Speech endpoint detection algorithm (new version)

[@Pauliomat]

Thanks for working on these projects

[@krosx278]

wow amazing, how do you set audio index?

[@Linguflex]

Use input_device_index property of AudioToTextRecorder class (hope that's what you mean)

[@krosx278]

@@Linguflex it works!! thank you...btw I want to make a question answering ai base on your RealtimeSTT transcribed. I am open for suggestion...

[@Linguflex]

@@krosx278 Please look at advanced_talk.py or openai_voice_interface.py in RealtimeSTT test folder or at my LocalAIVoiceChat project

[@krosx278]

@@Linguflex I want to use llama and fine-tune it... Is that possible?

[@Linguflex]

@@krosx278 Yes, of course

[@monsterfan-j2m]

is open api is needed?

[@Linguflex]

No, runs completely local

[@CodingPuff]

Interesting! What kind of models are you using

[@Linguflex]

faster_whisper for speech to text, Silero VAD + webrtcvad for voice activity detection, llama 3.1 8b for sentence end verification

[@parthpatwari3174]

@@Linguflex llama api?

[@Linguflex]

@@parthpatwari3174 LMStudio

[@Lucia-sy7le]

He's so creepy.