improve aqlm optim

Former-commit-id: 81be999b407e988c2f42764d827ac859d079ed3e

src/llmtuner/hparams/parser.py

@@ -226,6 +226,7 @@ def get_train_args(args: Optional[Dict[str, Any]] = None) -> _TRAIN_CLS:
         torch.bfloat16 if training_args.bf16 else (torch.float16 if training_args.fp16 else None)
     )
     model_args.model_max_length = data_args.cutoff_len
+    model_args.aqlm_optimization = not training_args.predict_with_generate
 
     # Log on each process the small summary:
     logger.info(
@@ -262,6 +263,7 @@ def get_eval_args(args: Optional[Dict[str, Any]] = None) -> _EVAL_CLS:
     _set_transformers_logging()
     _verify_model_args(model_args, finetuning_args)
     _check_dependencies(disabled=finetuning_args.disable_version_checking)
+    model_args.aqlm_optimization = True
 
     if data_args.template is None:
         raise ValueError("Please specify which `template` to use.")