[Model] Initial support for LLaVA-NeXT (#4199)

Co-authored-by: Roger Wang <ywang@roblox.com>

docs/source/models/supported_models.rst

@@ -89,7 +89,11 @@ Alongside each architecture, we include some popular models that use it.
     - ✅︎
   * - :code:`LlavaForConditionalGeneration`
     - LLaVA-1.5
-    - :code:`llava-hf/llava-1.5-7b-hf`\*, :code:`llava-hf/llava-1.5-13b-hf`\*, etc.
+    - :code:`llava-hf/llava-1.5-7b-hf`, :code:`llava-hf/llava-1.5-13b-hf`, etc.
+    -
+  * - :code:`LlavaNextForConditionalGeneration`
+    - LLaVA-NeXT
+    - :code:`llava-hf/llava-v1.6-mistral-7b-hf`, :code:`llava-hf/llava-v1.6-vicuna-7b-hf`, etc.
     -
   * - :code:`MiniCPMForCausalLM`
     - MiniCPM

tests/models/test_llava.py

@@ -39,8 +39,6 @@ def iter_llava_configs(model_name: str):
 
 model_and_vl_config = [
     *iter_llava_configs("llava-hf/llava-1.5-7b-hf"),
-    # Not enough memory
-    # *iter_llava_configs("llava-hf/llava-1.5-13b-hf"),
 ]
 
 

tests/models/test_llava_next.py

@@ -0,0 +1,123 @@
+from typing import List, Tuple
+
+import pytest
+from transformers import AutoTokenizer
+
+from vllm.config import VisionLanguageConfig
+
+from ..conftest import IMAGE_FILES
+
+pytestmark = pytest.mark.llava
+
+_PREFACE = (
+    "A chat between a curious human and an artificial intelligence assistant. "
+    "The assistant gives helpful, detailed, and polite answers to the human's "
+    "questions.")
+
+# The image token is placed before "user" on purpose so that the test can pass
+HF_IMAGE_PROMPTS = [
+    f"{_PREFACE} <image>\nUSER: What's the content of the image? ASSISTANT:",
+    f"{_PREFACE} <image>\nUSER: What is the season? ASSISTANT:",
+]
+
+assert len(HF_IMAGE_PROMPTS) == len(IMAGE_FILES)
+
+
+def iter_llava_next_configs(model_name: str):
+    image_hw_to_feature_size = {
+        (336, 336): 1176,
+        (672, 672): 2928,
+        (1344, 336): 1944,
+        (336, 1344): 1890,
+    }
+
+    for (h, w), f in image_hw_to_feature_size.items():
+        for input_type, input_shape in [
+            (VisionLanguageConfig.ImageInputType.PIXEL_VALUES, (1, 3, h, w)),
+        ]:
+            yield (model_name,
+                   VisionLanguageConfig(image_input_type=input_type,
+                                        image_feature_size=f,
+                                        image_token_id=32000,
+                                        image_input_shape=input_shape,
+                                        image_processor=model_name,
+                                        image_processor_revision=None))
+
+
+model_and_vl_config = [
+    *iter_llava_next_configs("llava-hf/llava-v1.6-vicuna-7b-hf"),
+]
+
+
+def vllm_to_hf_output(vllm_output: Tuple[List[int], str],
+                      vlm_config: VisionLanguageConfig, model_id: str):
+    """Sanitize vllm output to be comparable with hf output.
+    The function reduces `input_ids` from 1, 32000, 32000, ..., 32000,
