Batched code

app.py

@@ -224,7 +224,7 @@ def generate_base(subject, setting):
     speech_output = model.generate(input_ids=description_tokens, prompt_input_ids=story_tokens)
     speech_output = [output.cpu().numpy() for output in speech_output]
     gr.Info("Generated Audio")
-    return None, None, {"audio": speech_output, "text": model_input}
+    return None, None, {"audio": speech_output, "text": model_input_tokens}
     
 def stream_audio(state):
     speech_output = state["audio"]
@@ -237,7 +237,7 @@ def stream_audio(state):
 #         print(f"i, j, time: {i}, {j} {datetime.datetime.now()}")
         print(f"Sample of length: {round(new_audio.shape[0] / sampling_rate, 2)} seconds")
         story += f"{sentence}\n"
-        yield story,  numpy_to_mp3(new_audio, sampling_rate=sampling_rate)
+        yield story, numpy_to_mp3(new_audio, sampling_rate=sampling_rate)
 
     # BATCH_SIZE = 4
     # for i in range(0, len(model_input), BATCH_SIZE):