Remove metric savor for making benchmark

Author: mingukkang

src/worker.py

@@ -921,8 +921,6 @@ def evaluate(self, step, metrics, writing=True, training=False):
                     misc.save_dict_npy(directory=join(self.RUN.save_dir, "statistics", self.run_name, "train" if training else "eval"),
                                        name="metrics",
                                        dictionary=save_dict)
-                    with open("./eval_pickles/"+self.run_name + "-" + self.RUN.eval_backbone + "-" +self.RUN.ref_dataset +".pickle", "wb") as f:
-                        pickle.dump(metric_dict, f)
 
         misc.make_GAN_trainable(self.Gen, self.Gen_ema, self.Dis)
         return is_best