search

SevaSk / ecoute

Ecoute is a live transcription tool that provides real-time transcripts for both the user's microphone input (You) and the user's speakers output (Speaker) in a textbox. It also generates a suggested response using OpenAI's GPT-3.5 for the user to say based on the live transcription of the conversation.
https://github.com/SevaSk/ecoute
MIT License
5.85k stars 817 forks source link

Added Dropdown and "tried" to implement translation #21

Closed Adarsha-gg closed 1 year ago

Adarsha-gg commented 1 year ago

I created the dropbox it can be edited to add new languages. I also added the preffered_language variable so that the value of the dropbox goes to the class so it can be accessed but I couldn't find a way to actually translate the transcript in real time. This is same as the last pull request I closed with the only change being the whisper's model.

Adarsha-gg commented 1 year ago

Thank you for your feedback! I'll try implementing the changes. Also can you tell me how to actually "translate" it to different languages I couldn't find it in the whisper's documentation.

On Wed, 24 May 2023 at 06:00, Seva @.***> wrote:

@.**** requested changes on this pull request.

Thanks for contributing! I cannot merge this PR yet I have left some comment with feedback.

— Reply to this email directly, view it on GitHub https://github.com/SevaSk/ecoute/pull/21#pullrequestreview-1440760664, or unsubscribe https://github.com/notifications/unsubscribe-auth/AU6WWIOTZFC2VEHATRU26KTXHVHKDANCNFSM6AAAAAAYMSGH4Y . You are receiving this because you authored the thread.Message ID: @.***>

SevaSk commented 1 year ago

Here is the whisper documentation https://github.com/openai/whisper/blob/main/README.md

we would want to use the language flag so the line would look something like self.audio_model.transcribe(file_path, fp16=torch.cuda.is_available(), language= "The ISO 2 Letter Language Codes")

Adarsha-gg commented 1 year ago

Hey man I may be wrong but I think as of now whisper can only translate from different languages to english and not from english to different languages. I think the purpose of this PR is to perform the latter one which seems impossible to me or I am just dumb enough to not make it work.

On Wed, 24 May 2023 at 06:13, Seva @.***> wrote:

Here is the whisper documentation https://github.com/openai/whisper/blob/main/README.md

we would want to use the language flag so the line would look something like self.audio_model.transcribe(file_path, fp16=torch.cuda.is_available(), language= "The ISO 2 Letter Language Codes")

— Reply to this email directly, view it on GitHub https://github.com/SevaSk/ecoute/pull/21#issuecomment-1560292034, or unsubscribe https://github.com/notifications/unsubscribe-auth/AU6WWIIVH55G42GMAG2WXR3XHVI3VANCNFSM6AAAAAAYMSGH4Y . You are receiving this because you authored the thread.Message ID: @.***>

SevaSk commented 1 year ago

Whisper can definitely transcribe other languages if you get the feature complete and fully working please feel free to open another pull request! I cant merge this one if the feature is not fully implemented.