vllm.model_executor.models.deepseek_vl2

Inference-only Deepseek-VL2 model compatible with HuggingFace weights.

DeepseekVL2ImageInputs module-attribute

DeepseekVL2ImageInputs = Union[
    DeepseekVL2ImagePixelInputs,
    DeepseekVL2VImageEmbeddingInputs,
]

_IMAGE_TOKEN module-attribute

_IMAGE_TOKEN = '<image>'

DeepseekVL2DummyInputsBuilder

Bases: BaseDummyInputsBuilder[DeepseekVL2ProcessingInfo]

Source code in vllm/model_executor/models/deepseek_vl2.py

class DeepseekVL2DummyInputsBuilder(
        BaseDummyInputsBuilder[DeepseekVL2ProcessingInfo]):

    def get_dummy_text(self, mm_counts: Mapping[str, int]) -> str:
        num_images = mm_counts.get("image", 0)

        processor = self.info.get_hf_processor()
        image_token = processor.image_token

        return image_token * num_images

    def get_dummy_mm_data(
        self,
        seq_len: int,
        mm_counts: Mapping[str, int],
    ) -> MultiModalDataDict:
        num_images = mm_counts.get("image", 0)

        max_image_size = self.info.get_image_size_with_most_features()

        return {
            "image":
            self._get_dummy_images(width=max_image_size.width,
                                   height=max_image_size.height,
                                   num_images=num_images)
        }

get_dummy_mm_data

get_dummy_mm_data(
    seq_len: int, mm_counts: Mapping[str, int]
) -> MultiModalDataDict

Source code in vllm/model_executor/models/deepseek_vl2.py

def get_dummy_mm_data(
    self,
    seq_len: int,
    mm_counts: Mapping[str, int],
) -> MultiModalDataDict:
    num_images = mm_counts.get("image", 0)

    max_image_size = self.info.get_image_size_with_most_features()

    return {
        "image":
        self._get_dummy_images(width=max_image_size.width,
                               height=max_image_size.height,
                               num_images=num_images)
    }

get_dummy_text

get_dummy_text(mm_counts: Mapping[str, int]) -> str

Source code in vllm/model_executor/models/deepseek_vl2.py

def get_dummy_text(self, mm_counts: Mapping[str, int]) -> str:
    num_images = mm_counts.get("image", 0)

    processor = self.info.get_hf_processor()
    image_token = processor.image_token

    return image_token * num_images

DeepseekVL2ImagePixelInputs

Bases: TypedDict

Source code in vllm/model_executor/models/deepseek_vl2.py

class DeepseekVL2ImagePixelInputs(TypedDict):
    type: Literal["pixel_values"]
    data: Union[torch.Tensor, list[torch.Tensor]]
    """
    Shape: `(batch_size * num_images, num_channels, height, width)`
    """
    images_spatial_crop: torch.Tensor
    """
    Shape: `(batch_size * num_images, 2)`
    """

data instance-attribute

data: Union[Tensor, list[Tensor]]

Shape: (batch_size * num_images, num_channels, height, width)

images_spatial_crop instance-attribute

images_spatial_crop: Tensor

Shape: (batch_size * num_images, 2)

type instance-attribute

type: Literal['pixel_values']

DeepseekVL2MultiModalProcessor

Bases: BaseMultiModalProcessor[DeepseekVL2ProcessingInfo]

Source code in vllm/model_executor/models/deepseek_vl2.py

class DeepseekVL2MultiModalProcessor(
        BaseMultiModalProcessor[DeepseekVL2ProcessingInfo]):

    def _call_hf_processor(
        self,
        prompt: str,
        mm_data: Mapping[str, object],
        mm_kwargs: Mapping[str, object],
        tok_kwargs: Mapping[str, object],
    ) -> BatchFeature:
        if mm_data:
            processed_outputs = self.info.ctx.call_hf_processor(
                self.info.get_hf_processor(**mm_kwargs),
                dict(prompt=prompt, **mm_data),
                dict(**mm_kwargs, **tok_kwargs),
            )
            pixel_values = processed_outputs["pixel_values"]
            # split pixel values into patches corresponding to each image
            images_spatial_crop = processed_outputs["images_spatial_crop"]
            patches_per_image = [
                x.prod().item() + 1 for x in images_spatial_crop
            ]
            pixel_values = pixel_values.split(patches_per_image)
            processed_outputs["pixel_values"] = pixel_values
        else:
            tokenizer = self.info.get_tokenizer()
            processed_outputs = tokenizer(prompt,
                                          add_special_tokens=True,
                                          return_tensors="pt")

        return processed_outputs

    def _get_mm_fields_config(
        self,
        hf_inputs: BatchFeature,
        hf_processor_mm_kwargs: Mapping[str, object],
    ) -> Mapping[str, MultiModalFieldConfig]:
        return dict(
            pixel_values=MultiModalFieldConfig.batched("image"),
            images_spatial_crop=MultiModalFieldConfig.batched("image"),
            image_embeds=MultiModalFieldConfig.batched("image"),
        )

    def _get_prompt_updates(
        self,
        mm_items: MultiModalDataItems,
        hf_processor_mm_kwargs: Mapping[str, object],
        out_mm_kwargs: MultiModalKwargs,
    ) -> Sequence[PromptUpdate]:
        hf_processor = self.info.get_hf_processor(**hf_processor_mm_kwargs)

        image_token_id = hf_processor.image_token_id
        assert isinstance(image_token_id, int)

        def get_replacement_deepseek_vl2(item_idx: int):
            images = mm_items.get_items(
                "image", (ImageEmbeddingItems, ImageProcessorItems))

            if isinstance(images, ImageEmbeddingItems):
                num_image_tokens = images.get_feature_size(item_idx)
            else:
                image_size = images.get_image_size(item_idx)

                num_image_tokens = self.info.get_num_image_tokens(
                    image_width=image_size.width,
                    image_height=image_size.height,
                    cropping=len(images) <= 2,
                )
            return [image_token_id] * num_image_tokens

        return [
            PromptReplacement(
                modality="image",
                target=[image_token_id],
                replacement=get_replacement_deepseek_vl2,
            )
        ]

