change loader to PEFT

app.py

@@ -1,13 +1,18 @@
 import gradio as gr
 from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
+from peft import PeftModel  # Pastikan PEFT terinstall
 import torch
 
-# Gunakan model yang sudah diunggah ke Hugging Face
-model_path = "FadQ/gemma-2b-diary-consultaton-chatbot"
+# Model dasar dan adapter LoRA
+base_model = "google/gemma-2b-it"
+adapter_path = "FadQ/gemma-2b-diary-consultaton-chatbot"
 
-# Load model dan tokenizer dengan `trust_remote_code=True`
-model = AutoModelForCausalLM.from_pretrained(model_path, trust_remote_code=True, torch_dtype=torch.float16, device_map="auto")
-tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
+# Load model dasar dengan trust_remote_code=True
+model = AutoModelForCausalLM.from_pretrained(base_model, trust_remote_code=True, torch_dtype=torch.float16, device_map="auto")
+tokenizer = AutoTokenizer.from_pretrained(base_model, trust_remote_code=True)
+
+# Gabungkan adapter dengan model dasar
+model = PeftModel.from_pretrained(model, adapter_path)
 
 # Buat pipeline dengan tokenizer yang sesuai
 pipe = pipeline("text-generation", model=model, tokenizer=tokenizer, device=0)