Hugging Face's logo

Datasets:
jacktol
/
atc-dataset

Modalities:
Audio
Text
Formats:
parquet
Size:
10K - 100K
Libraries:
Datasets
Dask
Dataset card Viewer Files Community
2
Dataset Viewer
Auto-converted to Parquet
Full Screen
audio
audioduration (s)
0.34
15.3
text
stringlengths
2
247
o kilo hotel uniform u
hotel charlie papa frequency change approved bis spater
direct padka request three five zero
praha good morning lufthansa two five three
sa six mike alfa
csa three kilo foxtrot change proceed direct to lodz
csa four two zero praha radar contact climb to flight level three four zero
hotel hotel xray vacacated at zero and to the right and contact apron one two one decimal seven five five bye bye
proceed direct baltu oscar kilo kilo uniform november
high speed approved csa seven bravo juliett
csa nine two six roger line up runway three one cleared for takeoff wind two zero zero degrees four knots
affirm cleared ils approach
one zero zero wizz air four xray zulu
zero to holesova requested level one eight
fa
the ils for runway three one
roger climb to flight level three three zero
descending flight level one one zero csa five
thank you
good morning lufthansa two kilo victor we are leaving two six descending level
aero flot one four one runway three one clear to land wind one seven zero degrees three kno
praha radar volga dniepr six seven two zero good evening maintaining level three three zero
vel two eight zero air berlin two eight two gol
o kemad thomson three seven nine thank you
that s affirmative we would like to climb to flight level three two zero mike alfa kilo
lufthansa seven two six direct and warsaw one three four and decimal nine two five
csa five bravo b rad radar contact climb to flight level one six zero climbing flight level one six zero csa
hotel bravo papa oscar echo zurich tower go ahead
six seven
cathay zero six eight after passing omelo proceed romis
condor five two two contact
tact praha radar one two zero two seven five wizz air four
own navigation to csa three zero three
one three three four one zero qatari three zero si
yeah sorry it is not possible because of noise restrictions
three three eight six zero luft transport one five five good bye
line up and wait runway three one
three three three nine zero ryan air five three lima xray bye bye
lufthansa eight echo papa one three three three nine zero have a nice day bye
one three one three five zero oscar yankee
klm one eight four
swiss one one eight yankee descend flight level one zero zero
continuing appro
grosjet seven whiskey radar radar contact
praha good afternoon lufthansa three charlie hotel passing level three six three descending level three four zero inbound marem
bye bye
easy two five eight quebec resume own navigation to the left proceed direct to cerno left to cerno easy two five eight quebec
four zero austrian three three
left heading one two zero csa nine seven
lufthansa one four nine one descend flight level three zero zero given rate
nine zero beline
three four two six joining right hand downwind runway two five grass
speed bird eight two zero track distance to touch down six five miles
jobair zero seven one line up runway one three
i will call you
for runway one
hello air france two four eight three ruzyne tower expect one more landing
descend level three four zero lufthansa three kilo victor
descending nine zero one one nine zero csa
austrian seven zero three praha hello radar contact follow vlm one romeo
lufthansa five papa juliett one three one zero five five bye bye
lufthansa seven two six proceed direct lima delta z on course to lufthansa seven two six thank you very much
qnh one zero one eight after keep the proceeding to victor
praha radar csa five one seven reaching flight level one five zero and
to reach rapet three zero zero qatari
swiss four seven three charlie stop climb at flight level one two zero only
german wings seven fi
csa seven five nine contact ground one two one nine
german wings seven papa roger
six zero maintain
oscar yankee november
lufthansa four hotel romeo radar contact
sky travel six eight two contact praha radar one two seven decimal eight two five
csa five papa charlie radar radar contact no speed descend flight level one zero zero
one two one nine good bye swiss four seven two ech
romeo ki
four zero climbing
baltu two bravo runway three one copied six six two three wizz air five nine one
csa eight five seven descend to flight level one one zero
lufthansa nine four november contact praha one two seven one two five good day
oman air one zero two praha good morning radar contact
three one two passing flight level two two seven climbing three one zero
takeoff
wizz air four uniform zulu contact praha one one eight decimal three seven five
one one eight
via hotel lima holding two four sky travel
wien one three four four four zero india foxtrot bye
praha austrian five five alfa delta maintain level three four zero on for the last ten mintues
air berlin one one five quebec contact praha one three three three nine zero good day
negative one three three three nine zero
ne three five one three five three five three
by a left turn to heading
radar good aftenoon eight delta inbound to tusin maintaining three four zero
austrian three one one november call praha one three three point three nine zero good bye
csa three delta echo contact ruzyne ground one two one decimal nin nine two one nine csa
please direct varik
prague yangzte river se
climb flight level one six zero csa six zero five zero
praha departure lufthansa nine echo lima climbing five thousand passing three thousand six
one zero three
End of preview. Expand in Dataset Viewer.

ATC Dataset - Fine-Tuning Whisper

This dataset was created to fine-tune OpenAI's Whisper model for improving transcription accuracy in Air Traffic Control (ATC) communications. The dataset contains transcriptions and corresponding audio files from two main sources: ATCO2 and the UWB-ATCC corpus, specifically selected for aviation-related communications. The dataset is publicly available on Hugging Face for use in Automatic Speech Recognition (ASR) projects.

For more details on the fine-tuning process, check out the blog post and the corresponding GitHub repository.

Dataset Overview

  • Dataset Name: ATC Dataset
  • Total Samples: 11.9k (Training), 2.93k (Test)
  • Data Sources:
  • Format: Audio files (WAV format) with corresponding transcriptions.
  • The Transciption Ground Truth is within the text column.
  • The Audio Ground Truth is within the audio column.
  • License: MIT

This dataset is particularly useful for training speech recognition models like Whisper on short, domain-specific audio transmissions, such as those between pilots and air traffic controllers.

Key Features

  • Domain-specific: Tailored to ATC communications with specialized phraseology and terms.
  • Diverse accents: Contains multiple accent variations to reflect real-world international aviation communication.
  • Cleaned Data: Includes only high-quality samples after filtering erroneous or incomplete transcriptions.

Usage

  1. Install Dependencies:
    Use Hugging Face's datasets library to load the dataset:

    from datasets import load_dataset
dataset = load_dataset("jacktol/atc-dataset")

  2. Training:
    The dataset is ready for speech recognition tasks such as fine-tuning Whisper models. It includes training and test splits to evaluate models based on Word Error Rate (WER).

License

This dataset is shared under the MIT License. You are free to use, modify, and distribute it as long as you provide proper attribution.

Downloads last month
55
Size of downloaded dataset files:
822 MB
Size of the auto-converted Parquet files:
822 MB
Number of rows:
14,795

Models trained or fine-tuned on jacktol/atc-dataset