Rename clashing method names for vLLM model protocol (#27583)

Signed-off-by: Harry Mellor <[email protected]>

Author: hmellor

docs/contributing/model/basic.md

@@ -56,13 +56,13 @@ The initialization code should look like this:
 
 ### Computation Code
 
-- Add a `get_input_embeddings` method inside `MyModel` module that returns the text embeddings given `input_ids`. This is equivalent to directly calling the text embedding layer, but provides a unified interface in case `MyModel` is used within a composite multimodal model.
+- Add a `embed_input_ids` method inside `MyModel` module that returns the text embeddings given `input_ids`. This is equivalent to directly calling the text embedding layer, but provides a unified interface in case `MyModel` is used within a composite multimodal model.
 
 ```python
 class MyModel(nn.Module):
         ...
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
         ... 
 ```
 

docs/contributing/model/multimodal.md

@@ -36,7 +36,7 @@ Further update the model as follows:
 
   More conveniently, you can simply pass `**kwargs` to the [forward][torch.nn.Module.forward] method and retrieve the keyword parameters for multimodal inputs from it.
 
-- Implement [get_multimodal_embeddings][vllm.model_executor.models.interfaces.SupportsMultiModal.get_multimodal_embeddings] that returns the embeddings from running the multimodal inputs through the multimodal tokenizer of the model. Below we provide a boilerplate of a typical implementation pattern, but feel free to adjust it to your own needs.
+- Implement [embed_multimodal][vllm.model_executor.models.interfaces.SupportsMultiModal.embed_multimodal] that returns the embeddings from running the multimodal inputs through the multimodal tokenizer of the model. Below we provide a boilerplate of a typical implementation pattern, but feel free to adjust it to your own needs.
 
     ??? code
 
@@ -49,7 +49,7 @@ Further update the model as follows:
                 image_features = self.vision_encoder(image_input)
                 return self.multi_modal_projector(image_features)
 
-            def get_multimodal_embeddings(
+            def embed_multimodal(
                 self,
                 **kwargs: object,
             ) -> MultiModalEmbeddings | None:
@@ -69,7 +69,7 @@ Further update the model as follows:
 !!! note
     By default, vLLM merges the multimodal embeddings into text embeddings depending on the information of their locations defined in
     [PlaceholderRange][vllm.multimodal.inputs.PlaceholderRange] from input processing.
-    This logic can be found at [get_input_embeddings][vllm.model_executor.models.interfaces.SupportsMultiModal.get_input_embeddings].
+    This logic can be found at [embed_input_ids][vllm.model_executor.models.interfaces.SupportsMultiModal.embed_input_ids].
 
     You may override this method if additional logic is required for your model when merging embeddings. 
 

vllm/model_executor/models/apertus.py

@@ -382,7 +382,7 @@ def __init__(
             ["hidden_states", "residual"], config.hidden_size
         )
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
         return self.embed_tokens(input_ids)
 
     def forward(
@@ -396,7 +396,7 @@ def forward(
             if inputs_embeds is not None:
                 hidden_states = inputs_embeds
             else:
-                hidden_states = self.get_input_embeddings(input_ids)
+                hidden_states = self.embed_input_ids(input_ids)
             residual = None
         else:
             assert intermediate_tensors is not None
@@ -557,8 +557,8 @@ def _init_model(
             vllm_config=vllm_config, prefix=prefix, layer_type=layer_type
         )
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
-        return self.model.get_input_embeddings(input_ids)
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
+        return self.model.embed_input_ids(input_ids)
 
     def forward(
         self,

vllm/model_executor/models/arcee.py

@@ -239,7 +239,7 @@ def __init__(
             ["hidden_states", "residual"], config.hidden_size
         )
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
         return self.embed_tokens(input_ids)
 
     def forward(
@@ -254,7 +254,7 @@ def forward(
             hidden_states = (
                 inputs_embeds
                 if inputs_embeds is not None
-                else self.get_input_embeddings(input_ids)
+                else self.embed_input_ids(input_ids)
             )
             residual = None
         else:
@@ -423,8 +423,8 @@ def compute_logits(self, hidden_states: torch.Tensor) -> torch.Tensor | None:
         logits = self.logits_processor(self.lm_head, hidden_states)
         return logits
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
-        return self.model.get_input_embeddings(input_ids)
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
+        return self.model.embed_input_ids(input_ids)
 
     def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
         """Load weights into the model (delegates to inner model and handles

vllm/model_executor/models/arctic.py

@@ -442,7 +442,7 @@ def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
             ["hidden_states"], config.hidden_size
         )
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
         return self.embed_tokens(input_ids)
 