    def _cached_apply_hf_processor(
        self,
        prompt: Union[str, list[int]],
        mm_data_items: MultiModalDataItems,
        hf_processor_mm_kwargs: Mapping[str, object],
        tokenization_kwargs: Mapping[str, object],
        *,
        return_mm_hashes: bool,
    ) -> tuple[list[int], MultiModalKwargs, Optional[MultiModalHashes], bool]:
        # The processor logic is different for len(images) <= 2 vs > 2
        # Since the processing cache assumes that the processor output is
        # invariant of how many images are passed per prompt, we only
        # perform caching for the most common case
        if mm_data_items.get_count("image", strict=False) > 2:
            return self._apply_hf_processor(
                prompt=prompt,
                mm_data_items=mm_data_items,
                hf_processor_mm_kwargs=hf_processor_mm_kwargs,
                tokenization_kwargs=tokenization_kwargs,
                return_mm_hashes=return_mm_hashes,
            )

        return super()._cached_apply_hf_processor(
            prompt=prompt,
            mm_data_items=mm_data_items,
            hf_processor_mm_kwargs=hf_processor_mm_kwargs,
            tokenization_kwargs=tokenization_kwargs,
            return_mm_hashes=return_mm_hashes,
        )

_cached_apply_hf_processor

_cached_apply_hf_processor(
    prompt: Union[str, list[int]],
    mm_data_items: MultiModalDataItems,
    hf_processor_mm_kwargs: Mapping[str, object],
    tokenization_kwargs: Mapping[str, object],
    *,
    return_mm_hashes: bool,
) -> tuple[
    list[int],
    MultiModalKwargs,
    Optional[MultiModalHashes],
    bool,
]

Source code in vllm/model_executor/models/deepseek_vl2.py

def _cached_apply_hf_processor(
    self,
    prompt: Union[str, list[int]],
    mm_data_items: MultiModalDataItems,
    hf_processor_mm_kwargs: Mapping[str, object],
    tokenization_kwargs: Mapping[str, object],
    *,
    return_mm_hashes: bool,
) -> tuple[list[int], MultiModalKwargs, Optional[MultiModalHashes], bool]:
    # The processor logic is different for len(images) <= 2 vs > 2
    # Since the processing cache assumes that the processor output is
    # invariant of how many images are passed per prompt, we only
    # perform caching for the most common case
    if mm_data_items.get_count("image", strict=False) > 2:
        return self._apply_hf_processor(
            prompt=prompt,
            mm_data_items=mm_data_items,
            hf_processor_mm_kwargs=hf_processor_mm_kwargs,
            tokenization_kwargs=tokenization_kwargs,
            return_mm_hashes=return_mm_hashes,
        )

    return super()._cached_apply_hf_processor(
        prompt=prompt,
        mm_data_items=mm_data_items,
        hf_processor_mm_kwargs=hf_processor_mm_kwargs,
        tokenization_kwargs=tokenization_kwargs,
        return_mm_hashes=return_mm_hashes,
    )

_call_hf_processor

_call_hf_processor(
    prompt: str,
    mm_data: Mapping[str, object],
    mm_kwargs: Mapping[str, object],
    tok_kwargs: Mapping[str, object],
) -> BatchFeature

Source code in vllm/model_executor/models/deepseek_vl2.py

def _call_hf_processor(
    self,
    prompt: str,
    mm_data: Mapping[str, object],
    mm_kwargs: Mapping[str, object],
    tok_kwargs: Mapping[str, object],
) -> BatchFeature:
    if mm_data:
        processed_outputs = self.info.ctx.call_hf_processor(
            self.info.get_hf_processor(**mm_kwargs),
            dict(prompt=prompt, **mm_data),
            dict(**mm_kwargs, **tok_kwargs),
        )
        pixel_values = processed_outputs["pixel_values"]
        # split pixel values into patches corresponding to each image
        images_spatial_crop = processed_outputs["images_spatial_crop"]
        patches_per_image = [
            x.prod().item() + 1 for x in images_spatial_crop
        ]
        pixel_values = pixel_values.split(patches_per_image)
        processed_outputs["pixel_values"] = pixel_values
    else:
        tokenizer = self.info.get_tokenizer()
        processed_outputs = tokenizer(prompt,
                                      add_special_tokens=True,
                                      return_tensors="pt")

    return processed_outputs

_get_mm_fields_config

_get_mm_fields_config(
    hf_inputs: BatchFeature,
    hf_processor_mm_kwargs: Mapping[str, object],
) -> Mapping[str, MultiModalFieldConfig]

Source code in vllm/model_executor/models/deepseek_vl2.py

def _get_mm_fields_config(
    self,
    hf_inputs: BatchFeature,
    hf_processor_mm_kwargs: Mapping[str, object],
) -> Mapping[str, MultiModalFieldConfig]:
    return dict(
        pixel_values=MultiModalFieldConfig.batched("image"),
        images_spatial_crop=MultiModalFieldConfig.batched("image"),
        image_embeds=MultiModalFieldConfig.batched("image"),
    )

_get_prompt_updates

_get_prompt_updates(
    mm_items: MultiModalDataItems,
    hf_processor_mm_kwargs: Mapping[str, object],
    out_mm_kwargs: MultiModalKwargs,
) -> Sequence[PromptUpdate]

Source code in vllm/model_executor/models/deepseek_vl2.py

def _get_prompt_updates(
    self,
    mm_items: MultiModalDataItems,
    hf_processor_mm_kwargs: Mapping[str, object],
    out_mm_kwargs: MultiModalKwargs,
) -> Sequence[PromptUpdate]:
    hf_processor = self.info.get_hf_processor(**hf_processor_mm_kwargs)

    image_token_id = hf_processor.image_token_id
    assert isinstance(image_token_id, int)

    def get_replacement_deepseek_vl2(item_idx: int):
        images = mm_items.get_items(
            "image", (ImageEmbeddingItems, ImageProcessorItems))

        if isinstance(images, ImageEmbeddingItems):
            num_image_tokens = images.get_feature_size(item_idx)
        else:
            image_size = images.get_image_size(item_idx)

            num_image_tokens = self.info.get_num_image_tokens(
                image_width=image_size.width,
                image_height=image_size.height,
                cropping=len(images) <= 2,
            )
        return [image_token_id] * num_image_tokens

    return [
        PromptReplacement(
            modality="image",
            target=[image_token_id],
            replacement=get_replacement_deepseek_vl2,
        )
    ]

