Update LLMwithvoice.py

LLMwithvoice.py

@@ -1,7 +1,7 @@
 import requests
 import torch
 import numpy as np
-from transformers import AutoTokenizer
+from transformers import AutoTokenizer, AutoModelForQuestionAnswering
 from parler_tts import ParlerTTSForConditionalGeneration
 from pydub import AudioSegment
 import simpleaudio as sa
@@ -9,18 +9,15 @@ import simpleaudio as sa
 # Hugging Face API URL for Roberta model
 API_URL_ROBERTA = "https://api-inference.huggingface.co/models/deepset/roberta-base-squad2"
 
-device = "cpu"
-if torch.cuda.is_available():
-    device = "cuda:0"
-if torch.backends.mps.is_available():
-    device = "mps"
-if torch.xpu.is_available():
-    device = "xpu"
-torch_dtype = torch.float16 if device != "cpu" else torch.float32
+# Determine the device to run the models
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+torch_dtype = torch.float16 if device.type != "cpu" else torch.float32
 
+# Load the ParlerTTS model and tokenizer
 model = ParlerTTSForConditionalGeneration.from_pretrained("parler-tts/parler_tts_mini_v0.1").to(device, dtype=torch_dtype)
 tokenizer = AutoTokenizer.from_pretrained("parler-tts/parler_tts_mini_v0.1")
 
+# Function to query the RoBERTa model
 def query_roberta(api_token, prompt, context):
     payload = {
         "inputs": {
@@ -42,10 +39,11 @@ def query_roberta(api_token, prompt, context):
         print(f"Exception: {e}")
         return {"error": "An unexpected error occurred"}
 
+# Function to generate speech from text
 def generate_speech(answer):
     input_ids = tokenizer(answer, return_tensors="pt").input_ids.to(device)
 
-    generation = model.generate(input_ids=input_ids).to(torch.float32)
+    generation = model.generate(input_ids=input_ids)
     audio_arr = generation.cpu().numpy().squeeze()
 
     # Convert numpy array to audio segment
@@ -68,9 +66,19 @@ def generate_speech(answer):
     except Exception as e:
         print(f"Error playing audio: {e}")
 
+# Function to interface with Gradio
 def gradio_interface(api_token, prompt, context):
     answer = query_roberta(api_token, prompt, context)
     if 'error' in answer:
         return answer['error'], None
     generate_speech(answer.get('answer', ''))
-    return answer.get('answer', 'No answer found'), None
+    return answer.get('answer', 'No answer found'), None
+
+# Example usage
+if __name__ == "__main__":
+    api_token = "your_huggingface_api_token"
+    prompt = "What is the capital of France?"
+    context = "France, in Western Europe, encompasses medieval cities, alpine villages, and Mediterranean beaches. Paris, its capital, is famed for its fashion houses, classical art museums including the Louvre, and monuments like the Eiffel Tower."
+
+    answer, _ = gradio_interface(api_token, prompt, context)
+    print("Answer:", answer)