Add code

app.py

@@ -216,14 +216,13 @@ def generate_base(subject, setting, ):
 
     play_steps_in_s = 4.0
     play_steps = int(frame_rate * play_steps_in_s)
-    streamer = ParlerTTSStreamer(model, device=device, play_steps=play_steps)
 
     description = "Jenny speaks at an average pace with a calm delivery in a very confined sounding environment with clear audio quality."
     inputs = tokenizer(description, return_tensors="pt").to(device)
 
-    gr.Info("Reading story", duration=3)
 
-    for sentence in model_input:
+    for i, sentence in enumerate(model_input):
+        streamer = ParlerTTSStreamer(model, device=device, play_steps=play_steps)
 
         prompt = tokenizer(sentence, return_tensors="pt").to(device)
 
@@ -241,6 +240,8 @@ def generate_base(subject, setting, ):
         thread.start()
 
         for new_audio in streamer:
+            if i == 0:
+                gr.Info("Reading story", duration=3)
             print(f"Sample of length: {round(new_audio.shape[0] / sampling_rate, 2)} seconds")
             yield story, numpy_to_mp3(new_audio, sampling_rate=sampling_rate)