prumerge_llava_next.py

# ===================================   PruMerge and PruMerge+   ======================================
# Paste this code into src/transformers/models/llava_next/modeling_llava_next.py to use PruMerge and PruMerge+
# =====================================================================================================

# Credits:
# Code is copied from: https://github.com/42Shawn/LLaVA-PruMerge
# paper: https://arxiv.org/abs/2403.15388

import math, time
from dataclasses import dataclass
from typing import List, Optional, Tuple, Union

import numpy as np
import torch
import torch.utils.checkpoint
from torch import nn
import torch.nn.functional as F

from ... import PreTrainedModel
from ...activations import ACT2FN
from ...cache_utils import Cache
from ...image_processing_utils import select_best_resolution
from ...modeling_outputs import ModelOutput
from ...utils import (
    add_start_docstrings,
    add_start_docstrings_to_model_forward,
    logging,
    replace_return_docstrings,
)
from ..auto import AutoModel, AutoModelForCausalLM
from .configuration_llava_next import LlavaNextConfig


logger = logging.get_logger(__name__)

_CONFIG_FOR_DOC = "LlavaNextConfig"


# ===================================   PruMerge and PruMerge+ HELPER  ======================================

# for prumerge
outputs = {}

def hook_k(module, input, output):
    outputs['desired_k'] = output

def hook_q(module, input, output):
    outputs['desired_q'] = output

def complement_idx(idx, dim):
    a = torch.arange(dim, device=idx.device)
    ndim = idx.ndim
    dims = idx.shape
    n_idx = dims[-1]
    dims = dims[:-1] + (-1, )
    for i in range(1, ndim):
        a = a.unsqueeze(0)
    a = a.expand(*dims)
    masked = torch.scatter(a, -1, idx, 0)
    compl, _ = torch.sort(masked, dim=-1, descending=False)
    compl = compl.permute(-1, *tuple(range(ndim - 1)))
    compl = compl[n_idx:].permute(*(tuple(range(1, ndim)) + (0,)))
    return compl


def outlier_dectection(attn):
    attn_np = attn.to(dtype=torch.float32).cpu().numpy().flatten()

    Q1 = np.percentile(attn_np, 25)
    Q3 = np.percentile(attn_np, 75)
    IQR = Q3 - Q1

    # lower_bound = Q1 - 1.5 * IQR
    upper_bound = Q3 + 1.5 * IQR

    outlier_indices = np.where((attn_np > upper_bound))[0]

    ratio = len(outlier_indices) / len(attn_np)
    return ratio


# ===================================   PruMerge and PruMerge+ HELPER END  ======================================

def get_anyres_image_grid_shape(image_size, grid_pinpoints, patch_size):
    """
    Calculate the shape of the image patch grid after the preprocessing for images of any resolution.

    Args:
        image_size (`tuple`):
            The size of the input image in the format (width, height).
        grid_pinpoints (`List`):
            A list containing possible resolutions. Each item in the list should be a tuple or list
            of the form `(height, width)`.
        patch_size (`int`):
            The size of each image patch.

    Returns:
        tuple: The shape of the image patch grid in the format (width, height).
    """
    if not isinstance(grid_pinpoints, list):
        raise ValueError("grid_pinpoints should be a list of tuples or lists")

    # ! VERY IMPORTANT if image_size is tensor, must convert to into tuple, otherwise it will cause wrong calculate
    if not isinstance(image_size, (list, tuple)):
        if not isinstance(image_size, (torch.Tensor, np.ndarray)):
            raise ValueError(
                f"image_size invalid type: {type(image_size)} not valid, should be either list, tuple, np.ndarray or tensor"
            )
        image_size = image_size.tolist()

    height, width = select_best_resolution(image_size, grid_pinpoints)
    return height // patch_size, width // patch_size


def image_size_to_num_patches(image_size, grid_pinpoints, patch_size: int):
    """
    Calculate the number of patches after the preprocessing for images of any resolution.

    Args:
        image_size (`Union[torch.LongTensor, np.ndarray, Tuple[int, int]):
            The size of the input image in the format (height, width). ?
        grid_pinpoints (`List`):
            A list containing possible resolutions. Each item in the list should be a tuple or list
            of the form `(height, width)`.
        patch_size (`int`):
            The size of each image patch.

    Returns:
        int: the number of patches
    """
    if not isinstance(grid_pinpoints, list):
        raise ValueError("grid_pinpoints should be a list of tuples or lists")

    # ! VERY IMPORTANT if image_size is tensor, must convert to into tuple, otherwise it will cause wrong calculate
    if not isinstance(image_size, (list, tuple)):
        if not isinstance(image_size, (torch.Tensor, np.ndarray)):
            raise ValueError(f"image_size invalid type {type(image_size)} with value {image_size}")
        image_size = image_size.tolist()

    best_resolution = select_best_resolution(image_size, grid_pinpoints)
    height, width = best_resolution
    num_patches = 0
    # consider change to ceil(height/patch_size)*ceil(width/patch_size) + 1
    for i in range(0, height, patch_size):
        for j in range(0, width, patch_size):
            num_patches += 1
    # add the base patch
    num_patches += 1
    return num_patches


def unpad_image(tensor, original_size):
    """
    Unpads a PyTorch tensor of a padded and resized image or mask.
    A simple hack on top of the original implementation so that this
    function can handle both images and masks.

