Update app.py

app.py

@@ -6,9 +6,9 @@ import gradio as gr
 # Use GPU if available
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
-# Base model and adapter paths
+# Base model and adapter paths (updated for Hugging Face repo)
 base_model_name = "microsoft/phi-2"  # Pull from HF Hub directly
-adapter_path = "Shriti09/Microsoft-Phi-QLora"  # Your uploaded adapter folder in Space repo
+adapter_path = "Shriti09/phi2-qlora-adapter"  # Update with your Hugging Face repo path
 
 print("🔧 Loading base model...")
 base_model = AutoModelForCausalLM.from_pretrained(
@@ -61,7 +61,8 @@ def chat_fn(message, history):
 with gr.Blocks(theme=gr.themes.Soft()) as demo:
     gr.Markdown("<h1>🧠 Phi-2 QLoRA Chatbot</h1>")
 
-    chatbot = gr.Chatbot()
+    # Use 'type' parameter to specify message format for gr.Chatbot()
+    chatbot = gr.Chatbot(type="messages")  # Use 'messages' type for structured messages
     message = gr.Textbox(label="Your message:")
     clear = gr.Button("Clear chat")
 
@@ -71,5 +72,5 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
     clear.click(lambda: [], None, chatbot)
     clear.click(lambda: [], None, state)
 
-# Run with queue for multiple users
-demo.queue(concurrency_count=2).launch()
+# Run the app without the 'concurrency_count' argument
+demo.queue().launch()