DeepseekVL2ProcessingInfo

Bases: BaseProcessingInfo

Source code in vllm/model_executor/models/deepseek_vl2.py

class DeepseekVL2ProcessingInfo(BaseProcessingInfo):

    def get_hf_config(self):
        return self.ctx.get_hf_config(DeepseekVLV2Config)

    def get_hf_processor(self, **kwargs: object):
        return self.ctx.get_hf_processor(DeepseekVLV2Processor, **kwargs)

    def get_supported_mm_limits(self) -> Mapping[str, Optional[int]]:
        return {"image": None}

    def get_num_image_tokens(self,
                             *,
                             image_width: int,
                             image_height: int,
                             cropping: bool = True) -> int:
        hf_processor = self.get_hf_processor()
        image_size = hf_processor.image_size
        patch_size = hf_processor.patch_size
        downsample_ratio = hf_processor.downsample_ratio

        if cropping:
            best_width, best_height = hf_processor.select_best_resolution(
                (image_width, image_height))
            num_width_tiles, num_height_tiles = (best_width // image_size,
                                                 best_height // image_size)
        else:
            num_width_tiles = num_height_tiles = 1

        h = w = math.ceil((image_size // patch_size) / downsample_ratio)

        global_views_tokens = h * (w + 1)
        local_views_tokens = (num_height_tiles * h) * (num_width_tiles * w + 1)
        return global_views_tokens + local_views_tokens + 1

    def get_image_size_with_most_features(self) -> ImageSize:
        hf_config = self.get_hf_config()
        candidate_resolutions = hf_config.candidate_resolutions
        height, width = max(candidate_resolutions,
                            key=lambda x: self.get_num_image_tokens(
                                image_width=x[1], image_height=x[0]))
        return ImageSize(width=width, height=height)

get_hf_config

get_hf_config()

Source code in vllm/model_executor/models/deepseek_vl2.py

def get_hf_config(self):
    return self.ctx.get_hf_config(DeepseekVLV2Config)

get_hf_processor

get_hf_processor(**kwargs: object)

Source code in vllm/model_executor/models/deepseek_vl2.py

def get_hf_processor(self, **kwargs: object):
    return self.ctx.get_hf_processor(DeepseekVLV2Processor, **kwargs)

get_image_size_with_most_features

get_image_size_with_most_features() -> ImageSize

Source code in vllm/model_executor/models/deepseek_vl2.py

def get_image_size_with_most_features(self) -> ImageSize:
    hf_config = self.get_hf_config()
    candidate_resolutions = hf_config.candidate_resolutions
    height, width = max(candidate_resolutions,
                        key=lambda x: self.get_num_image_tokens(
                            image_width=x[1], image_height=x[0]))
    return ImageSize(width=width, height=height)

get_num_image_tokens

get_num_image_tokens(
    *,
    image_width: int,
    image_height: int,
    cropping: bool = True,
) -> int

Source code in vllm/model_executor/models/deepseek_vl2.py

def get_num_image_tokens(self,
                         *,
                         image_width: int,
                         image_height: int,
                         cropping: bool = True) -> int:
    hf_processor = self.get_hf_processor()
    image_size = hf_processor.image_size
    patch_size = hf_processor.patch_size
    downsample_ratio = hf_processor.downsample_ratio

    if cropping:
        best_width, best_height = hf_processor.select_best_resolution(
            (image_width, image_height))
        num_width_tiles, num_height_tiles = (best_width // image_size,
                                             best_height // image_size)
    else:
        num_width_tiles = num_height_tiles = 1

    h = w = math.ceil((image_size // patch_size) / downsample_ratio)

    global_views_tokens = h * (w + 1)
    local_views_tokens = (num_height_tiles * h) * (num_width_tiles * w + 1)
    return global_views_tokens + local_views_tokens + 1

get_supported_mm_limits

get_supported_mm_limits() -> Mapping[str, Optional[int]]

Source code in vllm/model_executor/models/deepseek_vl2.py

def get_supported_mm_limits(self) -> Mapping[str, Optional[int]]:
    return {"image": None}

DeepseekVL2VImageEmbeddingInputs

Bases: TypedDict

Source code in vllm/model_executor/models/deepseek_vl2.py

class DeepseekVL2VImageEmbeddingInputs(TypedDict):
    type: Literal["image_embeds"]
    data: Union[torch.Tensor, list[torch.Tensor]]
    """Shape: `(batch_size * num_images, image_feature_size, hidden_size)`

    `hidden_size` must match the hidden size of language model backbone.
    """

data instance-attribute

data: Union[Tensor, list[Tensor]]

Shape: (batch_size * num_images, image_feature_size, hidden_size)

hidden_size must match the hidden size of language model backbone.

type instance-attribute

type: Literal['image_embeds']

DeepseekVLV2ForCausalLM

Bases: Module, SupportsMultiModal, SupportsPP

Source code in vllm/model_executor/models/deepseek_vl2.py

@MULTIMODAL_REGISTRY.register_processor(
    DeepseekVL2MultiModalProcessor,
    info=DeepseekVL2ProcessingInfo,
    dummy_inputs=DeepseekVL2DummyInputsBuilder)
class DeepseekVLV2ForCausalLM(nn.Module, SupportsMultiModal, SupportsPP):

    hf_to_vllm_mapper = WeightsMapper(orig_to_new_prefix={
        "language.": "language_model.",
    })

    @classmethod
    def get_placeholder_str(cls, modality: str, i: int) -> Optional[str]:
        if modality.startswith("image"):
            return "<image>"

        raise ValueError("Only image modality is supported")

    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
        super().__init__()
        config: DeepseekVLV2Config = vllm_config.model_config.hf_config
        quant_config = vllm_config.quant_config
        multimodal_config = vllm_config.model_config.multimodal_config

        self.config = config
        self.multimodal_config = multimodal_config

        self.vision_config = config.vision_config
        self.projector_config = config.projector_config
        self.text_config = config.text_config

        model_config = vllm_config.model_config
        tokenizer = cached_tokenizer_from_config(model_config)
        self.image_token_id = tokenizer.vocab[_IMAGE_TOKEN]

        self.vision = self._init_vision_module(self.vision_config,
                                               quant_config,
                                               maybe_prefix(prefix, "vision"))

        self.projector = MlpProjector(self.projector_config)
        self.tile_tag = config.tile_tag
        self.global_view_pos = config.global_view_pos