+    x1, x2, x3 ... to 1, 32000, x1, x2, x3 ...
+    It also reduces `output_str` from "<image><image>bla" to "bla".
+    """
+    input_ids, output_str = vllm_output
+    image_token_id = vlm_config.image_token_id
+
+    tokenizer = AutoTokenizer.from_pretrained(model_id)
+    image_token_str = tokenizer.decode(image_token_id)
+
+    hf_input_ids = [
+        input_id for idx, input_id in enumerate(input_ids)
+        if input_id != image_token_id or input_ids[idx - 1] != image_token_id
+    ]
+    hf_output_str = output_str \
+        .replace(image_token_str * vlm_config.image_feature_size, " ")
+
+    return hf_input_ids, hf_output_str
+
+
+@pytest.mark.xfail(
+    reason="Inconsistent image processor being used due to lack "
+    "of support for dynamic image token replacement")
+@pytest.mark.parametrize("model_and_config", model_and_vl_config)
+@pytest.mark.parametrize("dtype", ["half"])
+@pytest.mark.parametrize("max_tokens", [128])
+def test_models(hf_runner, vllm_runner, hf_images, vllm_images,
+                model_and_config, dtype: str, max_tokens: int) -> None:
+    """Inference result should be the same between hf and vllm.
+
+    All the image fixtures for the test is under tests/images.
+    For huggingface runner, we provide the PIL images as input.
+    For vllm runner, we provide MultiModalData objects and corresponding
+    vision language config as input.
+    Note, the text input is also adjusted to abide by vllm contract.
+    The text output is sanitized to be able to compare with hf.
+    """
+    model_id, vlm_config = model_and_config
+
+    with hf_runner(model_id, dtype=dtype, is_vision_model=True) as hf_model:
+        hf_outputs = hf_model.generate_greedy(HF_IMAGE_PROMPTS,
+                                              max_tokens,
+                                              images=hf_images)
+
+    vllm_image_prompts = [
+        p.replace("<image>", "<image>" * vlm_config.image_feature_size)
+        for p in HF_IMAGE_PROMPTS
+    ]
+
+    with vllm_runner(
+            model_id,
+            dtype=dtype,
+            # should be greater than image_feature_size
+            max_model_len=4096,
+            enforce_eager=True,
+            **vlm_config.as_cli_args_dict(),
+    ) as vllm_model:
+        vllm_outputs = vllm_model.generate_greedy(vllm_image_prompts,
+                                                  max_tokens,
+                                                  images=vllm_images)
+
+    for i in range(len(HF_IMAGE_PROMPTS)):
+        hf_output_ids, hf_output_str = hf_outputs[i]
+        vllm_output_ids, vllm_output_str = vllm_to_hf_output(
+            vllm_outputs[i], vlm_config, model_id)
+        assert hf_output_str == vllm_output_str, (
+            f"Test{i}:\nHF: {hf_output_str!r}\nvLLM: {vllm_output_str!r}")
+        assert hf_output_ids == vllm_output_ids, (
+            f"Test{i}:\nHF: {hf_output_ids}\nvLLM: {vllm_output_ids}")

