feat: avoid the redundant words in the variables

modeling_jina_embeddings_v4.py

@@ -31,7 +31,7 @@ class PromptType(str, Enum):
 
 
 PREFIX_DICT = {"query": "Query", "passage": "Passage"}
-VECTOR_TYPES = ["single_vector", "multi_vector"]
+VECTOR_TYPES = ["single", "multi"]
 
 
 class JinaEmbeddingsV4Processor(Qwen2_5_VLProcessor):
@@ -284,8 +284,8 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
             attention_mask (torch.Tensor): The attention mask tensor.
         Returns:
             JinaEmbeddingsV4ModelOutput:
-                single_vector (torch.Tensor): Single-vector embeddings of shape (batch_size, dim).
-                multi_vector (torch.Tensor): Multi-vector embeddings of shape (batch_size, num_tokens, dim).
+                single (torch.Tensor): Single-vector embeddings of shape (batch_size, dim).
+                multi (torch.Tensor): Multi-vector embeddings of shape (batch_size, num_tokens, dim).
         """
         # Forward pass through the VLM
         hidden_states = self.get_last_hidden_states(
@@ -320,7 +320,7 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
         task_label: Union[str, List[str]],
         processor_fn: Callable,
         desc: str,
-        vector_type: str = "single_vector",
+        vector_type: str = "single",
         return_numpy: bool = False,
         batch_size: int = 32,
         truncate_dim: Optional[int] = None,
@@ -340,7 +340,7 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
                     device_type=torch.device(self.device).type, dtype=torch.bfloat16
                 ):
                     embeddings = self(**batch, task_label=task_label)
-                    if vector_type == "single_vector":
+                    if vector_type == "single":
                         embeddings = embeddings.single_vec_emb
                         if truncate_dim is not None:
                             embeddings = embeddings[:, :truncate_dim]
@@ -374,7 +374,7 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
                     else PREFIX_DICT["query"]
                 )
 
-        vector_type = vector_type or "single_vector"
+        vector_type = vector_type or "single"
         if vector_type not in VECTOR_TYPES:
             raise ValueError(
                 f"Invalid vector_type: {vector_type}. Must be one of {VECTOR_TYPES}."
@@ -425,7 +425,7 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
             texts: List of text strings to encode
             max_length: Maximum token length for text processing
             batch_size: Number of texts to process at once
-            vector_type: Type of embedding vector to generate ('single_vector' or 'multi_vector')
+            vector_type: Type of embedding vector to generate ('single' or 'multi')
             return_numpy: Whether to return numpy arrays instead of torch tensors
             truncate_dim: Dimension to truncate embeddings to (128, 256, 512, or 1024)
             prompt_name: Type of text being encoded ('query' or 'passage')
@@ -488,7 +488,7 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
         Args:
             images: List of PIL images, URLs, or local file paths to encode
             batch_size: Number of images to process at once
-            vector_type: Type of embedding vector to generate ('single_vector' or 'multi_vector')
+            vector_type: Type of embedding vector to generate ('single' or 'multi')
             return_numpy: Whether to return numpy arrays instead of torch tensors
             truncate_dim: Dimension to truncate embeddings to (128, 256, 512, or 1024)
             max_pixels: Maximum number of pixels to process per image