Update app.py

app.py

@@ -32,7 +32,7 @@ if not HUGGINGFACE_API_TOKEN:
     print("Go to your Space's Settings > Repository secrets to add it.")
 
 # The default AI model
-DEFAULT_MODEL = "mistralai/Mistral-7B-Instruct-v0.3"
+DEFAULT_MODEL = "Qwen/Qwen2.5-72B-Instruct"
 
 
 # ══════════════════════════════════════════════════════════════════
@@ -184,13 +184,8 @@ FEW_SHOT_OUTPUT = json.dumps({
 from huggingface_hub import InferenceClient
 
 def call_huggingface_model(referral_text: str, model_id: str = DEFAULT_MODEL) -> str:
-    """
-    Send the referral text to a Hugging Face model and get the response.
-    Uses InferenceClient which handles routing internally.
-    """
     client = InferenceClient(token=HUGGINGFACE_API_TOKEN)
 
-    # Build the conversation messages
     messages = [
         {"role": "system", "content": SYSTEM_PROMPT},
         {"role": "user", "content": f"Example referral:\n{FEW_SHOT_INPUT}"},
@@ -405,12 +400,12 @@ with gr.Blocks(
             gr.Markdown("### Input")
             model_dropdown = gr.Dropdown(
                 choices=[
-                    "mistralai/Mistral-7B-Instruct-v0.3",
-                    "HuggingFaceH4/zephyr-7b-beta",
-                    "meta-llama/Llama-2-13b-chat-hf",
-                    "tiiuae/falcon-7b-instruct",
+                    "Qwen/Qwen2.5-72B-Instruct",
+                    "meta-llama/Meta-Llama-3.1-8B-Instruct",
+                    "mistralai/Mistral-Nemo-Instruct-2407",
+                    "microsoft/Phi-3.5-mini-instruct",
                 ],
-                value="mistralai/Mistral-7B-Instruct-v0.3",
+                value="Qwen/Qwen2.5-72B-Instruct",
                 label="AI Model",
                 info="Try different models to compare results"
             )