tests/multimodal/test_processor.py

@@ -1,6 +1,6 @@
 import numpy as np
 import pytest
-from transformers import CLIPImageProcessor
+from transformers import CLIPImageProcessor, LlavaNextImageProcessor
 
 from vllm.config import ModelConfig, VisionLanguageConfig
 from vllm.multimodal import MULTIMODAL_REGISTRY
@@ -12,7 +12,7 @@ from ..conftest import _STR_DTYPE_TO_TORCH_DTYPE
 @pytest.mark.parametrize("dtype", ["half", "float"])
 def test_clip_image_processor(hf_images, dtype):
     MODEL_NAME = "llava-hf/llava-1.5-7b-hf"
-    IMAGE_HEIGHT = IMAGE_WIDTH = 33
+    IMAGE_HEIGHT = IMAGE_WIDTH = 560
 
     hf_processor = CLIPImageProcessor.from_pretrained(MODEL_NAME)
     assert isinstance(hf_processor, CLIPImageProcessor)
@@ -55,10 +55,61 @@ def test_clip_image_processor(hf_images, dtype):
             assert np.allclose(hf_arr, vllm_arr), f"Failed for key={key}"
 
 
+@pytest.mark.xfail(
+    reason="Inconsistent image processor being used due to lack "
+    "of support for dynamic image token replacement")
+@pytest.mark.parametrize("dtype", ["half", "float"])
+def test_llava_next_image_processor(hf_images, dtype):
+    MODEL_NAME = "llava-hf/llava-v1.6-34b-hf"
+    IMAGE_HEIGHT = IMAGE_WIDTH = 560
+
+    hf_processor = LlavaNextImageProcessor.from_pretrained(MODEL_NAME)
+    assert isinstance(hf_processor, LlavaNextImageProcessor)
+
+    model_config = ModelConfig(
+        model=MODEL_NAME,
+        tokenizer=MODEL_NAME,
+        tokenizer_mode="auto",
+        trust_remote_code=False,
+        seed=0,
+        dtype=dtype,
+        revision=None,
+    )
+    vlm_config = VisionLanguageConfig(
+        image_input_type=VisionLanguageConfig.ImageInputType.PIXEL_VALUES,
+        image_token_id=64000,
+        image_input_shape=(1, 3, IMAGE_HEIGHT, IMAGE_WIDTH),
+        image_feature_size=2928,
+        image_processor=MODEL_NAME,
+        image_processor_revision=None,
+    )
+
+    for image in hf_images:
+        hf_result = hf_processor.preprocess(
+            image,
+            return_tensors="pt",
+        ).to(dtype=_STR_DTYPE_TO_TORCH_DTYPE[dtype])
+        vllm_result = MULTIMODAL_REGISTRY.process_input(
+            ImagePixelData(image),
+            model_config=model_config,
+            vlm_config=vlm_config,
+        )
+
+        assert hf_result.keys() == vllm_result.keys()
+        for key, hf_tensor in hf_result.items():
+            hf_arr: np.ndarray = hf_tensor.numpy()
+            vllm_arr: np.ndarray = vllm_result[key].numpy()
+
+            assert hf_arr.shape == vllm_arr.shape, f"Failed for key={key}"
+            assert np.allclose(hf_arr, vllm_arr), f"Failed for key={key}"
+
+
+@pytest.mark.xfail(
+    reason="Example image pixels were not processed using HuggingFace")
 @pytest.mark.parametrize("dtype", ["float"])
 def test_image_pixel_types(hf_images, vllm_image_tensors, dtype):
     MODEL_NAME = "llava-hf/llava-1.5-7b-hf"
-    IMAGE_HEIGHT = IMAGE_WIDTH = 33
+    IMAGE_HEIGHT = IMAGE_WIDTH = 560
 
     model_config = ModelConfig(
         model=MODEL_NAME,
@@ -95,7 +146,4 @@ def test_image_pixel_types(hf_images, vllm_image_tensors, dtype):
             tensor_arr: np.ndarray = tensor_result[key].numpy()
 
             assert image_arr.shape == tensor_arr.shape, f"Failed for key={key}"
-
-            # The examples in PR#3042 have slightly different preprocessing from
-            # HuggingFace's LlavaProcessor, causing the test to fail.
-            # assert np.allclose(image_arr, tensor_arr), f"Failed for key={key}"
+            assert np.allclose(image_arr, tensor_arr), f"Failed for key={key}"

vllm/model_executor/models/__init__.py

@@ -33,6 +33,8 @@ _GENERATION_MODELS = {
     "LlamaForCausalLM": ("llama", "LlamaForCausalLM"),
     "LlavaForConditionalGeneration":
     ("llava", "LlavaForConditionalGeneration"),
+    "LlavaNextForConditionalGeneration":
+    ("llava_next", "LlavaNextForConditionalGeneration"),
     # For decapoda-research/llama-*
     "LLaMAForCausalLM": ("llama", "LlamaForCausalLM"),
     "MistralForCausalLM": ("llama", "LlamaForCausalLM"),

vllm/model_executor/models/llava.py

@@ -1,7 +1,7 @@
 from typing import Iterable, List, Literal, Optional, Tuple, TypedDict, Union
 
 import torch
-from torch import nn
+import torch.nn as nn
 # TODO(xwjiang): We should port CLIPVisionModel's code over to not depend on
 # transformers' impl.
 from transformers import CLIPVisionModel, LlavaConfig
@@ -51,10 +51,10 @@ class LlavaMultiModalProjector(nn.Module):
         return hidden_states
 
