Streamlit Openai Whisper 🗣️

Streamlit demo project for Openai Whisper.

Status ✔️

Streamlit App works - Has some issues - Last changed: 2023-03-10

Issues 🚩

• Dependencies of whisper are very heavy (pytorch, transformers, cuda...) and take a long time to install.
• Model is not pre-downloaded and very huge, so it takes a long time to download the first time.
• Larger models above the small model seems to crash on streamlit cloud, probably due to memory limits.
• The largest model that seems to run locally and on streamlit cloud is the small model.
• Usage of Git-LFS for pre-downloaded model also does not make sense, because at least in the free version of GitHub, you only get 1GB of storage space and 1GB of bandwidth per month.
• The transcribe process takes a long time, probably due to the lack of GPU support.
• The currently used audiorecorder component is not working. Have to check out other alternatives.

ToDo ☑️

• Test app locally with docker
• Test app locally within virtual environment
• Use temporary directory for uploaded file
• Try deployment on streamlit cloud
• Try larger models
• Add custom audio recording component to streamlit app
• Fix and test audiorecorder issue
• Try other audio recording components
• Improve visual layout of the frontend
• Restructure file structure

Resources 📚

• Streamlit ❤️
  • https://docs.streamlit.io/
• Openai Whisper ❤️
  • https://github.com/openai/whisper
• Streamlit Custom Components for Audio Recording
  • streamlit-audiorecorder 🤷‍♂️
    • https://github.com/theevann/streamlit-audiorecorder
  • audio-recorder-streamlit 🤷‍♂️
    • https://github.com/Joooohan/audio-recorder-streamlit
  • streamlit-mic-recorder 🤷‍♂️
    • https://github.com/B4PT0R/streamlit-mic-recorder
  • streamlit-webrtc 🤷‍♂️
    • https://github.com/whitphx/streamlit-webrtc
    • https://github.com/whitphx/streamlit-stt-app
  • streamlit_audio_recorder 🤔
    • https://github.com/stefanrmmr/streamlit_audio_recorder