limit on file size for audio transcription

edgenai / edgen

⚡ Edgen: Local, private GenAI server alternative to OpenAI. No GPU required. Run AI models locally: LLMs (Llama2, Mistral, Mixtral...), Speech-to-text (whisper) and many others.

https://docs.edgen.co/

Apache License 2.0

323 stars 14 forks source link

limit on file size for audio transcription #97

Closed prabirshrestha closed 5 months ago

prabirshrestha commented 5 months ago

Is there a set limit on the audio file when creating transcription?

I noticed that small files work but large files doesn't work. You can try a 30mb file from http://www.kiea.jp/english_voicefiles.html. Direct link to mp3: http://www.kiea.jp/hosenji.mp3

Looking at the code seems like file is a vec. https://github.com/edgenai/edgen/blob/66ee740d059ee0e8d2b52bab1e153bf2b5f1b7ba/crates/edgen_core/src/whisper.rs#L39-L40

pedro-devv commented 5 months ago

Apparently the default request size for axum is only 2MB, I added a new parameter in the settings to change this.