fix(strategy): fixed that old tune_cfg cannot be matched correctly in _find_tuning_history. (#1303)

Author: intel-zhangyi

neural_compressor/strategy/strategy.py

@@ -228,7 +228,7 @@ def traverse(self):
             logger.debug("Dump current tuning configuration:")
             logger.debug(tune_cfg)
             self.q_model = self.adaptor.quantize(
-                tune_cfg, self.model, self.calib_dataloader, self.q_func)
+                copy.deepcopy(tune_cfg), self.model, self.calib_dataloader, self.q_func)
             self.algo.calib_iter = tune_cfg['calib_iteration']
             self.algo.q_model = self.q_model
             # TODO align the api to let strategy has access to pre_optimized model