    Args:
        tensor (`torch.Tensor`):
            The image or mask tensor, assumed to be of shape (num_channels, height, width) for images or (height, width) for masks.
        original_size (`tuple`):
            The original size of the image (height, width).

    Returns:
        `torch.Tensor`: The unpadded image or mask tensor.
    """
    original_height, original_width = original_size
    if tensor.ndim == 3:
        current_height, current_width = tensor.shape[1:]
    else:
        current_height, current_width = tensor.shape

    original_aspect_ratio = original_width / original_height
    current_aspect_ratio = current_width / current_height

    if original_aspect_ratio > current_aspect_ratio:
        scale_factor = current_width / original_width
        new_height = int(original_height * scale_factor)
        padding = (current_height - new_height) // 2
        if tensor.ndim == 3:
            unpadded_tensor = tensor[:, padding : current_height - padding, :]
        else:
            unpadded_tensor = tensor[padding : current_height - padding, :]
    else:
        scale_factor = current_height / original_height
        new_width = int(original_width * scale_factor)
        padding = (current_width - new_width) // 2
        if tensor.ndim == 3:
            unpadded_tensor = tensor[:, :, padding : current_width - padding]
        else:
            unpadded_tensor = tensor[:, padding : current_width - padding]

    return unpadded_tensor


@dataclass
# Copied from transformers.models.idefics.modeling_idefics.IdeficsCausalLMOutputWithPast with Idefics->LlavaNext
class LlavaNextCausalLMOutputWithPast(ModelOutput):
    """
    Base class for LlavaNext causal language model (or autoregressive) outputs.

    Args:
        loss (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided):
            Language modeling loss (for next-token prediction).
        logits (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`):
            Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
        past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
            Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of shape
            `(batch_size, num_heads, sequence_length, embed_size_per_head)`)

            Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see
            `past_key_values` input) to speed up sequential decoding.
        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
            Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
            one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.

            Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
            sequence_length)`.

            Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
            heads.
        image_hidden_states (`tuple(torch.FloatTensor)`, *optional*):
            Tuple of `torch.FloatTensor` (one for the output of the image embeddings, `(batch_size, num_images,
            sequence_length, hidden_size)`.

            image_hidden_states of the model produced by the vision encoder, and optionally by the perceiver
    """

    loss: Optional[torch.FloatTensor] = None
    logits: torch.FloatTensor = None
    past_key_values: Optional[List[torch.FloatTensor]] = None
    hidden_states: Optional[Tuple[torch.FloatTensor]] = None
    attentions: Optional[Tuple[torch.FloatTensor]] = None
    image_hidden_states: Optional[Tuple[torch.FloatTensor]] = None


# Copied from transformers.models.llava.modeling_llava.LlavaMultiModalProjector with Llava->LlavaNext
class LlavaNextMultiModalProjector(nn.Module):
    def __init__(self, config: LlavaNextConfig):
        super().__init__()

        self.linear_1 = nn.Linear(config.vision_config.hidden_size, config.text_config.hidden_size, bias=True)
        self.act = ACT2FN[config.projector_hidden_act]
        self.linear_2 = nn.Linear(config.text_config.hidden_size, config.text_config.hidden_size, bias=True)

    def forward(self, image_features):
        hidden_states = self.linear_1(image_features)
        hidden_states = self.act(hidden_states)
        hidden_states = self.linear_2(hidden_states)
        return hidden_states


LLAVA_NEXT_START_DOCSTRING = r"""
    This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the
    library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
    etc.)

    This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
    Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
    and behavior.

    Parameters:
        config ([`LlavaNextConfig`] or [`LlavaNextVisionConfig`]):
            Model configuration class with all the parameters of the model. Initializing with a config file does not
            load the weights associated with the model, only the configuration. Check out the
            [`~PreTrainedModel.from_pretrained`] method to load the model weights.
"""


@add_start_docstrings(
    "The bare LLaMA Model outputting raw hidden-states without any specific head on top.",
    LLAVA_NEXT_START_DOCSTRING,
)
# Copied from transformers.models.llava.modeling_llava.LlavaPreTrainedModel with Llava->LlavaNext,llava->llava_next
class LlavaNextPreTrainedModel(PreTrainedModel):
    config_class = LlavaNextConfig
    base_model_prefix = "model"
    supports_gradient_checkpointing = True
    _no_split_modules = ["LlavaNextVisionAttention"]
    _skip_keys_device_placement = "past_key_values"
    _supports_flash_attn_2 = True

    def _init_weights(self, module):
        # important: this ported version of LlavaNext isn't meant for training from scratch - only
        # inference and fine-tuning - so the proper init weights code has been removed - the original codebase
        # https://github.com/haotian-liu/LLaVA/tree/main/llava_next should serve for that purpose
        std = (
            self.config.initializer_range
            if hasattr(self.config, "initializer_range")
            else self.config.text_config.initializer_range
        )

        if hasattr(module, "class_embedding"):
            module.class_embedding.data.normal_(mean=0.0, std=std)

        if isinstance(module, (nn.Linear, nn.Conv2d)):
            module.weight.data.normal_(mean=0.0, std=std)
            if module.bias is not None:
                module.bias.data.zero_()
        elif isinstance(module, nn.Embedding):
            module.weight.data.normal_(mean=0.0, std=std)
            if module.padding_idx is not None:
                module.weight.data[module.padding_idx].zero_()

    @property
    def _supports_sdpa(self):
        """
        Retrieve language_model's attribute to check whether the model supports
        SDPA or not.
        """
        return self.language_model._supports_sdpa


LLAVA_NEXT_INPUTS_DOCSTRING = r"""
    Args:
        input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
            Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
            it.

