common masked loss func, apply to all training script

library/train_util.py

@@ -3028,6 +3028,7 @@ def add_training_arguments(parser: argparse.ArgumentParser, support_dreambooth:
         "--full_bf16", action="store_true", help="bf16 training including gradients / 勾配も含めてbf16で学習する"
     )  # TODO move to SDXL training, because it is not supported by SD1/2
     parser.add_argument("--fp8_base", action="store_true", help="use fp8 for base model / base modelにfp8を使う")
+
     parser.add_argument(
         "--ddp_timeout",
         type=int,
@@ -3090,6 +3091,7 @@ def add_training_arguments(parser: argparse.ArgumentParser, support_dreambooth:
         default=None,
         help="specify WandB API key to log in before starting training (optional). / WandB APIキーを指定して学習開始前にログインする（オプション）",
     )
+
     parser.add_argument(
         "--noise_offset",
         type=float,
@@ -3252,6 +3254,20 @@ def add_training_arguments(parser: argparse.ArgumentParser, support_dreambooth:
         )
 
 
+def add_masked_loss_arguments(parser: argparse.ArgumentParser):
+    parser.add_argument(
+        "--conditioning_data_dir",
+        type=str,
+        default=None,
+        help="conditioning data directory / 条件付けデータのディレクトリ",
+    )
+    parser.add_argument(
+        "--masked_loss",
+        action="store_true",
+        help="apply mask for calculating loss. conditioning_data_dir is required for dataset. / 損失計算時にマスクを適用する。datasetにはconditioning_data_dirが必要",
+    )
+
+
 def verify_training_args(args: argparse.Namespace):
     r"""
     Verify training arguments. Also reflect highvram option to global variable