        # special token for image token sequence format
        embed_std = 1 / torch.sqrt(
            torch.tensor(self.projector_config.n_embed, dtype=torch.float32))
        if self.tile_tag == "2D":
            # <|view_separator|>, <|\n|>
            self.image_newline = nn.Parameter(
                torch.randn(self.projector_config.n_embed) * embed_std)
            # This is a typo in original implementation
            self.view_separator = nn.Parameter(
                torch.randn(self.projector_config.n_embed) * embed_std)
        else:
            raise ValueError(
                f"Only 2D tile_tag is supported currently, got: {self.tile_tag}"
            )

        if self.text_config.topk_method == "noaux_tc":
            architectures = ["DeepseekV3ForCausalLM"]
        elif not self.text_config.use_mla:
            architectures = ["DeepseekForCausalLM"]
        else:
            architectures = ["DeepseekV2ForCausalLM"]

        self.language_model = init_vllm_registered_model(
            vllm_config=vllm_config,
            hf_config=self.text_config,
            prefix=maybe_prefix(prefix, "language"),
            architectures=architectures,
        )

        self.make_empty_intermediate_tensors = (
            self.language_model.make_empty_intermediate_tensors)

    def _init_vision_module(
        self,
        vision_config: VisionEncoderConfig,
        quant_config: Optional[QuantizationConfig],
        prefix: str = "",
    ) -> nn.Module:
        # TODO: refactor vision model through timm wrapper from transformers
        try:
            import timm
        except ImportError:
            raise ImportError("Please install timm") from ImportError

        with set_default_torch_dtype(torch.float16):
            model = timm.create_model(
                "vit_so400m_patch14_siglip_384.webli",
                pretrained=False,
                num_classes=0,
                dynamic_img_size=True,
                dynamic_img_pad=True,
            )

        model = model.to(dtype=torch.get_default_dtype())
        return model

    def _validate_pixel_values(
        self, data: Union[torch.Tensor, list[torch.Tensor]]
    ) -> Union[torch.Tensor, list[torch.Tensor]]:

        h = w = self.vision_config.image_size
        expected_dims = (3, h, w)

        def _validate_shape(d: torch.Tensor):
            actual_dims = tuple(d.shape[1:])

            if actual_dims != expected_dims:
                expected_expr = ("num_patches", *map(str, expected_dims))
                raise ValueError(
                    "The expected shape of pixel values per image per batch "
                    f"is {expected_expr}. You supplied {tuple(d.shape)}.")

        for d in data:
            _validate_shape(d)

        return data

    def _validate_images_spatial_crop(
        self, data: Union[torch.Tensor, list[torch.Tensor]]
    ) -> Union[torch.Tensor, list[torch.Tensor]]:
        expected_dims = 2

        def _validate_shape(d: torch.Tensor):
            actual_dims = d.size(-1)

            if actual_dims != expected_dims:
                expected_expr = str(expected_dims)
                raise ValueError(
                    f"The expected shape of image sizes per image per batch "
                    f"is {expected_expr}. You supplied {tuple(d.shape)}.")

        for d in data:
            _validate_shape(d)

        return data

    def _parse_and_validate_image_input(
            self, **kwargs: object) -> Optional[DeepseekVL2ImageInputs]:
        pixel_values = kwargs.pop("pixel_values", None)
        images_spatial_crop = kwargs.pop("images_spatial_crop", None)
        image_embeds = kwargs.pop("image_embeds", None)

        if pixel_values is None and image_embeds is None:
            return None

        if pixel_values is not None:
            if not isinstance(pixel_values, (torch.Tensor, list)):
                raise ValueError("Incorrect type of pixel values. "
                                 f"Got type: {type(pixel_values)}")

            if not isinstance(images_spatial_crop, (torch.Tensor, list)):
                raise ValueError("Incorrect type of image sizes. "
                                 f"Got type: {type(images_spatial_crop)}")

            return DeepseekVL2ImagePixelInputs(
                type="pixel_values",
                data=self._validate_pixel_values(flatten_bn(pixel_values)),
                images_spatial_crop=self._validate_images_spatial_crop(
                    flatten_bn(images_spatial_crop, concat=True)))

        if image_embeds is not None:
            if not isinstance(image_embeds, (torch.Tensor, list)):
                raise ValueError("Incorrect type of image embeddings. "
                                 f"Got type: {type(image_embeds)}")

            return DeepseekVL2VImageEmbeddingInputs(
                type="image_embeds",
                data=flatten_bn(image_embeds),
            )

        raise AssertionError("This line should be unreachable.")

    def _pixel_values_to_embedding(
        self,
        pixel_values: NestedTensors,
        images_spatial_crop: torch.Tensor,
    ) -> NestedTensors:
        # Pixel_values: n_image * batch_size * [patch_per_img, 3, height, width]
        total_tiles = [x for x in pixel_values]

        # [batch_all_tiles, 3, height, width]
        total_tiles = torch.cat(total_tiles, dim=0)

        # [batch_all_tiles, vit_seq_len, c]
        images_feature = self.vision.forward_features(total_tiles)

        # [batch_all_tiles, hw, D]
        images_embeds = self.projector(images_feature)

        _, hw, n_dim = images_embeds.shape
        h = w = int(hw**0.5)

        # fill image token based on self.tile_tag & self.global_view_pos
        tile_index = 0
        vision_embeddings = []
        for jdx in range(images_spatial_crop.size(0)):
            # extra global & local features
            num_width_tiles, num_height_tiles = images_spatial_crop[jdx]
            if num_width_tiles == 0 or num_height_tiles == 0:
                break
            num_tiles_in_image = num_width_tiles * num_height_tiles

            # [hw, D]
            global_features = images_embeds[tile_index]

            # [num_height_tiles * num_width_tiles, hw, D]
            local_features = images_embeds[tile_index + 1:tile_index + 1 +
                                           num_tiles_in_image]
            tile_index += num_tiles_in_image + 1

            # format global and local features
            # ----------------- global view add newline -----------------
            # [hw, D] -> [h, w, D]
            global_features = global_features.view(h, w, n_dim)