            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
            [`PreTrainedTokenizer.__call__`] for details.

            [What are input IDs?](../glossary#input-ids)
        pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, image_size, image_size)):
            The tensors corresponding to the input images. Pixel values can be obtained using
            [`AutoImageProcessor`]. See [`LlavaNextImageProcessor.__call__`] for details. [`LlavaProcessor`] uses
            [`LlavaNextImageProcessor`] for processing images.
        image_sizes (`torch.LongTensor` of shape `(batch_size, 2)`, *optional*):
            The sizes of the images in the batch, being (height, width) for each image.
        attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
            Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:

            - 1 for tokens that are **not masked**,
            - 0 for tokens that are **masked**.

            [What are attention masks?](../glossary#attention-mask)

            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
            [`PreTrainedTokenizer.__call__`] for details.

            If `past_key_values` is used, optionally only the last `decoder_input_ids` have to be input (see
            `past_key_values`).

            If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`]
            and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more
            information on the default strategy.

            - 1 indicates the head is **not masked**,
            - 0 indicates the head is **masked**.
        position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
            config.n_positions - 1]`. [What are position IDs?](../glossary#position-ids)
        past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
            Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of shape
            `(batch_size, num_heads, sequence_length, embed_size_per_head)`) and 2 additional tensors of shape
            `(batch_size, num_heads, encoder_sequence_length, embed_size_per_head)`.

            Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention
            blocks) that can be used (see `past_key_values` input) to speed up sequential decoding.

            If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those that
            don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of all
            `decoder_input_ids` of shape `(batch_size, sequence_length)`.
        inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
            Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
            is useful if you want more control over how to convert `input_ids` indices into associated vectors than the
            model's internal embedding lookup matrix.
        vision_feature_layer (`int`, *optional*, defaults to -2):
            The index of the layer to select the vision feature.
        vision_feature_select_strategy (`str`, *optional*, defaults to `"default"`):
            The feature selection strategy used to select the vision feature from the vision backbone.
            Can be one of `"default"` or `"full"`. If `"default"`, the CLS token is removed from the vision features.
            If `"full"`, the full vision features are used.
        use_cache (`bool`, *optional*):
            If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see
            `past_key_values`).
        output_attentions (`bool`, *optional*):
            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
            tensors for more detail.
        output_hidden_states (`bool`, *optional*):
            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
            more detail.
        return_dict (`bool`, *optional*):
            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
"""


@add_start_docstrings(
    """The LLAVA-NeXT model which consists of a vision backbone and a language model.""",
    LLAVA_NEXT_START_DOCSTRING,
)
class LlavaNextForConditionalGeneration(LlavaNextPreTrainedModel):
    def __init__(self, config: LlavaNextConfig):
        super().__init__(config)
        self.vision_tower = AutoModel.from_config(config.vision_config)

        self.multi_modal_projector = LlavaNextMultiModalProjector(config)
        embed_std = 1 / math.sqrt(config.text_config.hidden_size)
        self.image_newline = nn.Parameter(torch.randn(config.text_config.hidden_size, dtype=self.dtype) * embed_std)

        self.vocab_size = config.text_config.vocab_size
        self.language_model = AutoModelForCausalLM.from_config(
            config.text_config, attn_implementation=config._attn_implementation
        )
        self.vit_to_llm_mapping = None
        self.pad_token_id = self.config.pad_token_id if self.config.pad_token_id is not None else -1
        self._padding_side = "left"  # set it to left by default, user can use setter to change padding_sides
        self.post_init()

    @property
    def padding_side(self):
        return self._padding_side

    @padding_side.setter
    def padding_side(self, padding_side: str):
        if padding_side not in ["left", "right"]:
            raise ValueError(f"{padding_side} is not `left` or `right`.")
        self._padding_side = padding_side

    # Copied from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration.get_input_embeddings
    def get_input_embeddings(self):
        return self.language_model.get_input_embeddings()

    # Copied from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration.set_input_embeddings
    def set_input_embeddings(self, value):
        self.language_model.set_input_embeddings(value)

    # Copied from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration.get_output_embeddings
    def get_output_embeddings(self):
        return self.language_model.get_output_embeddings()

    # Copied from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration.set_output_embeddings
    def set_output_embeddings(self, new_embeddings):
        self.language_model.set_output_embeddings(new_embeddings)

    # Copied from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration.set_decoder
    def set_decoder(self, decoder):
        self.language_model.set_decoder(decoder)

    # Copied from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration.get_decoder
    def get_decoder(self):
        return self.language_model.get_decoder()

    # Copied from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration.tie_weights
    def tie_weights(self):
        return self.language_model.tie_weights()

    # Copied from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration.resize_token_embeddings
    def resize_token_embeddings(self, new_num_tokens: Optional[int] = None, pad_to_multiple_of=None) -> nn.Embedding:
        model_embeds = self.language_model.resize_token_embeddings(new_num_tokens, pad_to_multiple_of)
        # update vocab size
        self.config.text_config.vocab_size = model_embeds.num_embeddings
        self.vocab_size = model_embeds.num_embeddings
        return model_embeds
    
    

    def _merge_input_ids_with_image_features(
        self,
        image_features,
        feature_lens,
        inputs_embeds,
        input_ids,
        attention_mask,
        position_ids=None,
        labels=None,
        image_token_index=None,
        ignore_index=-100,
    ):
        """
        Merge input_ids with with image features into final embeddings

