Update app.py

app.py

@@ -7,33 +7,30 @@ from threading import Thread
 hf_token = os.getenv("HF_TOKEN")
 model_id = "ZyperAI/Z-AI-0.1-1.1B-Code.web"
 
-@gr.cache_resource
-def load_assets():
-    # Adding use_fast=False can bypass some backend instantiation errors 
-    # while still providing full functionality.
-    tokenizer = AutoTokenizer.from_pretrained(
-        model_id, 
-        token=hf_token, 
-        use_fast=False 
-    )
-    model = AutoModelForCausalLM.from_pretrained(
-        model_id, 
-        torch_dtype=torch.float32, 
-        device_map="cpu",
-        token=hf_token
-    )
-    return tokenizer, model
-
-tokenizer, model = load_assets()
+# Professional loading logic: Since Gradio 6.x runs as a persistent server, 
+# global variables are naturally 'cached' for the duration of the process.
+print("Loading model and tokenizer...")
+tokenizer = AutoTokenizer.from_pretrained(
+    model_id, 
+    token=hf_token, 
+    use_fast=False
+)
+model = AutoModelForCausalLM.from_pretrained(
+    model_id, 
+    torch_dtype=torch.float32, 
+    device_map="cpu",
+    token=hf_token
+)
+print("Model loaded successfully.")
 
 def generate_code(prompt, history):
-    messages = []
-    for human, assistant in history:
-        messages.append({"role": "user", "content": human})
-        messages.append({"role": "assistant", "content": assistant})
-    messages.append({"role": "user", "content": prompt})
+    # Gradio 6.x uses the 'messages' format by default
+    # history will look like: [{"role": "user", "content": "..."}, {"role": "assistant", "content": "..."}]
     
-    # Apply the chat template specific to the ZyperAI model
+    # Prepend existing history to the current prompt
+    messages = history + [{"role": "user", "content": prompt}]
+    
+    # Prepare inputs using the model's chat template
     inputs = tokenizer.apply_chat_template(
         messages, 
         add_generation_prompt=True, 
@@ -59,23 +56,26 @@ def generate_code(prompt, history):
         response += new_text
         yield response
 
-# Professional Dark-Themed UI
-with gr.Blocks(theme=gr.themes.Default(primary_hue="blue", neutral_hue="zinc")) as demo:
+# Gradio 6.x UI setup
+with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue")) as demo:
     gr.Markdown("# ⚡ **Z-AI Web Coder**")
-    gr.Markdown("Optimized 1.1B model for high-speed web development.")
     
-    chatbot = gr.Chatbot(height=550, show_copy_button=True, type="messages")
+    # In Gradio 6, type="messages" is the standard for the chatbot component
+    chatbot = gr.Chatbot(height=500, show_copy_button=True, type="messages")
     
     with gr.Row():
         msg = gr.Textbox(
-            placeholder="E.g., Create a glassmorphism login form with Tailwind CSS...", 
+            placeholder="E.g., Create a responsive navigation bar with CSS...", 
             show_label=False, 
             scale=9
         )
-        submit = gr.Button("Generate", variant="primary", scale=1)
+        submit = gr.Button("Build", variant="primary", scale=1)
 
+    # Wire up the events
     msg.submit(generate_code, [msg, chatbot], [chatbot])
     submit.click(generate_code, [msg, chatbot], [chatbot])
+    
+    # Clear the input box after submission
     msg.submit(lambda: "", None, [msg])
     submit.click(lambda: "", None, [msg])