 
-def _merge_vision_embeddings(input_ids: torch.Tensor,
-                             inputs_embeds: torch.Tensor,
-                             vision_embeddings: torch.Tensor,
-                             image_token_id: int) -> torch.Tensor:
+def merge_vision_embeddings(input_ids: torch.Tensor,
+                            inputs_embeds: torch.Tensor,
+                            vision_embeddings: torch.Tensor,
+                            image_token_id: int) -> torch.Tensor:
     """In place merges in vision_embeddings with inputs_embeds."""
     mask = (input_ids == image_token_id)
 
@@ -151,7 +151,8 @@ class LlavaForConditionalGeneration(VisionLanguageModelBase):
                 return None
 
             if not isinstance(pixel_values, torch.Tensor):
-                raise ValueError("Incorrect type of pixel values")
+                raise ValueError("Incorrect type of pixel values. "
+                                 f"Got type: {type(pixel_values)}")
 
             return LlavaImagePixelInputs(
                 type="pixel_values",
@@ -166,7 +167,8 @@ class LlavaForConditionalGeneration(VisionLanguageModelBase):
                 return None
 
             if not isinstance(image_features, torch.Tensor):
-                raise ValueError("Incorrect type of image features")
+                raise ValueError("Incorrect type of image features. "
+                                 f"Got type: {type(image_features)}")
 
             return LlavaImageFeatureInputs(
                 type="image_features",
@@ -268,7 +270,7 @@ class LlavaForConditionalGeneration(VisionLanguageModelBase):
             vision_embeddings = self._process_image_input(image_input)
             inputs_embeds = self.language_model.get_input_embeddings(input_ids)
 
-            inputs_embeds = _merge_vision_embeddings(
+            inputs_embeds = merge_vision_embeddings(
                 input_ids, inputs_embeds, vision_embeddings,
                 self.vision_language_config.image_token_id)
 

