Update app.py

app.py

@@ -1,42 +1,30 @@
 import gradio as gr
-from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
+from llama_cpp import Llama
 
-# Load model & tokenizer dari HuggingFace Hub
-model_id = "deepseek-ai/DeepSeek-V3.1-Base"
+# Ganti dengan path model GGUF (download dulu ke Space atau pakai hf:// link)
+MODEL_PATH = "DeepSeek-V3.1-Chat-Q4_K_M.gguf"
 
-tokenizer = AutoTokenizer.from_pretrained(model_id)
-model = AutoModelForCausalLM.from_pretrained(
-    model_id,
-    device_map="auto",
-    torch_dtype="auto"
+# Load model quantized (ringan untuk CPU 16GB)
+llm = Llama(
+    model_path=MODEL_PATH,
+    n_ctx=2048,
+    n_threads=4
 )
 
-# Buat pipeline text-generation
-chatbot = pipeline(
-    "text-generation",
-    model=model,
-    tokenizer=tokenizer,
-    max_new_tokens=512
-)
-
-# Fungsi chatbot
 def respond(message, history):
     prompt = ""
     for user, bot in history:
         prompt += f"User: {user}\nAssistant: {bot}\n"
     prompt += f"User: {message}\nAssistant:"
 
-    output = chatbot(prompt, do_sample=True, temperature=0.7, top_p=0.9)[0]["generated_text"]
-    
-    # Ambil jawaban setelah "Assistant:"
-    response = output.split("Assistant:")[-1].strip()
+    output = llm(prompt, max_tokens=512, temperature=0.7, top_p=0.9)
+    response = output["choices"][0]["text"].strip()
     return response
 
-# Gradio UI
 with gr.Blocks() as demo:
-    gr.Markdown("# 🤖 DeepSeek V3.1 Chatbot")
-    chatbot_ui = gr.Chatbot()
-    msg = gr.Textbox(placeholder="Tulis pesanmu di sini...")
+    gr.Markdown("# 🤖 DeepSeek V3.1 Chatbot (Quantized, CPU)")
+    chatbot = gr.Chatbot()
+    msg = gr.Textbox(placeholder="Tulis pesan di sini...")
     clear = gr.Button("Clear")
 
     def user_input(message, history):
@@ -44,7 +32,7 @@ with gr.Blocks() as demo:
         history.append((message, response))
         return "", history
 
-    msg.submit(user_input, [msg, chatbot_ui], [msg, chatbot_ui])
-    clear.click(lambda: None, None, chatbot_ui, queue=False)
+    msg.submit(user_input, [msg, chatbot], [msg, chatbot])
+    clear.click(lambda: None, None, chatbot, queue=False)
 
 demo.launch()