feat-rename-vector-type-0622

modeling_jina_embeddings_v4.py

@@ -31,7 +31,6 @@ class PromptType(str, Enum):
 
 
 PREFIX_DICT = {"query": "Query", "passage": "Passage"}
-VECTOR_TYPES = ["single_vector", "multi_vector"]
 
 
 class JinaEmbeddingsV4Processor(Qwen2_5_VLProcessor):
@@ -284,8 +283,9 @@ 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).
+                vlm_last_hidden_states (torch.Tensor, optional): Last hidden states of the VLM.
+                single_vec_emb (torch.Tensor, optional): Single-vector embeddings.
+                multi_vec_emb (torch.Tensor, optional): Multi-vector embeddings.
         """
         # 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",
+        return_multivector: bool = False,
         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 not return_multivector:
                         embeddings = embeddings.single_vec_emb
                         if truncate_dim is not None:
                             embeddings = embeddings[:, :truncate_dim]
@@ -357,7 +357,6 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
 
     def _validate_encoding_params(
         self,
-        vector_type: Optional[str] = None,
         truncate_dim: Optional[int] = None,
         prompt_name: Optional[str] = None,
     ) -> Dict[str, Any]:
@@ -374,14 +373,6 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
                     else PREFIX_DICT["query"]
                 )
 
-        vector_type = vector_type or "single_vector"
-        if vector_type not in VECTOR_TYPES:
-            raise ValueError(
-                f"Invalid vector_type: {vector_type}. Must be one of {VECTOR_TYPES}."
-            )
-        else:
-            encode_kwargs["vector_type"] = vector_type
-
         truncate_dim = truncate_dim or self.config.truncate_dim
         if truncate_dim is not None and truncate_dim not in self.config.matryoshka_dims:
             raise ValueError(
@@ -407,36 +398,34 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
                 )
         return task
 
-    def encode_texts(
+    def encode_text(
         self,
-        texts: List[str],
+        texts: Union[str, List[str]],
         task: Optional[str] = None,
         max_length: int = 8192,
         batch_size: int = 8,
-        vector_type: Optional[str] = None,
+        return_multivector: bool = False,
         return_numpy: bool = False,
         truncate_dim: Optional[int] = None,
         prompt_name: Optional[str] = None,
-    ) -> List[torch.Tensor]:
+    ) -> Union[List[torch.Tensor], torch.Tensor]:
         """
         Encodes a list of texts into embeddings.
 
         Args:
-            texts: List of text strings to encode
+            texts: text or 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')
+            return_multivector: Whether to return multi-vector embeddings instead of single-vector embeddings
             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')
 
         Returns:
-            List of text embeddings as tensors or numpy arrays
+            List of text embeddings as tensors or numpy arrays when encoding multiple texts, or single text embedding as tensor when encoding a single text
         """
         prompt_name = prompt_name or "query"
-        encode_kwargs = self._validate_encoding_params(
-            vector_type, truncate_dim, prompt_name
-        )
+        encode_kwargs = self._validate_encoding_params(truncate_dim=truncate_dim, prompt_name=prompt_name)
 
         task = self._validate_task(task)
 
@@ -446,17 +435,23 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
             prefix=encode_kwargs.pop("prefix"),
         )
 
+        return_list = isinstance(texts, list)
+        
+        if isinstance(texts, str):
+            texts = [texts]
+
         embeddings = self._process_batches(
             data=texts,
             processor_fn=processor_fn,
             desc="Encoding texts...",
             task_label=task,
+            return_multivector=return_multivector,
             return_numpy=return_numpy,
             batch_size=batch_size,
             **encode_kwargs,
         )
 
-        return embeddings
+        return embeddings if return_list else embeddings[0] 
 
     def _load_images_if_needed(
         self, images: List[Union[str, Image.Image]]
@@ -472,37 +467,44 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
             loaded_images.append(image)
         return loaded_images
 
-    def encode_images(
+    def encode_image(
         self,
-        images: List[Union[str, Image.Image]],
+        images: Union[str, Image.Image, List[Union[str, Image.Image]]],
         task: Optional[str] = None,
         batch_size: int = 8,
-        vector_type: Optional[str] = None,
+        return_multivector: bool = False,
         return_numpy: bool = False,
         truncate_dim: Optional[int] = None,
         max_pixels: Optional[int] = None,
-    ) -> List[torch.Tensor]:
+    ) -> Union[List[torch.Tensor], torch.Tensor]:
         """
-        Encodes a list of images into embeddings.
+        Encodes a list of images or a single image into embedding(s).
 
         Args:
-            images: List of PIL images, URLs, or local file paths to encode
+            images: image(s) to encode, can be PIL Image(s), URL(s), or local file path(s)
             batch_size: Number of images to process at once
-            vector_type: Type of embedding vector to generate ('single_vector' or 'multi_vector')
+            return_multivector: Whether to return multi-vector embeddings instead of single-vector embeddings
             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
 
         Returns:
-            List of image embeddings as tensors or numpy arrays
+            List of image embeddings as tensors or numpy arrays when encoding multiple images, or single image embedding as tensor when encoding a single image
         """
         if max_pixels:
             default_max_pixels = self.processor.image_processor.max_pixels
             self.processor.image_processor.max_pixels = (
                 max_pixels  # change during encoding
             )
-        encode_kwargs = self._validate_encoding_params(vector_type, truncate_dim)
+        encode_kwargs = self._validate_encoding_params(truncate_dim=truncate_dim)
         task = self._validate_task(task)
+        
+        return_list = isinstance(images, list)
+
+        # Convert single image to list
+        if isinstance(images, (str, Image.Image)):
+            images = [images]
+            
         images = self._load_images_if_needed(images)
         embeddings = self._process_batches(
             data=images,
@@ -510,6 +512,7 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
             desc="Encoding images...",
             task_label=task,
             batch_size=batch_size,
+            return_multivector=return_multivector,
             return_numpy=return_numpy,
             **encode_kwargs,
         )
@@ -517,7 +520,7 @@ class JinaEmbeddingsV4Model(Qwen2_5_VLForConditionalGeneration):
         if max_pixels:
             self.processor.image_processor.max_pixels = default_max_pixels
 
-        return embeddings
+        return embeddings if return_list else embeddings[0]
 
     @classmethod
     def from_pretrained(