vllm/model_executor/models/llava_next.py

@@ -0,0 +1,445 @@
+from typing import (Dict, Iterable, List, Literal, Optional, Tuple, TypedDict,
+                    Union)
+
+import torch
+import torch.nn as nn
+from PIL import Image
+# TODO(xwjiang): We should port CLIPVisionModel's code over to not depend on
+# transformers' impl.
+from transformers import CLIPVisionModel, LlavaNextConfig
+from transformers.models.llava_next.modeling_llava_next import (
+    get_anyres_image_grid_shape, unpad_image)
+from typing_extensions import NotRequired
+
+from vllm.attention import AttentionMetadata
+from vllm.config import CacheConfig, ModelConfig, VisionLanguageConfig
+from vllm.logger import init_logger
+from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.quantization.base_config import (
+    QuantizationConfig)
+from vllm.model_executor.layers.sampler import Sampler
+from vllm.model_executor.layers.vocab_parallel_embedding import ParallelLMHead
+from vllm.model_executor.model_loader.weight_utils import default_weight_loader
+from vllm.model_executor.models.llama import LlamaModel
+from vllm.model_executor.sampling_metadata import SamplingMetadata
+from vllm.multimodal import MULTIMODAL_REGISTRY, MultiModalData
+from vllm.multimodal.image import ImagePixelData, get_dummy_image_data
+from vllm.sequence import SamplerOutput, SequenceData
+
+from .llava import LlavaMultiModalProjector, merge_vision_embeddings
+from .vlm_base import VisionLanguageModelBase
+
+logger = init_logger(__name__)
+
+_KEYS_TO_MODIFY_MAPPING = {
+    "language_model.lm_head": "lm_head",
+    "language_model.model": "language_model",
+}
+
+
+class LlavaNextImagePixelInputs(TypedDict):
+    type: Literal["pixel_values"]
+    data: torch.Tensor
+    """Shape: (batch_size, 1 + num_patches, num_channels, height, width)"""
+
+    image_sizes: NotRequired[torch.Tensor]
+    """Shape: (batch_size, 2)"""
+
+
+class LlavaNextImageFeatureInputs(TypedDict):
+    type: Literal["image_features"]
+    data: torch.Tensor
+    """Shape: (batch_size, 1 + num_patches, image_feature_size, hidden_size)"""
+
+    image_sizes: NotRequired[torch.Tensor]
+    """Shape: (batch_size, 2)"""
+
+
+LlavaNextImageInputs = Union[LlavaNextImagePixelInputs,
+                             LlavaNextImageFeatureInputs]
+
+
+def _get_dummy_image_data(
+    seq_len: int,
+    model_config: ModelConfig,
+    vlm_config: VisionLanguageConfig,
+) -> Tuple[SequenceData, MultiModalData]:
+    seq_data, fake_mm_data = get_dummy_image_data(seq_len, model_config,
+                                                  vlm_config)
+
+    config_input_type = vlm_config.image_input_type
+    ImageInputType = VisionLanguageConfig.ImageInputType
+
+    if config_input_type == ImageInputType.PIXEL_VALUES:
+        _, c, h, w = vlm_config.image_input_shape
+        mode = {1: "L", 3: "RGB"}[c]
+        fake_mm_data = ImagePixelData(Image.new(mode, (w, h), color=0))
+
+    return seq_data, fake_mm_data
+
+
+def _image_pixel_processor(
+    data: ImagePixelData,
+    model_config: ModelConfig,
+    vlm_config: VisionLanguageConfig,
+) -> Dict[str, torch.Tensor]:
+    image = data.image
+
+    if isinstance(image, torch.Tensor):
+        pixel_values = image.to(model_config.dtype)
+        batch_size, _, _, h, w = pixel_values.shape
+        image_sizes = torch.tensor([(w, h) for _ in range(batch_size)])
+
+        return {"pixel_values": pixel_values, "image_sizes": image_sizes}
+
+    # Temporary patch before dynamic number of image tokens is supported
+    _, _, h, w = vlm_config.image_input_shape
+    if (w, h) != (image.width, image.height):
+        logger.warning(
+            "Dynamic image shape is currently not supported. "
+            "Resizing input image to (%d, %d).", w, h)
+
+        data.image = image.resize((w, h))
+
+    return MULTIMODAL_REGISTRY._get_plugin_for_data_type(ImagePixelData) \
+        ._default_input_processor(data, model_config, vlm_config)
+
+
+@MULTIMODAL_REGISTRY.register_image_pixel_input(_image_pixel_processor)
+@MULTIMODAL_REGISTRY.register_dummy_data(_get_dummy_image_data)
+class LlavaNextForConditionalGeneration(VisionLanguageModelBase):
+    """
+    Args to `forward()`:
+        input_ids: Flattened (concatenated) input_ids corresponding to a
+            batch.
+        pixel_values: For PIXEL_VALUES, expects a batch with shape
+            [1, num_patches, 3, 336, 336].
+        image_features: For IMAGE_FEATURES, expects a batch with shape
+            [1, num_patches, 1176, 1024].
+    """
+
+    def __init__(self,
+                 config: LlavaNextConfig,
+                 vision_language_config: VisionLanguageConfig,
+                 cache_config: Optional[CacheConfig] = None,
+                 quant_config: Optional[QuantizationConfig] = None) -> None:
