README.md

---
title: Near Continuous Whispering
emoji: ⚡
colorFrom: yellow
colorTo: red
sdk: gradio
sdk_version: 3.4.1
app_file: app.py
pinned: false
license: apache-2.0
---

The near-continuous speech recognition demo using [OpenAI whisper](https://github.com/openai/whisper), built using [Gradio](https://gradio.app/).

### How to run?
Install openai/whisper

    pip install git+https://github.com/openai/whisper.git

Start the Gradio app

    python whisper_demo.py

### Simple Notes
1. The near-continuous recognition is implemented by incrementally recognizing all historical audio streaming every N seconds. The config is `REC_INTERVAL_IN_SECONDS`

2. The near-continuous recognition is in fact quite broken(slow) and only used for demo purpose. You should try a web socket way for real time recognition by referring to https://github.com/shirayu/whispering