        Args:
            image_features (`torch.Tensor` of shape `(all_feature_lens, embed_dim)`):
                All vision vectors of all images in the batch
            feature_lens (`torch.LongTensor` of shape `(num_images)`):
                The length of visual embeddings of each image as stacked in `image_features`
            inputs_embeds (`torch.Tensor` of shape `(batch_size, sequence_length, embed_dim)`):
                Token embeddings before merging with visual embeddings
            input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
                Input_ids of tokens, possibly filled with image token
            attention_mask (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
                Mask to avoid performing attention on padding token indices.
            position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
                Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
                config.n_positions - 1]`.
            labels (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*)
                :abels need to be recalculated to support training (if provided)
            image_token_index (`int`, *optional*)
                Token id used to indicate the special "image" token. Defaults to `config.image_token_index`
            ignore_index (`int`, *optional*)
                Value that is used to pad `labels` and will be ignored when calculated loss. Default: -100.
        Returns:
            final_embedding, final_attention_mask, position_ids, final_labels

        Explanation:
            each image has variable length embeddings, with length specified by feature_lens
            image_features is concatenation of all visual embed vectors
            task: fill each <image> with the correct number of visual embeddings
            Example:
                X (5 patches), Y (3 patches), Z (8)
                X, Y are in the same sequence (in-context learning)
            if right padding
                input_ids: [
                    a b c d e f X g h i j k Y l m
                    o p q r Z s t u v _ _ _ _ _ _
                ]
                input_ids should be: [
                    a b c d e f X X X X X g h i j k Y Y Y l m
                    o p q r Z Z Z Z Z Z Z Z s t u v _ _ _ _ _
                ]
                labels should be: [
                    a b c d e f _ _ _ _ _ g h i j k _ _ _ l m
                    o p q r _ _ _ _ _ _ _ _ s t u v _ _ _ _ _
                ]
            elif left padding
                input_ids: [
                    a b c d e f X g h i j k Y l m
                    _ _ _ _ _ _ o p q r Z s t u v
                ]
                input_ids should be: [
                    a b c d e f X X X X X g h i j k Y Y Y l m
                    _ _ _ _ _ o p q r Z Z Z Z Z Z Z Z s t u v
                ]
                labels should be: [
                    a b c d e f _ _ _ _ _ g h i j k _ _ _ l m
                    _ _ _ _ _ o p q r _ _ _ _ _ _ _ _ s t u v
                ]
            Edge cases:
                * If tokens are same but image token sizes are different, then cannot infer left or right padding
                ```python
                cat_img = Image.open(requests.get("http://images.cocodataset.org/val2017/000000039769.jpg", stream=True).raw)
                chart_img = Image.open(requests.get("https://github.com/haotian-liu/LLaVA/blob/1a91fc274d7c35a9b50b3cb29c4247ae5837ce39/images/llava_v1_5_radar.jpg?raw=true", stream=True).raw)
                prompts = [
                    "[INST] <image>\nWhat is shown in this image? [/INST]",
                    "[INST] <image>\nWhat is shown in this image? [/INST]",
                ]
                inputs = processor(prompts, [chart_img, cat_img], return_tensors='pt', padding=True).to("cuda")
                    chart_img has 2634 tokens, while cat_img has 2340 tokens
                ```

                input_ids: [
                    a b c d X g h
                    i j Y k l m n
                ]
                where X is 3 tokens while Y is 5, this mean after merge
                if left-padding (batched generation)
                    input_ids should be: [
                        _ _ a b c d X X X g h
                        i j Y Y Y Y Y k l m n
                    ]
                elif (right padding) (training)
                    input_ids should be: [
                        a b c d X X X g h _ _
                        i j Y Y Y Y Y k l m n
                    ]
        """
        image_token_index = image_token_index if image_token_index is not None else self.config.image_token_index
        ignore_index = ignore_index if ignore_index is not None else self.config.ignore_index

        with torch.no_grad():
            # ! in llava 1.6, number of patches is variable
            num_images = feature_lens.size(0)
            num_image_features, embed_dim = image_features.shape
            batch_size = input_ids.shape[0]
            _left_padding = torch.any(attention_mask[:, 0] == 0)
            _right_padding = torch.any(attention_mask[:, -1] == 0)

            left_padding = True
            if batch_size > 1:
                if _left_padding and not _right_padding:
                    left_padding = True
                elif not _left_padding and _right_padding:
                    left_padding = False
                elif not _left_padding and not _right_padding:
                    # both side is 1, so cannot tell
                    left_padding = self.padding_side == "left"
                else:
                    # invalid attention_mask
                    raise ValueError(f"both side of attention_mask has zero, invalid. {attention_mask}")