+        super().__init__(vision_language_config)
+
+        # Update the type annotation from that of its superclass
+        self.config = config
+
+        if self.vision_language_config.image_input_type == (
+                VisionLanguageConfig.ImageInputType.PIXEL_VALUES):
+            self.vision_tower = CLIPVisionModel(config.vision_config)
+        else:
+            raise TypeError("Image features are not supported by LLaVA-NeXT")
+
+        self.multi_modal_projector = LlavaMultiModalProjector(
+            vision_hidden_size=config.vision_config.hidden_size,
+            text_hidden_size=config.text_config.hidden_size,
+            projector_hidden_act=config.projector_hidden_act)
+
+        self.quant_config = quant_config
+        self.language_model = LlamaModel(config.text_config, cache_config,
+                                         quant_config)
+        self.unpadded_vocab_size = config.text_config.vocab_size
+        self.lm_head = ParallelLMHead(
+            self.unpadded_vocab_size,
+            config.text_config.hidden_size,
+            org_num_embeddings=self.language_model.org_vocab_size)
+        logit_scale = getattr(config, "logit_scale", 1.0)
+        self.logits_processor = LogitsProcessor(self.unpadded_vocab_size,
+                                                config.vocab_size, logit_scale)
+        self.sampler = Sampler()
+
+        self.image_newline = nn.Parameter(
+            torch.empty(config.text_config.hidden_size))
+
+    def _validate_image_pixels(self, data: torch.Tensor) -> torch.Tensor:
+        _, num_channels, _, _ = self.vision_language_config.image_input_shape
+
+        # Note that this is different from that of vLLM vision_language_config
+        # since the image is resized by the HuggingFace preprocessor
+        height = width = self.config.vision_config.image_size
+
+        if list(data.shape[2:]) != [num_channels, height, width]:
+            raise ValueError(
+                f"The expected image tensor shape is batch dimension plus "
+                f"num_patches plus {[num_channels, height, width]}. "
+                f"You supplied {data.shape}. "
+                f"If you are using vLLM's entrypoint, make sure your "
+                f"supplied image input is consistent with "
+                f"image_input_shape in engine args.")
+
+        return data
+
+    def _validate_image_sizes(self, data: torch.Tensor) -> torch.Tensor:
+        if list(data.shape[1:]) != [2]:
+            raise ValueError(
+                f"The expected image sizes shape is batch dimension plus "
+                f"{[2]}. You supplied {data.shape}.")
+
+        return data
+
+    def _parse_and_validate_image_input(
+            self, **kwargs: object) -> Optional[LlavaNextImageInputs]:
+        pixel_values = kwargs.pop("pixel_values", None)
+        image_sizes = kwargs.pop("image_sizes", None)
+        image_features = kwargs.pop("image_features", None)
+
+        expected_input_type = self.vision_language_config.image_input_type
+        ImageInputType = VisionLanguageConfig.ImageInputType
+
+        if expected_input_type == ImageInputType.PIXEL_VALUES:
+            if image_features is not None:
+                raise ValueError(
+                    "Expected pixel values but got image features")
+            if pixel_values is None:
+                return None
+
+            if not isinstance(pixel_values, torch.Tensor):
+                raise ValueError("Incorrect type of pixel values. "
+                                 f"Got type: {type(pixel_values)}")
+
+            if not isinstance(image_sizes, torch.Tensor):
+                raise ValueError("Incorrect type of image sizes. "
+                                 f"Got type: {type(image_sizes)}")
+
+            return LlavaNextImagePixelInputs(
+                type="pixel_values",
+                data=self._validate_image_pixels(pixel_values),
+                image_sizes=self._validate_image_sizes(image_sizes),
+            )
+
+        assert expected_input_type != ImageInputType.IMAGE_FEATURES, (
+            "Failed to validate this at initialization time")
+
+        return None
+
+    def _merge_image_patch_embeddings(self, image_size: torch.Tensor,
+                                      patch_embeddings: torch.Tensor, *,
+                                      strategy: str) -> torch.Tensor:
+        # Based on: https://github.com/haotian-liu/LLaVA/blob/main/llava/model/llava_arch.py
+        if strategy == "flat":
+            return patch_embeddings.flatten(0, 1)
+
+        if strategy.startswith("spatial"):
+            orig_width, orig_height = image_size
+            height = width = self.config.vision_config.image_size \
+                // self.config.vision_config.patch_size
+
