https://github.com/2003harsh/openai-whisper-automated-hindi-speech-recognition
This project adapts OpenAI's Whisper model to create an automated speech recognition system for Hindi. The goal is to accurately transcribe Hindi audio into text for applications like transcription, voice commands, and accessibility. By fine-tuning the model, the project aims to improve recognition accuracy and performance in Hindi-language context
https://github.com/2003harsh/openai-whisper-automated-hindi-speech-recognition
openai-whisper speech-recognition speech-to-text
Last synced: 7 months ago
JSON representation
This project adapts OpenAI's Whisper model to create an automated speech recognition system for Hindi. The goal is to accurately transcribe Hindi audio into text for applications like transcription, voice commands, and accessibility. By fine-tuning the model, the project aims to improve recognition accuracy and performance in Hindi-language context
- Host: GitHub
- URL: https://github.com/2003harsh/openai-whisper-automated-hindi-speech-recognition
- Owner: 2003HARSH
- Created: 2024-05-06T15:32:30.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-07-19T13:50:19.000Z (about 1 year ago)
- Last Synced: 2025-01-11T09:47:53.929Z (9 months ago)
- Topics: openai-whisper, speech-recognition, speech-to-text
- Language: Jupyter Notebook
- Homepage: https://www.kaggle.com/code/harshgupta2003/openai-whisper-automated-speech-recognition-hindi
- Size: 9.77 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# OpenAI Whisper Automated Speech Recognition (Hindi)
Welcome to the OpenAI Whisper Automated Speech Recognition (Hindi) project! This repository contains code and resources for building an automated speech recognition (ASR) system tailored for the Hindi language. The goal is to provide an efficient and accurate transcription tool for a variety of applications, from automatic transcriptions to voice commands and accessibility tools.
## Features
- **Custom ASR for Hindi**: Utilizes OpenAI's Whisper model to transcribe Hindi audio into text.
- **Word Error Rate (WER) Evaluation**: Provides a method to assess the accuracy of transcriptions.
- **Customizable**: Allows for fine-tuning and adaptation to various contexts within Hindi-language speech recognition.
## Getting Started
### Prerequisites
- Python 3.7+
- Kaggle account (for data and notebook execution)
- `OpenAI Whisper` and `torchaudio` libraries (for Whisper model)
### Code
For Detailed Code go here [https://www.kaggle.com/code/harshgupta2003/openai-whisper-automated-speech-recognition-hindi]
### Data
This project is hosted on Kaggle, where you can find datasets and sample notebooks to get started. Ensure you've downloaded the appropriate dataset for Hindi audio.
## Contributing
Contributions are welcome! If you'd like to contribute to the project, please fork the repository and submit a pull request. You can also report issues or suggest features via the GitHub issue tracker.
## Acknowledgments
Special thanks to OpenAI for the Whisper model and to the Kaggle community for providing the resources and datasets used in this project.