cjones/WhisperLive
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

update docs

Browse Source
This commit is contained in:
makaveli10
2023-07-17 23:30:47 +05:30
parent f4c6900afa
commit a84da4f79c
3 changed files with 27 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Audio-Transcription-Chrome/README.md
+10
View File
@@ -23,6 +23,13 @@ This Chrome extension allows you to send audio from your browser to a server for
### Capturing Audio
To capture the audio in the current tab, we used the chrome `tabCapture` API to obtain a `MediaStream` object of the current tab.
### Options
When using the Audio Transcription extension, you have the following options:
- **Use Collabora Server**: We provide a demo server which runs the whisper small model.
- **Use Multilingual Model**: Enable this option to utilize the multilingual capabilities of OpenAI-whisper.
- **Language**: Select the target language for transcription or translation. You can choose from a variety of languages supported by OpenAI-whisper.
- **Task:** Choose the specific task to perform on the audio. You can select either "transcribe" for transcription or "translate" to translate the audio to English.
### Getting Started
- Make sure the transcription server is running properly. To know more about how to start the server, see the [documentation here](https://github.com/collabora/whisper-live).
- Just click on the Chrome Extension which should show 2 options
@@ -33,3 +40,6 @@ To capture the audio in the current tab, we used the chrome `tabCapture` API to
## Limitations
This extension requires an internet connection to stream audio and receive transcriptions. The accuracy of the transcriptions may vary depending on the audio quality and the performance of the server-side transcription service. The extension may consume additional system resources while running, especially when streaming audio.
## Note
The extension relies on a properly running transcription server with multilingual support. Please follow the server documentation for setup and configuration.
Audio-Transcription-Firefox/README.md
+9
View File
@@ -21,6 +21,13 @@ This Firefox extension allows you to send audio from your browser to a server fo
### Capturing Audio
To capture the audio in the current tab, we used the chrome `tabCapture` API to obtain a `MediaStream` object of the current tab.
### Options
When using the Audio Transcription extension, you have the following options:
- **Use Collabora Server**: We provide a demo server which runs the whisper small model.
- **Use Multilingual Model**: Enable this option to utilize the multilingual capabilities of OpenAI-whisper.
- **Language**: Select the target language for transcription or translation. You can choose from a variety of languages supported by OpenAI-whisper.
- **Task:** Choose the specific task to perform on the audio. You can select either "transcribe" for transcription or "translate" to translate the audio to English.
### Getting Started
- Make sure the transcription server is running properly. To know more about how to start the server, see the [documentation here](https://github.com/collabora/whisper-live).
- Just click on the Firefox Extension which should show 2 options
@@ -31,3 +38,5 @@ To capture the audio in the current tab, we used the chrome `tabCapture` API to
## Limitations
This extension requires an internet connection to stream audio and receive transcriptions. The accuracy of the transcriptions may vary depending on the audio quality and the performance of the server-side transcription service. The extension may consume additional system resources while running, especially when streaming audio.
## Note
The extension relies on a properly running transcription server with multilingual support. Please follow the server documentation for setup and configuration.
README.md
+8 -4
View File
@@ -3,7 +3,7 @@ A nearly-live implementation of OpenAI's Whisper.
This project is a real-time transcription application that uses the OpenAI Whisper model to convert speech input into text output. It can be used to transcribe both live audio input from microphone and pre-recorded audio files.
Unlike traditional speech recognition systems that rely on continuous audio streaming, we use [voice activity detection (VAD)](https://github.com/snakers4/silero-vad) to detect the presence of speech and only send the audio data to whisper when speech is detected. This helps to reduce the amount of data sent to the API and improves the accuracy of the transcription output.
Unlike traditional speech recognition systems that rely on continuous audio streaming, we use [voice activity detection (VAD)](https://github.com/snakers4/silero-vad) to detect the presence of speech and only send the audio data to whisper when speech is detected. This helps to reduce the amount of data sent to the whisper model and improves the accuracy of the transcription output.
## Installation
- Install PyAudio and ffmpeg
@@ -30,13 +30,17 @@ Unlike traditional speech recognition systems that rely on continuous audio stre
- On the client side
- To transcribe an audio file:
```bash
python client.py --audio "audio.wav" --host "localhost" --port "9090"
python client.py --audio "audio.wav" --host "localhost" --port "9090" --multilingual --language "hi" --task "transcribe"
"translate"
```
This command transcribes the specified audio file (audio.wav) using the Whisper model. It connects to the server running on localhost at port 9090. It also enables the multilingual feature, allowing transcription in multiple languages. The --language flag specifies the target language for transcription, in this case, Hindi ("hi"). The --task flag is set to "transcribe" to indicate that transcription is the desired task. Also, --task can be set to "translate" to translate source language to English.
- To transcribe from microphone:
```bash
python client.py --host "localhost" --port "9090"
python client.py --host "localhost" --port "9090" --multilingual --language "en" --task "transcribe"
```
This command captures audio from the microphone and sends it to the server for transcription. It uses the same options as the previous command, enabling the multilingual feature and specifying the target language and task.
## Transcribe audio from browser
- Run the server
@@ -63,7 +67,7 @@ This would start the websocket server on port ```9090```.
```
## Future Work
- [ ] Update Documentation.
- [x] Update Documentation.
- [x] Keep only a single server implementation i.e. websockets and get rid of the socket implementation in ```server.py```. Also, update ```client.py``` to websockets-client implemenation.
- [ ] Add translation to other languages on top of transcription.