Update app.py

app.py

@@ -6,10 +6,6 @@ import json
 ACCESS_TOKEN = os.getenv("HF_TOKEN")
 print("Access token loaded.")
 
-# Initialize the HF Inference Client
-client = InferenceClient(token=ACCESS_TOKEN)
-print("Hugging Face Inference Client initialized.")
-
 def respond(
     message,
     history: list[tuple[str, str]],
@@ -30,6 +26,11 @@ def respond(
     print(f"Selected model (custom_model): {custom_model}")
     print(f"Selected provider: {provider}")
 
+    # Initialize the Inference Client with the provider
+    # Provider is specified during initialization, not in the method call
+    client = InferenceClient(token=ACCESS_TOKEN, provider=provider)
+    print(f"Hugging Face Inference Client initialized with {provider} provider.")
+
     # Convert seed to None if -1 (meaning random)
     if seed == -1:
         seed = None
@@ -72,14 +73,14 @@ def respond(
     if seed is not None:
         parameters["seed"] = seed
 
-    # Use the InferenceClient for making the request with proper provider selection
+    # Use the InferenceClient for making the request
     try:
         # Create a generator for the streaming response
+        # The provider is already set when initializing the client
         stream = client.chat_completion(
             model=model_to_use,
             messages=messages,
             stream=True,
-            provider=provider,  # Use the selected provider
             **parameters  # Pass all other parameters
         )