Update app.py

app.py

@@ -1,47 +1,23 @@
-import gradio as gr
-from transformers import AutoTokenizer, AutoModelForCausalLM
-import torch
+from huggingface_hub import InferenceClient
+import os
 
-# Load Model dan Tokenizer
-model_name = "meta-llama/Llama-3.3-70B-Instruct"
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(
-    model_name,
-    device_map="auto",
-    torch_dtype=torch.float16
-)
+# Ambil token dari Repository secrets
+api_token = os.getenv("LLM")
+
+# Inisialisasi Inference Client
+client = InferenceClient(api_key=api_token)
 
-# Fungsi Chatbot
-def chat(input_text):
-    # Tokenisasi input
-    inputs = tokenizer(
-        input_text,
-        return_tensors="pt",
-        truncation=True,  # Potong jika terlalu panjang
-        max_length=512    # Batasi panjang input
-    ).to("cuda")
-    
-    # Hasilkan output menggunakan model
-    outputs = model.generate(
-        **inputs,
-        max_length=200,       # Panjang maksimal output
-        top_p=0.9,            # Sampling berdasarkan probabilitas kumulatif
-        temperature=0.7,      # Kontrol kreativitas output
-        num_return_sequences=1 # Jumlah respons yang dihasilkan
-    )
-    
-    # Decode output menjadi teks
-    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
-    return response
+# Pesan dari user
+messages = [
+    {"role": "user", "content": "What is the capital of France?"}
+]
 
-# Membuat antarmuka Gradio
-iface = gr.Interface(
-    fn=chat,
-    inputs="text",
-    outputs="text",
-    title="Llama Chatbot",
-    description="Chatbot AI menggunakan model Llama-3.3-70B-Instruct."
+# Inferensi menggunakan model
+completion = client.chat.completions.create(
+    model="meta-llama/Llama-3.3-70B-Instruct",
+    messages=messages,
+    max_tokens=500
 )
 
-# Meluncurkan Gradio
-iface.launch()
+# Cetak jawaban dari model
+print(completion.choices[0].message)