[Llama4] Remove image_sizes arg and deprecate vision_feature_layer

src/transformers/models/llama4/configuration_llama4.py

@@ -14,6 +14,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import warnings
 
 from ...configuration_utils import PretrainedConfig, layer_type_validation
 from ...utils import logging
@@ -56,7 +57,6 @@ class Llama4VisionConfig(PretrainedConfig):
             The size (resolution) of each patch.
         norm_eps (`float`, *optional*, defaults to 1e-05):
             The epsilon used by the layer normalization layers.
-        vision_feature_layer (``, *optional*, defaults to -1): TODO
         vision_feature_select_strategy (`int`, *optional*, defaults to `"default"`): TODO
         initializer_range (`float`, *optional*, defaults to 0.02):
             The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
@@ -93,7 +93,6 @@ def __init__(
         image_size: int = 448,
         patch_size: int = 14,
         norm_eps: float = 1e-5,
-        vision_feature_layer=-1,
         vision_feature_select_strategy="default",
         initializer_range: float = 0.02,
         pixel_shuffle_ratio=0.5,
@@ -122,9 +121,23 @@ def __init__(
         self.multi_modal_projector_bias = multi_modal_projector_bias
         self.projector_dropout = projector_dropout
         self.attention_dropout = attention_dropout
-        self.vision_feature_layer = vision_feature_layer
         self.vision_feature_select_strategy = vision_feature_select_strategy
         self.rope_theta = rope_theta
+
+        self._vision_feature_layer = kwargs.get("vision_feature_layer", -1)
+
+        @property
+        def vision_feature_layer(self):
+            warnings.warn(
+                "The `vision_feature_layer` attribute is deprecated and will be removed in v4.58.0.",
+                FutureWarning,
+            )
+            return self._vision_feature_layer
+
+        @vision_feature_layer.setter
+        def vision_feature_layer(self, value):
+            self._vision_feature_layer = value
+
         super().__init__(**kwargs)
 
 

src/transformers/models/llama4/modeling_llama4.py

@@ -1173,7 +1173,6 @@ def get_decoder(self):
     def get_image_features(
         self,
         pixel_values: torch.FloatTensor,
-        vision_feature_layer: Union[int, list[int]],
         vision_feature_select_strategy: str,
         **kwargs,
     ):
@@ -1183,10 +1182,6 @@ def get_image_features(
         Args:
             pixel_values (`torch.FloatTensor]` of shape `(batch_size, channels, height, width)`)
                The tensors corresponding to the input images.
-            vision_feature_layer (`Union[int, list[int]]`):
-                The index of the layer to select the vision feature. If multiple indices are provided,
-                the vision feature of the corresponding indices will be concatenated to form the
-                vision features.
             vision_feature_select_strategy (`str`):
                 The feature selection strategy used to select the vision feature from the vision backbone.
                 Can be one of `"default"` or `"full"`
@@ -1224,6 +1219,7 @@ def get_placeholder_mask(
         return special_image_mask
 
     @auto_docstring
+    @deprecate_kwarg("vision_feature_layer", version="4.58")
     def forward(
         self,
         input_ids: Optional[torch.LongTensor] = None,
@@ -1241,7 +1237,6 @@ def forward(
         return_dict: Optional[bool] = None,
         cache_position: Optional[torch.LongTensor] = None,
         logits_to_keep: Union[int, torch.Tensor] = 0,
-        image_sizes: Optional[torch.Tensor] = None,
         **kwargs: Unpack[TransformersKwargs],
     ) -> Union[tuple, Llama4CausalLMOutputWithPast]:
         r"""
@@ -1277,11 +1272,6 @@ def forward(
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
         )
         return_dict = return_dict if return_dict is not None else self.config.use_return_dict
-        vision_feature_layer = (
-            vision_feature_layer
-            if vision_feature_layer is not None
-            else self.config.vision_config.vision_feature_layer
-        )
         vision_feature_select_strategy = (
             vision_feature_select_strategy
             if vision_feature_select_strategy is not None
@@ -1302,9 +1292,7 @@ def forward(
         if pixel_values is not None:
             image_features = self.get_image_features(
                 pixel_values=pixel_values,
-                vision_feature_layer=vision_feature_layer,
                 vision_feature_select_strategy=vision_feature_select_strategy,
-                image_sizes=image_sizes,
             )
 
             vision_flat = image_features.view(-1, image_features.size(-1))