            # [D]     -> [h, 1, D]
            new_lines_in_global = repeat(self.image_newline, "d -> h 1 d", h=h)

            # cat([h, w, D], [h, 1, D], dim=1) -> [h, w + 1, D]
            global_features = torch.cat([global_features, new_lines_in_global],
                                        dim=1)

            # [h, w + 1, D] -> [h * (w + 1), D]
            global_features = global_features.view(-1, n_dim)

            # ----------------- local view add newline -----------------
            # [num_height_tiles * num_width_tiles, h * w, D] ->
            # [num_height_tiles * h, num_width_tiles * w, D]
            local_features = rearrange(local_features,
                                       "(th tw) (h w) d -> (th h) (tw w) d",
                                       th=num_height_tiles,
                                       tw=num_width_tiles,
                                       h=h,
                                       w=w)

            # [D] -> [num_height_tiles * h, 1, D]
            new_lines_in_local = repeat(self.image_newline,
                                        "d -> (th h) 1 d",
                                        th=num_height_tiles,
                                        h=h)

            # [num_height_tiles * h, num_width_tiles * w + 1, D]
            local_features = torch.cat([local_features, new_lines_in_local],
                                       dim=1)

            # [num_height_tiles * h, num_width_tiles * w + 1, D]
            #   --> [(num_height_tiles * h) * (num_width_tiles * w + 1), D]
            local_features = local_features.view(-1, n_dim)

            # merge global and local tiles
            if self.global_view_pos == "head":
                global_local_features = torch.cat([
                    global_features,
                    self.view_separator[None, :],
                    local_features,
                ])
            else:
                global_local_features = torch.cat([
                    local_features,
                    self.view_separator[None, :],
                    global_features,
                ])

            vision_embeddings.append(global_local_features)
        return vision_embeddings

    def _process_image_input(
            self, image_input: DeepseekVL2ImageInputs) -> torch.Tensor:
        if image_input["type"] == "image_embeds":
            image_data = image_input["data"]
            if is_list_of(image_data, torch.Tensor):
                # it's already a list of tensors
                return image_data
            if len(image_data.shape) == 3:
                # 3D tensor
                return list(torch.unbind(image_data, dim=0))
            raise ValueError(
                "We expect batched 2D tensors; "
                "this can be either a list of 2D tensors or a single 3D tensor."
            )

        pixel_values = image_input["data"]
        images_spatial_crop = image_input["images_spatial_crop"]

        return self._pixel_values_to_embedding(
            pixel_values=pixel_values, images_spatial_crop=images_spatial_crop)

    def get_language_model(self) -> torch.nn.Module:
        return self.language_model

    def get_multimodal_embeddings(self,
                                  **kwargs: object) -> MultiModalEmbeddings:
        image_input = self._parse_and_validate_image_input(**kwargs)
        if image_input is None:
            return []
        vision_embeddings = self._process_image_input(image_input)
        return vision_embeddings

    def get_input_embeddings(
        self,
        input_ids: torch.Tensor,
        multimodal_embeddings: Optional[MultiModalEmbeddings] = None,
    ) -> torch.Tensor:
        inputs_embeds = self.language_model.get_input_embeddings(input_ids)
        if multimodal_embeddings is not None \
            and len(multimodal_embeddings) != 0:
            inputs_embeds = merge_multimodal_embeddings(
                input_ids, inputs_embeds, multimodal_embeddings,
                self.image_token_id)
        return inputs_embeds

    def forward(self,
                input_ids: torch.Tensor,
                positions: torch.Tensor,
                intermediate_tensors: Optional[IntermediateTensors] = None,
                inputs_embeds: Optional[torch.Tensor] = None,
                **kwargs: object):

        if intermediate_tensors is not None:
            inputs_embeds = None

        # NOTE: In v1, inputs_embeds is always generated at model runner, this
        # condition is for v0 compatibility
        elif inputs_embeds is None:
            vision_embeddings = self.get_multimodal_embeddings(**kwargs)
            inputs_embeds = self.get_input_embeddings(input_ids,
                                                      vision_embeddings)
            input_ids = None

        hidden_states = self.language_model(input_ids,
                                            positions,
                                            intermediate_tensors,
                                            inputs_embeds=inputs_embeds)

        return hidden_states

    def compute_logits(
        self,
        hidden_states: torch.Tensor,
        sampling_metadata: SamplingMetadata,
    ) -> Optional[torch.Tensor]:
        return self.language_model.compute_logits(hidden_states,
                                                  sampling_metadata)

    def load_weights(self, weights: Iterable[tuple[str,
                                                   torch.Tensor]]) -> set[str]:

        loader = AutoWeightsLoader(self)
        autoloaded_weights = loader.load_weights(weights,
                                                 mapper=self.hf_to_vllm_mapper)
        return autoloaded_weights

config instance-attribute

config = config

global_view_pos instance-attribute

global_view_pos = global_view_pos

hf_to_vllm_mapper class-attribute instance-attribute

hf_to_vllm_mapper = WeightsMapper(
    orig_to_new_prefix={"language.": "language_model."}
)

image_newline instance-attribute

image_newline = Parameter(randn(n_embed) * embed_std)

image_token_id instance-attribute

image_token_id = vocab[_IMAGE_TOKEN]

language_model instance-attribute

language_model = init_vllm_registered_model(
    vllm_config=vllm_config,
    hf_config=text_config,
    prefix=maybe_prefix(prefix, "language"),
    architectures=architectures,
)

make_empty_intermediate_tensors instance-attribute

make_empty_intermediate_tensors = (
    make_empty_intermediate_tensors
)

multimodal_config instance-attribute

multimodal_config = multimodal_config

projector instance-attribute

projector = MlpProjector(projector_config)

projector_config instance-attribute

projector_config = projector_config

text_config instance-attribute

text_config = text_config

tile_tag instance-attribute

tile_tag = tile_tag

view_separator instance-attribute

view_separator = Parameter(randn(n_embed) * embed_std)

vision instance-attribute

vision = _init_vision_module(
    vision_config,
    quant_config,
    maybe_prefix(prefix, "vision"),
)

