diff --git a/vllm/model_executor/models/llava.py b/vllm/model_executor/models/llava.py
index 4b3fd5d3cd681..4090e1979fe53 100644
--- a/vllm/model_executor/models/llava.py
+++ b/vllm/model_executor/models/llava.py
@@ -596,7 +596,7 @@ def apply(
         result = super().apply(prompt_text, mm_data, hf_processor_mm_kwargs)
 
         mm_items = self._to_mm_items(mm_data)
-        mm_item_counts = mm_items.get_item_counts()
+        mm_item_counts = mm_items.get_all_counts()
         mm_kwargs = result["mm_kwargs"]
 
         # We reimplement the functionality of MLlavaProcessor from
diff --git a/vllm/model_executor/models/phi3v.py b/vllm/model_executor/models/phi3v.py
index 413655ff60fe7..0402c556296ce 100644
--- a/vllm/model_executor/models/phi3v.py
+++ b/vllm/model_executor/models/phi3v.py
@@ -392,7 +392,7 @@ def get_replacement_phi3v(item_idx: int):
 
             return [_IMAGE_TOKEN_ID] * num_tokens + [bos_token_id]
 
-        num_images = mm_items.get_item_count("image", strict=False)
+        num_images = mm_items.get_count("image", strict=False)
 
         return [
             PromptReplacement(
diff --git a/vllm/multimodal/parse.py b/vllm/multimodal/parse.py
index 1e190a6999f81..21f3ca8f3487a 100644
--- a/vllm/multimodal/parse.py
+++ b/vllm/multimodal/parse.py
@@ -169,7 +169,7 @@ class MultiModalDataItems(UserDict[str, ModalityDataItems[Any, Any]]):
     corresponds to a list.
     """
 
-    def get_item_count(self, modality: str, *, strict: bool = True) -> int:
+    def get_count(self, modality: str, *, strict: bool = True) -> int:
         """
         Get the number of data items belonging to a modality.
         
@@ -186,7 +186,7 @@ def get_item_count(self, modality: str, *, strict: bool = True) -> int:
 
         return self[modality].get_count()
 
-    def get_item_counts(self) -> Mapping[str, int]:
+    def get_all_counts(self) -> Mapping[str, int]:
         """Get the number of items belonging to each modality."""
         return {m: items.get_count() for m, items in self.items()}
 
diff --git a/vllm/multimodal/processing.py b/vllm/multimodal/processing.py
index fe6448d99da98..7c523735a73d7 100644
--- a/vllm/multimodal/processing.py
+++ b/vllm/multimodal/processing.py
@@ -747,7 +747,7 @@ def _apply_hf_processor_missing(
         cached items; instead, we rely on our own prompt replacement logic
         for the full text.
         """
-        mm_missing_counts = mm_missing_data_items.get_item_counts()
+        mm_missing_counts = mm_missing_data_items.get_all_counts()
 
         prompt_ids, _ = self._apply_hf_processor(
             prompt_text=prompt_text,
@@ -844,7 +844,7 @@ def _cached_apply_hf_processor(
             mm_merged_field_items[modality] = merged_modal_items_lst
 
         if self.enable_sanity_checks:
-            mm_missing_counts = mm_missing_data_items.get_item_counts()
+            mm_missing_counts = mm_missing_data_items.get_all_counts()
             assert all(
                 item_count == mm_missing_counts[modality]
                 for modality, item_count in mm_missing_next_idx.items()), dict(
@@ -857,7 +857,7 @@ def _cached_apply_hf_processor(
         )
 
         if self.enable_sanity_checks:
-            mm_item_counts = mm_data_items.get_item_counts()
+            mm_item_counts = mm_data_items.get_all_counts()
 
             for modality, item_count in mm_item_counts.items():
                 for item_idx in range(item_count):
@@ -967,7 +967,7 @@ def apply(
 
         # If HF processor already inserts placeholder tokens,
         # there is no need for us to insert them
-        mm_item_counts = mm_items.get_item_counts()
+        mm_item_counts = mm_items.get_all_counts()
         all_placeholders = self._find_placeholders(prompt_repls, prompt_ids,
                                                    mm_item_counts)