Create app.py

app.py

@@ -0,0 +1,22 @@
+import gradio as gr
+import whisper
+
+model_size = 'small'
+model = whisper.load_model(model_size)
+
+def speech_to_text(tmp_filename, _model_size):
+    if model_size != _model_size:
+        model_size = _model_size
+        model = whisper.load_model(model_size)
+
+    result = model.transcribe(tmp_filename)
+
+    return result["text"]
+
+gr.Interface(
+    fn=speech_to_text,
+    inputs=[
+        gr.Audio(source="microphone", type="filepath"),
+        gr.Dropdown(choices=["tiny", "base", "small", "medium", "large"]),
+        ],
+    outputs="text").launch()