Update README.md

README.md

@@ -10,4 +10,19 @@ pinned: false
 license: apache-2.0
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+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
+