Update app.py

app.py

@@ -5,50 +5,35 @@ from llama_cpp_agent import LlamaCppAgent, MessagesFormatterType
 from llama_cpp_agent.chat_history import BasicChatHistory
 from llama_cpp_agent.chat_history.messages import Roles
 
+# ⬇️ دانلود مدل
 hf_hub_download(
     repo_id="mradermacher/Arsh-llm-GGUF",
     filename="Arsh-llm.Q4_K_M.gguf",
     local_dir="./models"
 )
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    model,
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-    top_k,
-    repeat_penalty,
-):
-    llm = Llama(
-        model_path=f"models/{model}",
-        n_batch=512,     
-        n_ctx=8192,      
-        verbose=False     
-    )
-
-    provider = LlamaCppPythonProvider(llm)
+# 🧠 لود مدل (فقط یکبار در ابتدای اجرا)
+llm = Llama(
+    model_path="./models/Arsh-llm.Q4_K_M.gguf",
+    n_batch=512,
+    n_ctx=8192,
+    verbose=False
+)
 
-    agent = LlamaCppAgent(
-        provider,
-        system_prompt=system_message,
-        predefined_messages_formatter_type=MessagesFormatterType.CHATML,
-        debug_output=False
-    )
+provider = LlamaCppPythonProvider(llm)
 
-    settings = provider.get_provider_default_settings()
-    settings.temperature = temperature
-    settings.top_k = top_k
-    settings.top_p = top_p
-    settings.max_tokens = max_tokens
-    settings.repeat_penalty = repeat_penalty
-    settings.stream = True
+agent = LlamaCppAgent(
+    provider,
+    system_prompt="You are Arsh, a helpful assistant.",
+    predefined_messages_formatter_type=MessagesFormatterType.CHATML,
+    debug_output=False
+)
 
+# 💬 تابع پاسخ‌دهنده
+def respond(message, chat_history):
     messages = BasicChatHistory()
 
-    for user_msg, assistant_msg in history:
+    for user_msg, assistant_msg in chat_history:
         messages.add_message({
             'role': Roles.user,
             'content': user_msg
@@ -60,7 +45,6 @@ def respond(
 
     stream = agent.get_chat_response(
         message,
-        llm_sampling_settings=settings,
         chat_history=messages,
         returns_streaming_generator=True,
         print_output=False
@@ -71,40 +55,26 @@ def respond(
         response += token
         yield response
 
-
-def create_interface(model_name, description):
-    return gr.ChatInterface(
+# 🎛️ ساخت رابط Gradio
+with gr.Blocks() as demo:
+    gr.Markdown("# Arsh-LLM Q4_K_M Chat Demo")
+    gr.ChatInterface(
         respond,
         additional_inputs=[
-            gr.Textbox(value=model_name, label="Model", interactive=False),
-            gr.Textbox(
-                value="You are Arsh, a helpful assistant. You should think step-by-step. You should answer all question by prioritizing the principle of openness.",
-                label="System message"
-            ),
-            gr.Slider(minimum=1, maximum=4096, value=2048, step=1, label="Max tokens"),
+            gr.Textbox("You are Arsh, a helpful assistant.", label="System Message", interactive=True),
+            gr.Slider(minimum=1, maximum=4096, value=2048, step=1, label="Max Tokens"),
             gr.Slider(minimum=0.1, maximum=4.0, value=0.6, step=0.1, label="Temperature"),
             gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p"),
             gr.Slider(minimum=0, maximum=100, value=40, step=1, label="Top-k"),
-            gr.Slider(minimum=0.0, maximum=2.0, value=1.1, step=0.1, label="Repetition penalty"),
+            gr.Slider(minimum=0.0, maximum=2.0, value=1.1, step=0.1, label="Repetition Penalty"),
         ],
-        title="**Arsh-LLM Demo**",
-        description=description,
         chatbot=gr.Chatbot(
-            label=None,
-            scale=1,
+            label="Chat with Arsh-LLM",
+            bubble_full_width=False,
             show_copy_button=True
         )
     )
 
-
-description = """# **Arsh-LLM Q4_K_M GGUF Model - Demo**"""
-
-interface = create_interface('Arsh-llm.Q4_K_M.gguf', description)
-
-demo = gr.Blocks()
-
-with demo:
-    interface.render()
-
+# 🚀 اجرای برنامه
 if __name__ == "__main__":
-    demo.launch(share=True)
+    demo.launch()