vision_config instance-attribute

vision_config = vision_config

__init__

__init__(*, vllm_config: VllmConfig, prefix: str = '')

Source code in vllm/model_executor/models/deepseek_vl2.py

def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
    super().__init__()
    config: DeepseekVLV2Config = vllm_config.model_config.hf_config
    quant_config = vllm_config.quant_config
    multimodal_config = vllm_config.model_config.multimodal_config

    self.config = config
    self.multimodal_config = multimodal_config

    self.vision_config = config.vision_config
    self.projector_config = config.projector_config
    self.text_config = config.text_config

    model_config = vllm_config.model_config
    tokenizer = cached_tokenizer_from_config(model_config)
    self.image_token_id = tokenizer.vocab[_IMAGE_TOKEN]

    self.vision = self._init_vision_module(self.vision_config,
                                           quant_config,
                                           maybe_prefix(prefix, "vision"))

    self.projector = MlpProjector(self.projector_config)
    self.tile_tag = config.tile_tag
    self.global_view_pos = config.global_view_pos

    # special token for image token sequence format
    embed_std = 1 / torch.sqrt(
        torch.tensor(self.projector_config.n_embed, dtype=torch.float32))
    if self.tile_tag == "2D":
        # <|view_separator|>, <|\n|>
        self.image_newline = nn.Parameter(
            torch.randn(self.projector_config.n_embed) * embed_std)
        # This is a typo in original implementation
        self.view_separator = nn.Parameter(
            torch.randn(self.projector_config.n_embed) * embed_std)
    else:
        raise ValueError(
            f"Only 2D tile_tag is supported currently, got: {self.tile_tag}"
        )

    if self.text_config.topk_method == "noaux_tc":
        architectures = ["DeepseekV3ForCausalLM"]
    elif not self.text_config.use_mla:
        architectures = ["DeepseekForCausalLM"]
    else:
        architectures = ["DeepseekV2ForCausalLM"]

    self.language_model = init_vllm_registered_model(
        vllm_config=vllm_config,
        hf_config=self.text_config,
        prefix=maybe_prefix(prefix, "language"),
        architectures=architectures,
    )

    self.make_empty_intermediate_tensors = (
        self.language_model.make_empty_intermediate_tensors)

_init_vision_module

_init_vision_module(
    vision_config: VisionEncoderConfig,
    quant_config: Optional[QuantizationConfig],
    prefix: str = "",
) -> Module

Source code in vllm/model_executor/models/deepseek_vl2.py

def _init_vision_module(
    self,
    vision_config: VisionEncoderConfig,
    quant_config: Optional[QuantizationConfig],
    prefix: str = "",
) -> nn.Module:
    # TODO: refactor vision model through timm wrapper from transformers
    try:
        import timm
    except ImportError:
        raise ImportError("Please install timm") from ImportError

    with set_default_torch_dtype(torch.float16):
        model = timm.create_model(
            "vit_so400m_patch14_siglip_384.webli",
            pretrained=False,
            num_classes=0,
            dynamic_img_size=True,
            dynamic_img_pad=True,
        )

    model = model.to(dtype=torch.get_default_dtype())
    return model

_parse_and_validate_image_input

_parse_and_validate_image_input(
    **kwargs: object,
) -> Optional[DeepseekVL2ImageInputs]

Source code in vllm/model_executor/models/deepseek_vl2.py

def _parse_and_validate_image_input(
        self, **kwargs: object) -> Optional[DeepseekVL2ImageInputs]:
    pixel_values = kwargs.pop("pixel_values", None)
    images_spatial_crop = kwargs.pop("images_spatial_crop", None)
    image_embeds = kwargs.pop("image_embeds", None)

    if pixel_values is None and image_embeds is None:
        return None

    if pixel_values is not None:
        if not isinstance(pixel_values, (torch.Tensor, list)):
            raise ValueError("Incorrect type of pixel values. "
                             f"Got type: {type(pixel_values)}")

        if not isinstance(images_spatial_crop, (torch.Tensor, list)):
            raise ValueError("Incorrect type of image sizes. "
                             f"Got type: {type(images_spatial_crop)}")

        return DeepseekVL2ImagePixelInputs(
            type="pixel_values",
            data=self._validate_pixel_values(flatten_bn(pixel_values)),
            images_spatial_crop=self._validate_images_spatial_crop(
                flatten_bn(images_spatial_crop, concat=True)))

    if image_embeds is not None:
        if not isinstance(image_embeds, (torch.Tensor, list)):
            raise ValueError("Incorrect type of image embeddings. "
                             f"Got type: {type(image_embeds)}")

        return DeepseekVL2VImageEmbeddingInputs(
            type="image_embeds",
            data=flatten_bn(image_embeds),
        )

    raise AssertionError("This line should be unreachable.")

_pixel_values_to_embedding

_pixel_values_to_embedding(
    pixel_values: NestedTensors, images_spatial_crop: Tensor
) -> NestedTensors

Source code in vllm/model_executor/models/deepseek_vl2.py

def _pixel_values_to_embedding(
    self,
    pixel_values: NestedTensors,
    images_spatial_crop: torch.Tensor,
) -> NestedTensors:
    # Pixel_values: n_image * batch_size * [patch_per_img, 3, height, width]
    total_tiles = [x for x in pixel_values]

    # [batch_all_tiles, 3, height, width]
    total_tiles = torch.cat(total_tiles, dim=0)

    # [batch_all_tiles, vit_seq_len, c]
    images_feature = self.vision.forward_features(total_tiles)

    # [batch_all_tiles, hw, D]
    images_embeds = self.projector(images_feature)

    _, hw, n_dim = images_embeds.shape
    h = w = int(hw**0.5)

    # fill image token based on self.tile_tag & self.global_view_pos
    tile_index = 0
    vision_embeddings = []
    for jdx in range(images_spatial_crop.size(0)):
        # extra global & local features
        num_width_tiles, num_height_tiles = images_spatial_crop[jdx]
        if num_width_tiles == 0 or num_height_tiles == 0:
            break
        num_tiles_in_image = num_width_tiles * num_height_tiles

        # [hw, D]
        global_features = images_embeds[tile_index]

        # [num_height_tiles * num_width_tiles, hw, D]
        local_features = images_embeds[tile_index + 1:tile_index + 1 +
                                       num_tiles_in_image]
        tile_index += num_tiles_in_image + 1

