diff --git a/src/transformers/trainer.py b/src/transformers/trainer.py
index 6567de9186201f..0c3ea207bb140b 100755
--- a/src/transformers/trainer.py
+++ b/src/transformers/trainer.py
@@ -2770,6 +2770,7 @@ def init_git_repo(self, at_init: bool = False):
                 self.args.output_dir,
                 clone_from=repo_name,
                 use_auth_token=use_auth_token,
+                private=self.args.hub_private_repo,
             )
         except EnvironmentError:
             if self.args.overwrite_output_dir and at_init:
diff --git a/src/transformers/training_args.py b/src/transformers/training_args.py
index 3f05b772721237..adaf50da8ffe41 100644
--- a/src/transformers/training_args.py
+++ b/src/transformers/training_args.py
@@ -416,6 +416,8 @@ class TrainingArguments:
         hub_token (`str`, *optional*):
             The token to use to push the model to the Hub. Will default to the token in the cache folder obtained with
             `huggingface-cli login`.
+        hub_private_repo (`bool`, *optional*, defaults to `False`):
+            If True, the Hub repo will be set to private.
         gradient_checkpointing (`bool`, *optional*, defaults to `False`):
             If True, use gradient checkpointing to save memory at the expense of slower backward pass.
         include_inputs_for_metrics (`bool`, *optional*, defaults to `False`):
@@ -738,6 +740,7 @@ class TrainingArguments:
         metadata={"help": "The hub strategy to use when `--push_to_hub` is activated."},
     )
     hub_token: str = field(default=None, metadata={"help": "The token to use to push to the Model Hub."})
+    hub_private_repo: bool = field(default=False, metadata={"help": "Whether the model repository is private or not."})
     gradient_checkpointing: bool = field(
         default=False,
         metadata={