Update app.py

app.py

@@ -3,9 +3,16 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
 
 # Load your model and tokenizer from Hugging Face Hub (forcing CPU usage)
-model_name = "PierreJousselin/lora_model"  # Replace with the name you used on Hugging Face
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(model_name, device_map="cpu")  # Force model to load on CPU
+
+# model_name = "PierreJousselin/lora_model"  # Replace with the name you used on Hugging Face
+# model = AutoModelForCausalLM.from_pretrained(model_name, device_map="cpu")  # Force model to load on CPU
+
+model_id = "unsloth/llama-3.2-3b-instruct-bnb-4bit"
+peft_model_id = "PierreJousselin/lora_model"
+
+model = AutoModelForCausalLM.from_pretrained(model_id)
+model.load_adapter(peft_model_id)
+tokenizer = AutoTokenizer.from_pretrained(model_id)
 
 # Ensure pad_token_id is set to eos_token_id to avoid errors
 model.config.pad_token_id = model.config.eos_token_id