Update Gradio settings for ZeroGPU

app.py

@@ -11,7 +11,7 @@ login(token=hf_token)
 
 # Model repository IDs
 base_model_id = "meta-llama/Llama-3.2-3B-Instruct"
-peft_model_id = "ubiodee/Plutuslearn-Llama-3.2-3B-Instruct" 
+peft_model_id = "ubiodee/Plutuslearn-Llama-3.2-3B-Instruct"  # Replace with your model repo
 
 # Load the tokenizer from the fine-tuned model
 tokenizer = AutoTokenizer.from_pretrained(peft_model_id, token=hf_token)
@@ -26,7 +26,7 @@ base_model = AutoModelForCausalLM.from_pretrained(
     trust_remote_code=True
 )
 
-# Resize the base model's embeddings to match the fine-tuned tokenizer
+# Resize embeddings to match fine-tuned tokenizer
 base_model.resize_token_embeddings(len(tokenizer))
 
 # Load the PEFT adapter
@@ -42,7 +42,7 @@ def predict(text, max_length=100):
     except Exception as e:
         return f"Error during inference: {str(e)}"
 
-# Create Gradio interface for ZeroGPU
+# Create Gradio interface
 demo = gr.Interface(
     fn=predict,
     inputs=[
@@ -52,7 +52,8 @@ demo = gr.Interface(
     outputs=gr.Textbox(label="Model Output"),
     title="LearnPlutus Demo",
     description="Test the fine-tuned Llama-3.2-3B-Instruct model on ZeroGPU.",
-    allow_flagging="never"
+    flagging_mode="never"  # Updated from allow_flagging
 )
 
-demo.launch(server_name="0.0.0.0", server_port=7860)
+# Launch with ZeroGPU settings
+demo.launch(server_name="0.0.0.0", server_port=7860, share=False, ssr_mode=False)