        # format global and local features
        # ----------------- global view add newline -----------------
        # [hw, D] -> [h, w, D]
        global_features = global_features.view(h, w, n_dim)

        # [D]     -> [h, 1, D]
        new_lines_in_global = repeat(self.image_newline, "d -> h 1 d", h=h)

        # cat([h, w, D], [h, 1, D], dim=1) -> [h, w + 1, D]
        global_features = torch.cat([global_features, new_lines_in_global],
                                    dim=1)

        # [h, w + 1, D] -> [h * (w + 1), D]
        global_features = global_features.view(-1, n_dim)

        # ----------------- local view add newline -----------------
        # [num_height_tiles * num_width_tiles, h * w, D] ->
        # [num_height_tiles * h, num_width_tiles * w, D]
        local_features = rearrange(local_features,
                                   "(th tw) (h w) d -> (th h) (tw w) d",
                                   th=num_height_tiles,
                                   tw=num_width_tiles,
                                   h=h,
                                   w=w)

        # [D] -> [num_height_tiles * h, 1, D]
        new_lines_in_local = repeat(self.image_newline,
                                    "d -> (th h) 1 d",
                                    th=num_height_tiles,
                                    h=h)

        # [num_height_tiles * h, num_width_tiles * w + 1, D]
        local_features = torch.cat([local_features, new_lines_in_local],
                                   dim=1)

        # [num_height_tiles * h, num_width_tiles * w + 1, D]
        #   --> [(num_height_tiles * h) * (num_width_tiles * w + 1), D]
        local_features = local_features.view(-1, n_dim)

        # merge global and local tiles
        if self.global_view_pos == "head":
            global_local_features = torch.cat([
                global_features,
                self.view_separator[None, :],
                local_features,
            ])
        else:
            global_local_features = torch.cat([
                local_features,
                self.view_separator[None, :],
                global_features,
            ])

        vision_embeddings.append(global_local_features)
    return vision_embeddings

_process_image_input

_process_image_input(
    image_input: DeepseekVL2ImageInputs,
) -> Tensor

Source code in vllm/model_executor/models/deepseek_vl2.py

def _process_image_input(
        self, image_input: DeepseekVL2ImageInputs) -> torch.Tensor:
    if image_input["type"] == "image_embeds":
        image_data = image_input["data"]
        if is_list_of(image_data, torch.Tensor):
            # it's already a list of tensors
            return image_data
        if len(image_data.shape) == 3:
            # 3D tensor
            return list(torch.unbind(image_data, dim=0))
        raise ValueError(
            "We expect batched 2D tensors; "
            "this can be either a list of 2D tensors or a single 3D tensor."
        )

    pixel_values = image_input["data"]
    images_spatial_crop = image_input["images_spatial_crop"]

    return self._pixel_values_to_embedding(
        pixel_values=pixel_values, images_spatial_crop=images_spatial_crop)

_validate_images_spatial_crop

_validate_images_spatial_crop(
    data: Union[Tensor, list[Tensor]],
) -> Union[Tensor, list[Tensor]]

Source code in vllm/model_executor/models/deepseek_vl2.py

def _validate_images_spatial_crop(
    self, data: Union[torch.Tensor, list[torch.Tensor]]
) -> Union[torch.Tensor, list[torch.Tensor]]:
    expected_dims = 2

    def _validate_shape(d: torch.Tensor):
        actual_dims = d.size(-1)

        if actual_dims != expected_dims:
            expected_expr = str(expected_dims)
            raise ValueError(
                f"The expected shape of image sizes per image per batch "
                f"is {expected_expr}. You supplied {tuple(d.shape)}.")

    for d in data:
        _validate_shape(d)

    return data

_validate_pixel_values

_validate_pixel_values(
    data: Union[Tensor, list[Tensor]],
) -> Union[Tensor, list[Tensor]]

Source code in vllm/model_executor/models/deepseek_vl2.py

def _validate_pixel_values(
    self, data: Union[torch.Tensor, list[torch.Tensor]]
) -> Union[torch.Tensor, list[torch.Tensor]]:

    h = w = self.vision_config.image_size
    expected_dims = (3, h, w)

    def _validate_shape(d: torch.Tensor):
        actual_dims = tuple(d.shape[1:])

        if actual_dims != expected_dims:
            expected_expr = ("num_patches", *map(str, expected_dims))
            raise ValueError(
                "The expected shape of pixel values per image per batch "
                f"is {expected_expr}. You supplied {tuple(d.shape)}.")

    for d in data:
        _validate_shape(d)

    return data

compute_logits

compute_logits(
    hidden_states: Tensor,
    sampling_metadata: SamplingMetadata,
) -> Optional[Tensor]

Source code in vllm/model_executor/models/deepseek_vl2.py

def compute_logits(
    self,
    hidden_states: torch.Tensor,
    sampling_metadata: SamplingMetadata,
) -> Optional[torch.Tensor]:
    return self.language_model.compute_logits(hidden_states,
                                              sampling_metadata)

forward

forward(
    input_ids: Tensor,
    positions: Tensor,
    intermediate_tensors: Optional[
        IntermediateTensors
    ] = None,
    inputs_embeds: Optional[Tensor] = None,
    **kwargs: object,
)

Source code in vllm/model_executor/models/deepseek_vl2.py

def forward(self,
            input_ids: torch.Tensor,
            positions: torch.Tensor,
            intermediate_tensors: Optional[IntermediateTensors] = None,
            inputs_embeds: Optional[torch.Tensor] = None,
            **kwargs: object):

    if intermediate_tensors is not None:
        inputs_embeds = None

    # NOTE: In v1, inputs_embeds is always generated at model runner, this
    # condition is for v0 compatibility
    elif inputs_embeds is None:
        vision_embeddings = self.get_multimodal_embeddings(**kwargs)
        inputs_embeds = self.get_input_embeddings(input_ids,
                                                  vision_embeddings)
        input_ids = None

    hidden_states = self.language_model(input_ids,
                                        positions,
                                        intermediate_tensors,
                                        inputs_embeds=inputs_embeds)

    return hidden_states

get_input_embeddings

get_input_embeddings(
    input_ids: Tensor,
    multimodal_embeddings: Optional[
        MultiModalEmbeddings
    ] = None,
) -> Tensor

