Update LLMwithvoice.py

LLMwithvoice.py

@@ -19,7 +19,13 @@ torch_dtype = torch.float16 if device != "cpu" else torch.float32
 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")
 
-def query_roberta(api_token, payload):
+def query_roberta(api_token, prompt, context):
+    payload = {
+        "inputs": {
+            "question": prompt,
+            "context": context
+        }
+    }
     headers = {"Authorization": f"Bearer {api_token}"}
     response = requests.post(API_URL_ROBERTA, headers=headers, json=payload)
     try:
@@ -27,11 +33,10 @@ def query_roberta(api_token, payload):
     except ValueError:
         return {"error": "Invalid JSON response"}
 
-def generate_speech(prompt, description):
-    input_ids = tokenizer(description, return_tensors="pt").input_ids.to(device)
-    prompt_input_ids = tokenizer(prompt, return_tensors="pt").input_ids.to(device)
+def generate_speech(answer):
+    input_ids = tokenizer(answer, return_tensors="pt").input_ids.to(device)
 
-    generation = model.generate(input_ids=input_ids, prompt_input_ids=prompt_input_ids).to(torch.float32)
+    generation = model.generate(input_ids=input_ids).to(torch.float32)
     audio_arr = generation.cpu().numpy().squeeze()
 
     # Construct the audio path dynamically
@@ -46,18 +51,7 @@ def generate_speech(prompt, description):
 
     return audio_path
 
-def chat_with_roberta(api_token, question, context):
-    payload = {
-        "inputs": {
-            "question": question,
-            "context": context
-        }
-    }
-    response = query_roberta(api_token, payload)
-    if "error" in response:
-        return f"Error: {response['error']}"
-    else:
-        try:
-            return response['answer']
-        except (IndexError, KeyError):
-            return f"Unexpected response structure: {response}"
+def gradio_interface(api_token, prompt, context):
+    answer = query_roberta(api_token, prompt, context)
+    audio_path = generate_speech(answer)
+    return answer, audio_path