+            base_patch_embeds = patch_embeddings[0]
+            if height * width != base_patch_embeds.shape[0]:
+                raise ValueError(
+                    "The number of patches is not consistent with the "
+                    "image size.")
+
+            if patch_embeddings.shape[0] > 1:
+                other_patch_embeds = patch_embeddings[1:]
+
+                # image_aspect_ratio == "anyres"
+                num_patch_width, num_patch_height = get_anyres_image_grid_shape(
+                    (orig_width, orig_height),
+                    self.config.image_grid_pinpoints,
+                    self.config.vision_config.image_size,
+                )
+                other_patch_embeds = other_patch_embeds \
+                    .view(num_patch_width, num_patch_height, height, width, -1)
+
+                if "unpad" in strategy:
+                    other_patch_embeds = other_patch_embeds \
+                        .permute(4, 0, 2, 1, 3).contiguous() \
+                        .flatten(1, 2).flatten(2, 3)
+                    other_patch_embeds = unpad_image(other_patch_embeds,
+                                                     image_size)
+                    other_patch_embeds = torch.cat((
+                        other_patch_embeds,
+                        self.image_newline[:, None, None] \
+                            .expand(*other_patch_embeds.shape[:-1], 1) \
+                            .to(other_patch_embeds.device),
+                    ), dim=-1)
+                    other_patch_embeds = other_patch_embeds \
+                        .flatten(1, 2).transpose(0, 1)
+                else:
+                    other_patch_embeds = other_patch_embeds \
+                        .permute(0, 2, 1, 3, 4).contiguous() \
+                        .flatten(0, 3)
+
+                merged_patch_embeddings = torch.cat(
+                    (base_patch_embeds, other_patch_embeds), dim=0)
+            else:
+                if "unpad" in strategy:
+                    merged_patch_embeddings = torch.cat(
+                        (base_patch_embeds,
+                         self.image_newline[None] \
+                            .to(base_patch_embeds.device)
+                    ), dim=0)
+                else:
+                    merged_patch_embeddings = base_patch_embeds
+
+            return merged_patch_embeddings
+
+        raise ValueError(f"Unexpected patch merge strategy: {strategy}")
+
+    def _process_image_pixels(
+            self, inputs: LlavaNextImagePixelInputs) -> torch.Tensor:
+        assert self.vision_tower is not None
+
+        pixel_values = inputs["data"]
+
+        b, num_patches, c, h, w = pixel_values.shape
+        stacked_pixel_values = pixel_values.view(b * num_patches, c, h, w)
+
+        stacked_image_features = self._image_pixels_to_features(
+            self.vision_tower, stacked_pixel_values)
+
+        return stacked_image_features.view(b, num_patches,
+                                           *stacked_image_features.shape[-2:])
+
+    def _process_image_input(
+            self, image_input: LlavaNextImageInputs) -> torch.Tensor:
+        if image_input["type"] == "pixel_values":
+            assert self.vision_tower is not None
+            image_features = self._process_image_pixels(image_input)
+        else:
+            image_features = image_input["data"]
+
+        patch_embeddings = self.multi_modal_projector(image_features)
+
+        image_sizes = image_input.get("image_sizes")
+        if image_sizes is None:
+            batch_size = image_input["data"].shape[0]
+            vision_config = self.config.vision_config
+            default_width = default_height = vision_config.image_size
+            image_sizes = torch.as_tensor([[default_width, default_height]
+                                           for _ in range(batch_size)])
+
+        merged_patch_embeddings = [
+            self._merge_image_patch_embeddings(image_sizes[i],
+                                               patch_features,
+                                               strategy="spatial_unpad")
+            for i, patch_features in enumerate(patch_embeddings)
+        ]
+
+        return torch.stack(merged_patch_embeddings, dim=0)
+
+    def forward(
+        self,
+        input_ids: torch.Tensor,
+        positions: torch.Tensor,
+        kv_caches: List[torch.Tensor],
+        attn_metadata: AttentionMetadata,
+        **kwargs: object,
+    ) -> SamplerOutput:
+        """Run forward pass for Llava 1.5.
+
+        One key thing to understand is the `input_ids` already accounts for the
+        positions of the to-be-inserted image embeddings.
+        Concretely, consider a text prompt:
+        "<image>\nUSER: What's the content of the image?\nASSISTANT:".
