Update app.py

app.py

@@ -1,15 +1,15 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
-import torch
+import soundfile as sf
 from transformers import pipeline
-from datasets import load_dataset
-
-# Set up your TTS model (as before)
-synthesiser = pipeline("text-to-speech", "Futuresony/output")
+import torch
 
-# Set up your text generation client
+# Initialize the client for the text generation model
 client = InferenceClient("Futuresony/future_ai_12_10_2024.gguf")
 
+# Initialize the TTS pipeline from Huggingface
+synthesizer = pipeline("text-to-speech", "Futuresony/output")
+
 def respond(
     message,
     history: list[tuple[str, str]],
@@ -18,18 +18,21 @@ def respond(
     temperature,
     top_p,
 ):
-    # Generate text response from your model
+    # Prepare the messages for the chatbot
     messages = [{"role": "system", "content": system_message}]
-
+    
+    # Add history of previous conversation
     for val in history:
         if val[0]:
             messages.append({"role": "user", "content": val[0]})
         if val[1]:
             messages.append({"role": "assistant", "content": val[1]})
-
+    
     messages.append({"role": "user", "content": message})
 
     response = ""
+
+    # Generate the response from the model
     for message in client.chat_completion(
         messages,
         max_tokens=max_tokens,
@@ -41,27 +44,26 @@ def respond(
         response += token
         yield response
 
-    # Convert the generated text into speech (Text-to-Speech)
-    # Get speaker embedding (optional, if you want to control the speaker)
-    embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
-    speaker_embedding = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze(0)
-
-    # Generate speech from the text response
-    speech = synthesiser(response, forward_params={"speaker_embeddings": speaker_embedding})
-
-    # Return the audio as a Gradio audio component
-    return response, speech["audio"]
+    # Convert the generated text to speech
+    speech = synthesizer(response)
+    
+    # Save the generated speech to a file
+    sf.write("generated_speech.wav", speech["audio"], samplerate=speech["sampling_rate"])
+    
+    # Return both the text and the audio for playback
+    return response, "generated_speech.wav"
 
 
 # Create the Gradio interface
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
+demo = gr.Interface(
+    fn=respond,
+    inputs=[
         gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
         gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
         gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
     ],
+    outputs=[gr.Textbox(), gr.Audio()],
 )
 
 if __name__ == "__main__":