WhisperLive
WhisperLive copied to clipboard
collabora
A nearly-live implementation of OpenAI's Whisper.
When I run: from whisper_live.client import TranscriptionClient client = TranscriptionClient( "localhost", 9090, lang="hi", translate=True, model="small" ) client() I get the transcribed text printed to the screen. But how can I...
I was wondering why writing the chunks to disk is necessary for the client. I could not really find a reason in the code. If it's optional it might make...
Building on [Yasmin Moslem's tutorial on opennmt](https://forum.opennmt.net/t/nllb-200-with-ctranslate2/5090), I made a first pass at incorporating translation into whisperlive using ctranslate2, sentencepiece, and nllb-200. This is working, although it is a little...
Hello! First of all, amazing work on the library, I really appreciate what you're doing here! I'm trying to understand how (and if it's even possible) to generate SRT with...
I have a situation where the last few words of the transcript shouldn't change as much. For example if I would say 400 044 right now system could return 400...
What I mean is that the transcriptions up to e.g. 10 seconds of where the video is currently playing are computed. This is helpful for languages such as Japanese (and...
With option to record and transcribe system audio it would be possible to cover all sorts of useful use cases like: * Transcribe videos in any random video player or...
I'm using a cloud server to transcribe parts of the audio and thus encryption of the connection is vital for me. I simply implemented options for server and client to...
Hi there, I was wondering why you chose to use _ubuntu:focal_ as the base for the cpu Dockerfile. Why not use the default python3.11 images. These should be based on...
