Update app.py

app.py

@@ -20,7 +20,8 @@ if not hf_token:
 
 tokenizer = AutoTokenizer.from_pretrained(model_path)
 quantization_config = BitsAndBytesConfig(load_in_8bit=True)
-model = AutoModelForCausalLM.from_pretrained( model_path, quantization_config=quantization_config).to("cuda")
+model = AutoModelForCausalLM.from_pretrained( model_path, quantization_config=quantization_config)
+
 @spaces.GPU
 def generate_text(prompt, temperature, max_length):
     inputs = tokenizer.encode(prompt, return_tensors="pt").to("cuda")