            # Whether to turn off right padding
            # 1. Create a mask to know where special image tokens are
            special_image_token_mask = input_ids == image_token_index
            # special_image_token_mask: [bsz, seqlen]
            num_special_image_tokens = torch.sum(special_image_token_mask, dim=-1)
            # num_special_image_tokens: [bsz]
            # Reserve for padding of num_images
            total_num_special_image_tokens = torch.sum(special_image_token_mask)
            if total_num_special_image_tokens != num_images:
                raise ValueError(
                    f"Number of image tokens in input_ids ({total_num_special_image_tokens}) different from num_images ({num_images})."
                )
            # Compute the maximum embed dimension
            # max_image_feature_lens is max_feature_lens per batch
            feature_lens_batch = feature_lens.split(num_special_image_tokens.tolist(), dim=0)
            feature_lens_batch_sum = torch.tensor([x.sum() for x in feature_lens_batch], device=feature_lens.device)
            embed_sequence_lengths = (
                (attention_mask == 1).long().sum(-1) - num_special_image_tokens + feature_lens_batch_sum
            )
            max_embed_dim = embed_sequence_lengths.max()

            batch_indices, non_image_indices = torch.where((input_ids != image_token_index) & (attention_mask == 1))
            # 2. Compute the positions where text should be written
            # Calculate new positions for text tokens in merged image-text sequence.
            # `special_image_token_mask` identifies image tokens. Each image token will be replaced by `nb_text_tokens_per_images` text tokens.
            # `torch.cumsum` computes how each image token shifts subsequent text token positions.
            # - 1 to adjust for zero-based indexing, as `cumsum` inherently increases indices by one.
            # ! instead of special_image_token_mask * (num_image_patches - 1)
            #   special_image_token_mask * (num_feature_len - 1)
            special_image_token_mask = special_image_token_mask.long()
            special_image_token_mask[special_image_token_mask == 1] = feature_lens - 1
            new_token_positions = torch.cumsum((special_image_token_mask + 1), -1) - 1
            if left_padding:
                # shift right token positions so that they are ending at the same number
                # the below here was incorrect? new_token_positions += new_token_positions[:, -1].max() - new_token_positions[:, -1:]
                new_token_positions += max_embed_dim - 1 - new_token_positions[:, -1:]

            text_to_overwrite = new_token_positions[batch_indices, non_image_indices]

        # 3. Create the full embedding, already padded to the maximum position
        final_embedding = torch.zeros(
            batch_size, max_embed_dim, embed_dim, dtype=inputs_embeds.dtype, device=inputs_embeds.device
        )
        final_attention_mask = torch.zeros(
            batch_size, max_embed_dim, dtype=attention_mask.dtype, device=inputs_embeds.device
        )
        final_labels = None
        if labels is not None:
            final_labels = torch.full_like(final_attention_mask, ignore_index).to(torch.long)
        # In case the Vision model or the Language model has been offloaded to CPU, we need to manually
        # set the corresponding tensors into their correct target device.
        target_device = inputs_embeds.device
        batch_indices, non_image_indices, text_to_overwrite = (
            batch_indices.to(target_device),
            non_image_indices.to(target_device),
            text_to_overwrite.to(target_device),
        )
        attention_mask = attention_mask.to(target_device)

        # 4. Fill the embeddings based on the mask. If we have ["hey" "<image>", "how", "are"]
        # we need to index copy on [0, 577, 578, 579] for the text and [1:576] for the image features
        final_embedding[batch_indices, text_to_overwrite] = inputs_embeds[batch_indices, non_image_indices]
        final_attention_mask[batch_indices, text_to_overwrite] = attention_mask[batch_indices, non_image_indices]
        if labels is not None:
            final_labels[batch_indices, text_to_overwrite] = labels[batch_indices, non_image_indices]

        # 5. Fill the embeddings corresponding to the images. Anything that is not `text_positions` needs filling (#29835)
        with torch.no_grad():
            image_to_overwrite = torch.full(
                (batch_size, max_embed_dim), True, dtype=torch.bool, device=inputs_embeds.device
            )
            image_to_overwrite[batch_indices, text_to_overwrite] = False
            embed_indices = torch.arange(max_embed_dim).unsqueeze(0).to(target_device)
            embed_indices = embed_indices.expand(batch_size, max_embed_dim)
            embed_seq_lens = embed_sequence_lengths[:, None].to(target_device)

            if left_padding:
                # exclude padding on the left
                val = (max_embed_dim - embed_indices) <= embed_seq_lens
            else:
                # exclude padding on the right
                val = embed_indices < embed_seq_lens
            image_to_overwrite &= val

            if image_to_overwrite.sum() != num_image_features:
                raise ValueError(
                    f"{image_to_overwrite.sum()=} != {num_image_features=} The input provided to the model are wrong. "
                    f"The number of image tokens is {torch.sum(special_image_token_mask)} while"
                    f" the number of image given to the model is {num_images}. "
                    f"This prevents correct indexing and breaks batch generation."
                )
        final_embedding[image_to_overwrite] = image_features.contiguous().reshape(-1, embed_dim).to(target_device)
        final_attention_mask |= image_to_overwrite
        position_ids = (final_attention_mask.cumsum(-1) - 1).masked_fill_((final_attention_mask == 0), 1)

        # Calculate the image feature indices in the final embedding
        image_indices = []
        for i in range(batch_size):
            image_positions = torch.where(image_to_overwrite[i])[0]
            image_indices.append(image_positions)

        return final_embedding, final_attention_mask, position_ids, final_labels, image_indices

    def pack_image_features(self, image_features, image_sizes, image_newline=None, image_features_masks=None):
        """
        Reshape, unpad and then pack each image_feature into a single image_features tensor containing all visual vectors.

        Args:
            image_features (`List[torch.Tensor]` of length num_images, each of shape `(num_patches, image_length, embed_dim)`):
                List of image feature tensor, each contains all the visual feature of all patches.
            image_sizes (`torch.Tensor` of shape `(num_images, 2)`):
                Actual image size of each image (H, W).
            image_newline (`torch.Tensor` of shape `(embed_dim)`):
                New line embedding vector.
            image_features_masks Tuple of (`torch.Tensor` of shape `(num_patches, num_tokens)`):
                Mask for selecting tokens.
        Returns:
            image_features (`torch.Tensor` of shape `(all_feat_len, embed_dim)`):
                Concatenated image features.
            feature_lens (`List[int]`):
                Token length of each image in image_features.
        """
        new_image_features = []
        feature_lens = []
        for image_idx, (image_feature, image_features_mask) in enumerate(zip(image_features, image_features_masks)):
            if image_feature.shape[0] > 1:
                base_image_feature = image_feature[0] # [576, 4096]
                base_image_mask = image_features_mask[0] # [576]
                reduced_base_image_feature = base_image_feature[base_image_mask.bool()] # [reduced, 4096]
                image_feature = image_feature[1:] # [4, 576, 4096]
                other_images_mask = image_features_mask[1:] # [4, 576]
                height = width = self.config.vision_config.image_size // self.config.vision_config.patch_size # 24
                if height * width != base_image_feature.shape[0]:
                    raise ValueError("The number of patches is not consistent with the image size.")
                
                
                num_patch_width, num_patch_height = get_anyres_image_grid_shape(
                    image_sizes[image_idx],
                    self.config.image_grid_pinpoints,
                    self.config.vision_config.image_size,
                ) # (2, 2)
                image_feature = image_feature.view(num_patch_height, num_patch_width, height, width, -1) # [2, 2, 24, 24, 4096]
                other_images_mask = other_images_mask.view(num_patch_height, num_patch_width, height, width) # [2, 2, 24, 24]
                image_feature = image_feature.permute(4, 0, 2, 1, 3).contiguous() # [4096, 2, 24, 2, 24]
                other_images_mask = other_images_mask.permute(0, 2, 1, 3).contiguous() # [2, 24, 2, 24]
                image_feature = image_feature.flatten(1, 2).flatten(2, 3) # [4096, 48, 48]
                other_images_mask = other_images_mask.flatten(0, 1).flatten(1, 2) # [48, 48]
                image_feature = unpad_image(image_feature, image_sizes[image_idx]) # [4096, 48, 48] assuming no unpadding
                other_images_mask = unpad_image(other_images_mask, image_sizes[image_idx]) # [48, 48] assuming no unpadding

                if image_newline is not None:
                    image_feature = torch.cat(
                        (
                            image_feature,
                            image_newline[:, None, None].expand(*image_feature.shape[:-1], 1).to(image_feature.dtype),
                        ),
                        dim=-1,
                    ) # [4096, 48, 49]
                    other_images_mask = torch.cat(
                        (
                            other_images_mask,
                            torch.ones(other_images_mask.shape[0], 1, device=other_images_mask.device),  # [42, 1]
                        ),
                        dim=-1,
                    ) # [48,49]

                # applying mask to reduce
                image_feature = image_feature.flatten(1, 2) # [4096, 48*49]
                other_images_mask = other_images_mask.flatten() # [48*49]
                reduced_image_feature = image_feature[:, other_images_mask.bool()] # [4096, reduced]
                reduced_image_feature = reduced_image_feature.transpose(0, 1) # [reduced, 4096]
                image_feature = torch.cat((reduced_base_image_feature, reduced_image_feature), dim=0) # [reduced+reduced, 4096]
            else:
                image_feature = image_feature[0]
                if image_newline is not None:
                    image_feature = torch.cat((image_feature, image_newline[None].to(image_feature)), dim=0)
            new_image_features.append(image_feature)
            feature_lens.append(image_feature.size(0))
        image_features = torch.cat(new_image_features, dim=0)
        feature_lens = torch.tensor(feature_lens, dtype=torch.long, device=image_features.device)
        return image_features, feature_lens
    
    # ========================= PruMerge+ =========================

    def token_prune_merge_advanced_plus(self, images, if_adaptive=True, reduction_ratio = 0.2):
        '''
        version 24/03/2024 using the spacially smapled tokens to supplement the pruned tokens
        '''
        # token_indix_list = []
        # token_indix_dict = {}

        #set hooks for extracting desired layer's k and q
        #set hooks for extracting desired layer's k and q
        hook_handle_k = self.vision_tower.vision_model.encoder.layers[23].self_attn.k_proj.register_forward_hook(hook_k)
        hook_handle_q = self.vision_tower.vision_model.encoder.layers[23].self_attn.q_proj.register_forward_hook(hook_q)

        #forward pass
        image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype), output_hidden_states=True)
        image_features = image_forward_outs.hidden_states[self.config.vision_feature_layer][:, 1:]

        B, N, C = image_features.shape

        #extract desired layer's k and q and remove hooks; calculate attention
        desired_layer_k = outputs["desired_k"]
        desired_layer_q = outputs["desired_q"]

        hook_handle_k.remove()
        hook_handle_q.remove()

        attn = (desired_layer_q @ desired_layer_k.transpose(-2, -1)) * C ** -0.5
        attn = F.softmax(attn, dim=-1)

        cls_attn = attn[:, 0, 1:]  

        if if_adaptive:
            reduction_ratio = outlier_dectection(cls_attn)#*3.5
        _, idx = torch.topk(cls_attn, int(N*reduction_ratio), dim=1, largest=True)  # [B, left_tokens] , sorted=True
        
        # # # print("idx: ", idx)
        if if_adaptive:
            step_length = int(1/reduction_ratio)
            arithmetic_sequence = torch.arange(0, 575, int(step_length/3)).to(device=self.device)
            original_tensor_1d = idx.flatten().to(device=self.device)
            filtered_sequence = torch.tensor([x for x in arithmetic_sequence if x not in original_tensor_1d]).to(device=self.device)
            concatenated_tensor = torch.cat((idx, filtered_sequence.unsqueeze(0)), dim=1)
            idx = concatenated_tensor
            # # print("idx_new: ", idx)
        else:
            # # this is for training
            step_length = int(1/reduction_ratio)
            new_idx = torch.zeros((idx.size(0), idx.size(1)*2), dtype=torch.long).to(device=self.device)
            for i in range(idx.size(0)):
                arithmetic_sequence = torch.arange(int(step_length/2), 575, int(step_length)).to(device=self.device)
                original_tensor_1d = idx[i].flatten().to(device=self.device)
                filtered_sequence = arithmetic_sequence
                # filtered_sequence = torch.tensor([x for x in arithmetic_sequence if x not in original_tensor_1d]).to(device=self.device)
                concatenated_tensor = torch.cat((original_tensor_1d, filtered_sequence), dim=0)
                new_idx[i] = concatenated_tensor
            idx = new_idx

        index = idx.unsqueeze(-1).expand(-1, -1, C)  # [B, left_tokens, C]

        Key_wo_cls = desired_layer_k[:, 1:]  # [B, N-1, C]

        x_others = torch.gather(image_features, dim=1, index=index)  # [B, left_tokens, C]
        x_others_attn = torch.gather(cls_attn, dim=1, index=idx)  
        Key_others = torch.gather(Key_wo_cls, dim=1, index=index)  # [B, left_tokens, C]
        compl = complement_idx(idx, N)  # [B, N-1-left_tokens]
        non_topk = torch.gather(image_features, dim=1, index=compl.unsqueeze(-1).expand(-1, -1, C))  # [B, N-1-left_tokens, C]
        non_topk_Key = torch.gather(Key_wo_cls, dim=1, index=compl.unsqueeze(-1).expand(-1, -1, C))
        non_topk_attn = torch.gather(cls_attn, dim=1, index=compl)  # [B, N-1-left_tokens]

        Key_others_norm = F.normalize(Key_others, p=2, dim=-1)
        non_topk_Key_norm = F.normalize(non_topk_Key, p=2, dim=-1)

        # cos_sim = torch.bmm(Key_others_norm, non_topk_Key_norm.transpose(1, 2)) # [B, left_tokens, N-1-left_tokens]

        # _, cluster_indices = torch.topk(cos_sim, k=4, dim=2, largest=True)

        B, left_tokens, C = x_others.size()
        updated_x_others = torch.zeros_like(x_others)

        for b in range(B):
            for i in range(left_tokens):
                key_others_norm = Key_others_norm[b,i,:].unsqueeze(0).unsqueeze(0)

                before_i_Key = Key_others_norm[b, :i, :].unsqueeze(0)  
                after_i_Key = Key_others_norm[b, i+1:, :].unsqueeze(0) 

                before_i_x_others = x_others[b, :i, :].unsqueeze(0)  
                after_i_x_others = x_others[b, i+1:, :].unsqueeze(0)   
                rest_x_others = torch.cat([before_i_x_others, after_i_x_others, non_topk[b,:,:].unsqueeze(0)], dim=1)   
                before_i_x_others_attn = x_others_attn[b, :i].unsqueeze(0)  
                after_i_x_others_attn = x_others_attn[b, i+1:].unsqueeze(0)  
                rest_x_others_attn = torch.cat([before_i_x_others_attn, after_i_x_others_attn, non_topk_attn[b,:].unsqueeze(0)], dim=1)  

                rest_Keys = torch.cat([before_i_Key, after_i_Key, non_topk_Key_norm[b,:,:].unsqueeze(0)], dim=1)
                cos_sim_matrix = torch.bmm(key_others_norm, rest_Keys.transpose(1, 2))

                _, cluster_indices = torch.topk(cos_sim_matrix, k=int(32), dim=2, largest=True)

                cluster_tokens = rest_x_others[:,cluster_indices.squeeze(),:]
                weights = rest_x_others_attn[:,cluster_indices.squeeze()].unsqueeze(-1)

                # update cluster centers
                weighted_avg = torch.sum(cluster_tokens * weights, dim=1) #/ torch.sum(weights)
                updated_center = x_others[b, i, :]  + weighted_avg 
                updated_x_others[b, i, :] = updated_center 
            
        extra_one_token = torch.sum(non_topk * non_topk_attn.unsqueeze(-1), dim=1, keepdim=True)  # [B, 1, C]
        updated_x_others = torch.cat([updated_x_others, extra_one_token],dim=1)
        image_features = updated_x_others
        return image_features

    # ========================= PruMerge =========================
    
    def token_prune_merge_advanced(self, images, if_adaptive=True, reduction_ratio = 0.2):
        '''
        version 10/03/2024 using the key*key matrix to calculate the cosine similarity
        '''
        # token_indix_list = []
        # token_indix_dict = {}

        #set hooks for extracting desired layer's k and q
        hook_handle_k = self.vision_tower.vision_model.encoder.layers[23].self_attn.k_proj.register_forward_hook(hook_k)
        hook_handle_q = self.vision_tower.vision_model.encoder.layers[23].self_attn.q_proj.register_forward_hook(hook_q)