Source code in vllm/model_executor/models/deepseek_vl2.py

def get_input_embeddings(
    self,
    input_ids: torch.Tensor,
    multimodal_embeddings: Optional[MultiModalEmbeddings] = None,
) -> torch.Tensor:
    inputs_embeds = self.language_model.get_input_embeddings(input_ids)
    if multimodal_embeddings is not None \
        and len(multimodal_embeddings) != 0:
        inputs_embeds = merge_multimodal_embeddings(
            input_ids, inputs_embeds, multimodal_embeddings,
            self.image_token_id)
    return inputs_embeds

get_language_model

get_language_model() -> Module

Source code in vllm/model_executor/models/deepseek_vl2.py

def get_language_model(self) -> torch.nn.Module:
    return self.language_model

get_multimodal_embeddings

get_multimodal_embeddings(
    **kwargs: object,
) -> MultiModalEmbeddings

Source code in vllm/model_executor/models/deepseek_vl2.py

def get_multimodal_embeddings(self,
                              **kwargs: object) -> MultiModalEmbeddings:
    image_input = self._parse_and_validate_image_input(**kwargs)
    if image_input is None:
        return []
    vision_embeddings = self._process_image_input(image_input)
    return vision_embeddings

get_placeholder_str classmethod

get_placeholder_str(modality: str, i: int) -> Optional[str]

Source code in vllm/model_executor/models/deepseek_vl2.py

@classmethod
def get_placeholder_str(cls, modality: str, i: int) -> Optional[str]:
    if modality.startswith("image"):
        return "<image>"

    raise ValueError("Only image modality is supported")

load_weights

load_weights(
    weights: Iterable[tuple[str, Tensor]],
) -> set[str]

Source code in vllm/model_executor/models/deepseek_vl2.py

def load_weights(self, weights: Iterable[tuple[str,
                                               torch.Tensor]]) -> set[str]:

    loader = AutoWeightsLoader(self)
    autoloaded_weights = loader.load_weights(weights,
                                             mapper=self.hf_to_vllm_mapper)
    return autoloaded_weights

MlpProjector

Bases: Module

Source code in vllm/model_executor/models/deepseek_vl2.py

class MlpProjector(nn.Module):

    def __init__(self, cfg: MlpProjectorConfig):

        super().__init__()

        self.cfg = cfg
        assert not cfg.token_pooling, (
            "Token pooling is not supported currently.")

        if cfg.projector_type == "downsample_mlp_gelu":
            mlp_depth = cfg.depth
            mlp_ratio = cfg.mlp_ratio
            modules = [
                nn.Linear(
                    cfg.input_dim * cfg.downsample_ratio *
                    cfg.downsample_ratio, cfg.n_embed * mlp_ratio)
            ]
            for _ in range(1, mlp_depth - 1):
                modules.append(nn.GELU())
                modules.append(
                    nn.Linear(cfg.n_embed * mlp_ratio,
                              cfg.n_embed * mlp_ratio))
            modules.append(nn.GELU())
            modules.append(nn.Linear(cfg.n_embed * mlp_ratio, cfg.n_embed))
            modules = nn.Sequential(*modules)

        else:
            raise NotImplementedError(
                f"Unsupported projector type: {cfg.projector_type}")

        self.layers = modules

    def forward(self, x):
        bs, hw, input_dim = x.shape
        h = w = int((hw)**0.5)
        """compute padding"""
        if h % self.cfg.downsample_ratio:
            pad = self.cfg.downsample_ratio - h % self.cfg.downsample_ratio
        else:
            pad = 0
        x = x.reshape(bs, h, w, input_dim)
        if pad > 0:
            x = F.pad(x, (0, 0, 0, pad, 0, pad), "constant", 0)
        """4 to 1 concat"""
        x = x.permute(0, 3, 1, 2)  # B, C, H, W
        x = F.unfold(x,
                     kernel_size=self.cfg.downsample_ratio,
                     stride=self.cfg.downsample_ratio,
                     padding=0)  # B, C*4, HW // 4
        x = x.permute(0, 2, 1)

        return self.layers(x)

cfg instance-attribute

cfg = cfg

layers instance-attribute

layers = modules

__init__

__init__(cfg: MlpProjectorConfig)

Source code in vllm/model_executor/models/deepseek_vl2.py

def __init__(self, cfg: MlpProjectorConfig):

    super().__init__()

    self.cfg = cfg
    assert not cfg.token_pooling, (
        "Token pooling is not supported currently.")

    if cfg.projector_type == "downsample_mlp_gelu":
        mlp_depth = cfg.depth
        mlp_ratio = cfg.mlp_ratio
        modules = [
            nn.Linear(
                cfg.input_dim * cfg.downsample_ratio *
                cfg.downsample_ratio, cfg.n_embed * mlp_ratio)
        ]
        for _ in range(1, mlp_depth - 1):
            modules.append(nn.GELU())
            modules.append(
                nn.Linear(cfg.n_embed * mlp_ratio,
                          cfg.n_embed * mlp_ratio))
        modules.append(nn.GELU())
        modules.append(nn.Linear(cfg.n_embed * mlp_ratio, cfg.n_embed))
        modules = nn.Sequential(*modules)

    else:
        raise NotImplementedError(
            f"Unsupported projector type: {cfg.projector_type}")

    self.layers = modules

forward

forward(x)

Source code in vllm/model_executor/models/deepseek_vl2.py

def forward(self, x):
    bs, hw, input_dim = x.shape
    h = w = int((hw)**0.5)
    """compute padding"""
    if h % self.cfg.downsample_ratio:
        pad = self.cfg.downsample_ratio - h % self.cfg.downsample_ratio
    else:
        pad = 0
    x = x.reshape(bs, h, w, input_dim)
    if pad > 0:
        x = F.pad(x, (0, 0, 0, pad, 0, pad), "constant", 0)
    """4 to 1 concat"""
    x = x.permute(0, 3, 1, 2)  # B, C, H, W
    x = F.unfold(x,
                 kernel_size=self.cfg.downsample_ratio,
                 stride=self.cfg.downsample_ratio,
                 padding=0)  # B, C*4, HW // 4
    x = x.permute(0, 2, 1)

    return self.layers(x)