fix app.py to remove redundant generate_kwargs in transcription calls and enable demo sharing during launch

app.py

@@ -53,7 +53,7 @@ def stream_transcribe(stream, new_chunk):
         else:
             stream = y
             
-        transcription = pipe({"sampling_rate": sr, "raw": stream}, generate_kwargs={"condition_on_previous_text": False})["text"]
+        transcription = pipe({"sampling_rate": sr, "raw": stream})["text"]
         end_time = time.time()
         latency = end_time - start_time
 
@@ -70,7 +70,7 @@ def transcribe(inputs, previous_transcription):
         sample_rate, audio_data = inputs
         scipy.io.wavfile.write(filename, sample_rate, audio_data)
 
-        transcription = pipe(filename, generate_kwargs={"condition_on_previous_text": False})["text"]
+        transcription = pipe(filename)["text"]
         previous_transcription += transcription
 
         end_time = time.time()
@@ -121,4 +121,4 @@ with gr.Blocks() as demo:
     gr.TabbedInterface([microphone, file], ["Microphone", "Transcribe from file"])
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch(share=True)