Update app.py

app.py

@@ -25,7 +25,7 @@ class MessageRequest(BaseModel):
     message: str
     language: str
 
-repo_id = "microsoft/phi-4"
+repo_id = "HuggingFaceH4/zephyr-7b-alpha"
 llm_client = InferenceClient(
     model=repo_id,
     token=os.getenv("HF_TOKEN"),
@@ -61,8 +61,8 @@ templates = Jinja2Templates(directory="static")
 
 # Configure Llama index settings
 Settings.llm = HuggingFaceInferenceAPI(
-    model_name="microsoft/phi-4",
-    tokenizer_name="microsoft/phi-4",
+    model_name="HuggingFaceH4/zephyr-7b-alpha",
+    tokenizer_name="HuggingFaceH4/zephyr-7b-alpha",
     context_window=3000,
     token=os.getenv("HF_TOKEN"),
     max_new_tokens=512,
@@ -185,7 +185,7 @@ async def chat(request: MessageRequest):
     audio_path = "response.mp3"
     tts.save(audio_path)
     return {"response": response1,
-           "audioUrl": f"http://localhost:5000/audio/{audio_path}"}
+           "audioUrl": f"https://redfernstech-hotel-gtts.hf.space/audio/{audio_path}"}
 
 @app.get("/")
 def read_root(request: Request):