Update utils.py

utils.py

@@ -19,6 +19,7 @@ from pygments.lexers import guess_lexer,get_lexer_by_name
 from pygments.formatters import HtmlFormatter
 import transformers
 from transformers import AutoTokenizer, AutoModelForCausalLM, GPT2Tokenizer, GPT2LMHeadModel
+import auto_gptq
 from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
 
 
@@ -129,6 +130,7 @@ def load_tokenizer_and_model_gpt2(base_model,load_8bit=False):
 
 
 def load_tokenizer_and_model_bloke_gpt(base_model, model_basename):
+    use_triton = False
     if torch.cuda.is_available():
         device = "cuda"
     else: