Update app.py

app.py

@@ -7,8 +7,7 @@ import spaces
 # Load the model and tokenizer
 peft_model_id = "rootxhacker/CodeAstra-7B"
 config = PeftConfig.from_pretrained(peft_model_id)
-model = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, return_dict=True, load_in_4bit=True, device_map='auto')
-model.to('cuda')
+model = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, return_dict=True, load_in_4bit=True,device_map={"":0})
 tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)
 
 # Load the Lora model