0.00002699129981920123,-0.0000580663327127695(...TRUNCATED)
|
1159994_chunk001.wav
| "nimangiz bo'lsa chik ko'tarsangiz ham aqlingiz bo'lsa chik ko'tarsangiz ham keyin xm dan bonusga ha(...TRUNCATED)
| {"array":[0.018005073070526123,0.03174597769975662,0.025372151285409927,0.030240945518016815,0.02672(...TRUNCATED)
|
1159994_chunk002.wav
|
agar aniq aqlingiz yetsa 5 dollarlik pul solib turib
| {"array":[0.0011414553737267852,0.002868989482522011,0.002911577932536602,0.004182690754532814,0.005(...TRUNCATED)
|
1159994_chunk004.wav
|
50 60 ga ko'tarib 30 dollarni qaytarib turib 25 20 larini yechib olsangiz
| {"array":[-0.00032577477395534515,-0.00044592004269361496,0.000027468428015708923,0.0006958702579140(...TRUNCATED)
|
1159994_chunk005.wav
|
keyin exness akkaunt oching exness akkauntdan
| {"array":[0.011245672591030598,0.017075415700674057,0.01262490451335907,0.015633599832654,0.01557401(...TRUNCATED)
|
1159994_chunk006.wav
| "standartini ochsangiz spread 02 chiqadi bu ishlashingiz osonroq bo'ladi foydangiz tezroq bo'ladi de(...TRUNCATED)
| {"array":[0.003018748015165329,0.0019762825686484575,-0.0010100125800818205,-0.0013246247544884682,-(...TRUNCATED)
|
1159994_chunk007.wav
|
shunaqa hozircha
| {"array":[-0.0029898802749812603,-0.005870899185538292,-0.005631991196423769,-0.004958765115588903,-(...TRUNCATED)
|
1159995_chunk000.wav
|
unda ham 10 dollarda ham foydangiz normalni bilimingizga bilim qo'shilishi mumkinda
| {"array":[-8.088652975857258e-6,-0.000024083739845082164,-0.00002844582195393741,-0.0000364665174856(...TRUNCATED)
|
End of preview. Expand
in Data Studio
Speech-to-Text Evaluation Dataset
Dataset Overview
This dataset is designed for evaluating Uzbek speech-to-text (STT) models on real-world conversational speech data. The audio samples were collected from various open Telegram groups, capturing natural voice messages in diverse acoustic conditions and speaking styles.
Key Statistics
- Total Samples: 745 audio files
- Total Duration: 1 hour 40 minutes (~100 minutes)
- Average Duration: ~8 seconds per sample
- Source: Voice messages from various open Telegram groups
- Transcriptions: Manually annotated
Dataset Structure
The dataset is saved as a datasets.Dataset object in Arrow format, containing the following fields:
name: Name of audio fileaudio: Audio file data (dict witharray, andsampling_rate)transcription: Ground truth text transcription (manually annotated)
Loading the Dataset
Installation
To use this dataset, you need to install the Hugging Face datasets library:
pip install datasets
Basic Loading
from datasets import load_dataset
# Load the dataset from the Arrow files
dataset = load_dataset("OvozifyLabs/asr_evaluate_set")
# View dataset information
print(dataset)
print(f"Number of samples: {len(dataset)}")
Data Characteristics
Audio Properties
- Source Domain: Conversational voice messages from Telegram
- Variability: Multiple speakers, diverse acoustic environments
- Recording Conditions: Real-world
- Language: Uzbek
Transcription Details
- Annotation Method: Manual transcription
- Quality: Human-verified ground truth labels
- Convention: punctuation removed, lowercased
Use Cases
This dataset is suitable for:
- Evaluating speech-to-text model performance on conversational speech
- Benchmarking ASR systems on real-world voice messages
- Testing model robustness to varied acoustic conditions
- Comparing different STT models
- Downloads last month
- 1