+        Tokenizer outputs:
+        [1, 32000, 29871, 13, 11889, 29901, 1724, 29915, 29879, 278,
+        2793, 310, 278, 1967, 29973, 13, 22933, 9047, 13566, 29901].
+        The to-be-inserted image has a size of 576 (24 * 24) along the context
+        length dimension.
+        `input_ids` is thus [1, 32000, ..., 32000, 29871, 13, 11889, 29901,
+        1724, 29915, 29879, 278, 2793, 310, 278, 1967, 29973, 13, 22933,
+        9047, 13566, 29901].
+        There will be 576 `32000` in the `input_ids`.
+        (32000 is the token id for `<image>`.)
+
+        This way, the `positions` and `attn_metadata` are consistent
+        with the `input_ids`.
+
+        The model takes two types of image inputs:
+        PIXEL_VALUES and IMAGE_FEATURES.
+        The following shows how each maps to huggingface implementation.
+        PIXEL_VALUES:
+        - https://github.com/huggingface/transformers/blob/07bdbeb/src/transformers/models/llava/modeling_llava.py#L353
+        IMAGE_FEATURES:
+        - https://github.com/huggingface/transformers/blob/07bdbeb/src/transformers/models/llava/modeling_llava.py#L430
+        before going through the multi modal projector.
+
+        Args:
+            input_ids: Flattened (concatenated) input_ids corresponding to a
+                batch.
+            pixel_values: For PIXEL_VALUES, expects a batch with shape
+                [1, 3, 336, 336].
+            image_features: For IMAGE_FEATURES, expects a batch with shape
+                [1, 576, 1024].
+        """
+        image_input = self._parse_and_validate_image_input(**kwargs)
+
+        if image_input is not None:
+            vision_embeddings = self._process_image_input(image_input)
+            inputs_embeds = self.language_model.get_input_embeddings(input_ids)
+
+            inputs_embeds = merge_vision_embeddings(
+                input_ids, inputs_embeds, vision_embeddings,
+                self.vision_language_config.image_token_id)
+
+            input_ids = None
+        else:
+            inputs_embeds = None
+
+        hidden_states = self.language_model(input_ids,
+                                            positions,
+                                            kv_caches,
+                                            attn_metadata,
+                                            inputs_embeds=inputs_embeds)
+
+        return hidden_states
+
+    def compute_logits(self, hidden_states: torch.Tensor,
+                       sampling_metadata: SamplingMetadata) -> torch.Tensor:
+        logits = self.logits_processor(self.lm_head.weight, hidden_states,
+                                       sampling_metadata)
+        return logits
+
+    def sample(
+        self,
+        logits: torch.Tensor,
+        sampling_metadata: SamplingMetadata,
+    ) -> Optional[SamplerOutput]:
+        next_tokens = self.sampler(logits, sampling_metadata)
+        return next_tokens
+
+    def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
+        # only doing this for language model part for now.
+        stacked_params_mapping = [
+            # (param_name, shard_name, shard_id)
+            ("qkv_proj", "q_proj", "q"),
+            ("qkv_proj", "k_proj", "k"),
+            ("qkv_proj", "v_proj", "v"),
+            ("gate_up_proj", "gate_proj", 0),
+            ("gate_up_proj", "up_proj", 1),
+        ]
+        params_dict = dict(self.named_parameters())
+        for name, loaded_weight in weights:
+            if "rotary_emb.inv_freq" in name:
+                continue
+            for key_to_modify, new_key in _KEYS_TO_MODIFY_MAPPING.items():
+                if key_to_modify in name:
+                    name = name.replace(key_to_modify, new_key)
+            use_default_weight_loading = False
+            if "vision" in name:
+                if self.vision_tower is not None:
+                    # We only do sharding for language model and
+                    # not vision model for now.
+                    use_default_weight_loading = True
+            else:
+                for (param_name, weight_name,
+                     shard_id) in stacked_params_mapping:
+                    if weight_name not in name:
+                        continue
+                    param = params_dict[name.replace(weight_name, param_name)]
+                    weight_loader = param.weight_loader
+                    weight_loader(param, loaded_weight, shard_id)
+                    break
+                else:
+                    use_default_weight_loading = True
+            if use_default_weight_loading:
+                param = params_dict[name]
+                weight_loader = getattr(param, "weight_loader",
+                                        default_weight_loader)
+                weight_loader(param, loaded_weight)