     def forward(
@@ -456,7 +456,7 @@ def forward(
             if inputs_embeds is not None:
                 hidden_states = inputs_embeds
             else:
-                hidden_states = self.get_input_embeddings(input_ids)
+                hidden_states = self.embed_input_ids(input_ids)
         else:
             assert intermediate_tensors is not None
             hidden_states = intermediate_tensors["hidden_states"]
@@ -496,8 +496,8 @@ def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
             self.model.make_empty_intermediate_tensors
         )
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
-        return self.model.get_input_embeddings(input_ids)
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
+        return self.model.embed_input_ids(input_ids)
 
     def forward(
         self,

vllm/model_executor/models/aria.py

@@ -613,7 +613,7 @@ def _process_image_input(
     def get_language_model(self) -> torch.nn.Module:
         return self.language_model
 
-    def get_multimodal_embeddings(self, **kwargs: object) -> MultiModalEmbeddings:
+    def embed_multimodal(self, **kwargs: object) -> MultiModalEmbeddings:
         image_input = self._parse_and_validate_image_input(**kwargs)
         if image_input is None:
             return []
@@ -629,8 +629,8 @@ def forward(
         **kwargs: object,
     ) -> torch.Tensor | IntermediateTensors:
         if inputs_embeds is None:
-            multimodal_embeddings = self.get_multimodal_embeddings(**kwargs)
-            inputs_embeds = self.get_input_embeddings(
+            multimodal_embeddings = self.embed_multimodal(**kwargs)
+            inputs_embeds = self.embed_input_ids(
                 input_ids,
                 multimodal_embeddings,
                 is_multimodal=input_ids == self.config.image_token_index,

vllm/model_executor/models/aya_vision.py

@@ -417,7 +417,7 @@ def _parse_and_validate_image_input(
     def get_language_model(self) -> torch.nn.Module:
         return self.language_model
 
-    def get_multimodal_embeddings(self, **kwargs: object) -> MultiModalEmbeddings:
+    def embed_multimodal(self, **kwargs: object) -> MultiModalEmbeddings:
         image_input = self._parse_and_validate_image_input(**kwargs)
         if image_input is None:
             return []

vllm/model_executor/models/baichuan.py

@@ -309,7 +309,7 @@ def __init__(
             ["hidden_states", "residual"], config.hidden_size
         )
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
         return self.embed_tokens(input_ids)
 
     def forward(
@@ -323,7 +323,7 @@ def forward(
             if inputs_embeds is not None:
                 hidden_states = inputs_embeds
             else:
-                hidden_states = self.get_input_embeddings(input_ids)
+                hidden_states = self.embed_input_ids(input_ids)
             residual = None
         else:
             assert intermediate_tensors is not None
@@ -426,8 +426,8 @@ def __init__(
             self.model.make_empty_intermediate_tensors
         )
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
-        return self.model.get_input_embeddings(input_ids)
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
+        return self.model.embed_input_ids(input_ids)
 
     def forward(
         self,

vllm/model_executor/models/bailing_moe.py

@@ -438,7 +438,7 @@ def __init__(
         else:
             self.norm = PPMissingLayer()
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
         return self.word_embeddings(input_ids)
 
     def forward(
@@ -452,7 +452,7 @@ def forward(
             if inputs_embeds is not None:
                 hidden_states = inputs_embeds
             else:
-                hidden_states = self.get_input_embeddings(input_ids)
+                hidden_states = self.embed_input_ids(input_ids)
             residual = None
         else:
             assert intermediate_tensors is not None
@@ -608,8 +608,8 @@ def __init__(
             self.model.make_empty_intermediate_tensors
         )
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
-        return self.model.get_input_embeddings(input_ids)
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
+        return self.model.embed_input_ids(input_ids)
 
     def forward(
         self,

vllm/model_executor/models/bamba.py

@@ -314,7 +314,7 @@ def get_layer(prefix: str):
 
         self.final_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
         return self.embed_tokens(input_ids)
 
     def forward(
@@ -328,7 +328,7 @@ def forward(
             if inputs_embeds is not None:
                 hidden_states = inputs_embeds
             else:
-                hidden_states = self.get_input_embeddings(input_ids)
+                hidden_states = self.embed_input_ids(input_ids)
             residual = None
         else:
             assert intermediate_tensors is not None
@@ -493,8 +493,8 @@ def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
             self.model.make_empty_intermediate_tensors
         )
 
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
-        return self.model.get_input_embeddings(input_ids)
+    def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
+        return self.model.embed_input_ids(input_ids)
 
     def forward(
         self,