Whisper Medium.en Fine Tuned For ATC

🚀 Whisper Medium EN Fine-Tuned for Air Traffic Control (ATC)

This model is a fine - tuned version of OpenAI's Whisper Medium EN model, specifically tailored for Air Traffic Control (ATC) communication. It offers high - precision transcription for aviation - related communications, significantly enhancing the accuracy compared to the original model.

🚀 Quick Start

You can access the fine - tuned model on Hugging Face:

• Whisper Medium EN Fine - Tuned for ATC
• Whisper Medium EN Fine - Tuned for ATC (Faster Whisper)

You can also test the model online using the ATC Transcription Assistant, which allows you to upload audio files and generate transcriptions.

✨ Features

• Enhanced Transcription Accuracy: The fine - tuning process reduces the Word Error Rate (WER) by 84% on domain - specific aviation communications compared to the original pretrained model.
• Handling Variations: It is particularly effective at dealing with accent variations and ambiguous phrasing commonly found in ATC communications.
• Optimized for ATC: Specifically designed to handle short, distinct transmissions between pilots and air traffic controllers.

📚 Documentation

Model Overview

This model is a fine - tuned version of OpenAI's Whisper Medium EN model, trained on Air Traffic Control (ATC) communication datasets. The fine - tuning brings remarkable improvements in transcription accuracy for aviation - specific communications.

Model Description

Whisper Medium EN fine - tuned for ATC is optimized for short transmissions between pilots and controllers. It is fine - tuned using data from:

• ATCO2 corpus (1 - hour test subset)
• UWB - ATCC corpus

The ATC Dataset combines and cleans data from these two sources, filtering and refining to enhance transcription accuracy for ATC communications.

Intended Use

• Transcribing aviation communication: Provide accurate transcriptions for ATC communications, accounting for accents and English phrasing variations.
• Air Traffic Control Systems: Assist in real - time transcription of pilot - ATC conversations to improve situational awareness.
• Research and training: Useful for researchers, developers, or aviation professionals studying ATC communication or developing new aviation - safety tools.

Training Procedure

• Hardware: Fine - tuning was conducted on two A100 GPUs with 80GB memory.
• Epochs: 10
• Learning Rate: 1e - 5
• Batch Size: 32 (effective batch size with gradient accumulation)
• Augmentation: Dynamic data augmentation techniques (Gaussian noise, pitch shifting, etc.) were applied during training.
• Evaluation Metric: Word Error Rate (WER)

Limitations

While the fine - tuned model performs well in ATC - specific communications, it may not generalize effectively to other speech domains. Also, like most speech - to - text models, transcription accuracy can be affected by extremely poor - quality audio or heavily accented speech not encountered during training.

🔧 Technical Details

The model - index information is as follows:

• Name: Whisper Medium EN Fine - Tuned for ATC
• Results:
  • Task:
    • Type: automatic - speech - recognition
  • Dataset:
    • Name: ATC Dataset
    • Type: jacktol/atc - dataset
  • Metrics:
    • Name: Word Error Rate (WER)
    • Type: wer
    • Value: 15.08
  • Source:
    • Name: ATC Transcription Evaluation
    • URL: https://jacktol.net/posts/fine - tuning_whisper_for_atc/

📄 License

This model is released under the MIT license.

References

• Blog Post: [Fine - Tuning Whisper for ATC: 84% Improvement in Transcription Accuracy](https://jacktol.net/posts/fine - tuning_whisper_for_atc/)
• GitHub Repository: [Fine - Tuning Whisper on ATC Data](https://github.com/jack - tol/fine - tuning - whisper - on - atc - data/tree/main)