        #forward pass
        image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype), output_hidden_states=True)
        image_features = image_forward_outs.hidden_states[self.config.vision_feature_layer][:, 1:]

        B, N, C = image_features.shape

        #extract desired layer's k and q and remove hooks; calculate attention
        desired_layer_k = outputs["desired_k"]
        desired_layer_q = outputs["desired_q"]

        hook_handle_k.remove()
        hook_handle_q.remove()

        attn = (desired_layer_q @ desired_layer_k.transpose(-2, -1)) * C ** -0.5
        attn = F.softmax(attn, dim=-1)

        cls_attn = attn[:, 0, 1:]  

        if if_adaptive:
            reduction_ratio = outlier_dectection(cls_attn)#*3.5
        _, idx = torch.topk(cls_attn, int(N*reduction_ratio), dim=1, largest=True)  # [B, left_tokens] , sorted=True
        index = idx.unsqueeze(-1).expand(-1, -1, C)  # [B, left_tokens, C]

        Key_wo_cls = desired_layer_k[:, 1:]  # [B, N-1, C]

        x_others = torch.gather(image_features, dim=1, index=index)  # [B, left_tokens, C]
        x_others_attn = torch.gather(cls_attn, dim=1, index=idx)  
        Key_others = torch.gather(Key_wo_cls, dim=1, index=index)  # [B, left_tokens, C]
        compl = complement_idx(idx, N)  # [B, N-1-left_tokens]
        non_topk = torch.gather(image_features, dim=1, index=compl.unsqueeze(-1).expand(-1, -1, C))  # [B, N-1-left_tokens, C]
        non_topk_Key = torch.gather(Key_wo_cls, dim=1, index=compl.unsqueeze(-1).expand(-1, -1, C))
        non_topk_attn = torch.gather(cls_attn, dim=1, index=compl)  # [B, N-1-left_tokens]

        Key_others_norm = F.normalize(Key_others, p=2, dim=-1)
        non_topk_Key_norm = F.normalize(non_topk_Key, p=2, dim=-1)

        # cos_sim = torch.bmm(Key_others_norm, non_topk_Key_norm.transpose(1, 2)) # [B, left_tokens, N-1-left_tokens]

        # _, cluster_indices = torch.topk(cos_sim, k=4, dim=2, largest=True)

        B, left_tokens, C = x_others.size()
        updated_x_others = torch.zeros_like(x_others)

        for b in range(B):
            for i in range(left_tokens):
                key_others_norm = Key_others_norm[b,i,:].unsqueeze(0).unsqueeze(0)

                before_i_Key = Key_others_norm[b, :i, :].unsqueeze(0)  
                after_i_Key = Key_others_norm[b, i+1:, :].unsqueeze(0) 

                before_i_x_others = x_others[b, :i, :].unsqueeze(0)  
                after_i_x_others = x_others[b, i+1:, :].unsqueeze(0)   
                rest_x_others = torch.cat([before_i_x_others, after_i_x_others, non_topk[b,:,:].unsqueeze(0)], dim=1)   
                before_i_x_others_attn = x_others_attn[b, :i].unsqueeze(0)  
                after_i_x_others_attn = x_others_attn[b, i+1:].unsqueeze(0)  
                rest_x_others_attn = torch.cat([before_i_x_others_attn, after_i_x_others_attn, non_topk_attn[b,:].unsqueeze(0)], dim=1)  

                rest_Keys = torch.cat([before_i_Key, after_i_Key, non_topk_Key_norm[b,:,:].unsqueeze(0)], dim=1)
                cos_sim_matrix = torch.bmm(key_others_norm, rest_Keys.transpose(1, 2))

                _, cluster_indices = torch.topk(cos_sim_matrix, k=int(32), dim=2, largest=True)


                cluster_tokens = rest_x_others[:,cluster_indices.squeeze(),:]
                weights = rest_x_others_attn[:,cluster_indices.squeeze()].unsqueeze(-1)

                # update cluster centers
                weighted_avg = torch.sum(cluster_tokens * weights, dim=1) #/ torch.sum(weights)
                updated_center = weighted_avg + x_others[b, i, :]  
                updated_x_others[b, i, :] = updated_center 
            

        extra_one_token = torch.sum(non_topk * non_topk_attn.unsqueeze(-1), dim=1, keepdim=True)  # [B, 1, C]
        updated_x_others = torch.cat([updated_x_others, extra_one_token],dim=1)
        image_features = updated_x_others
        return image_features

    @add_start_docstrings_to_model_forward(LLAVA_NEXT_INPUTS_DOCSTRING)
    @replace_return_docstrings(output_type=LlavaNextCausalLMOutputWithPast, config_class=_CONFIG_FOR_DOC)
    def forward(
        self,
        input_ids: torch.LongTensor = None,
        pixel_values: torch.FloatTensor = None,
        image_sizes: Optional[torch.LongTensor] = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[List[torch.FloatTensor]] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        vision_feature_layer: Optional[int] = None,
        vision_feature_select_strategy: Optional[str] = None,
        labels: Optional[torch.LongTensor] = None,
        use_cache: Optional[bool] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        return_dict: Optional[bool] = None,
    ) -> Union[Tuple, LlavaNextCausalLMOutputWithPast]:
        r"""
        Args:
            labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
                Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
                config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
                (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.

        Returns:

        Example:

        ```python
        >>> from PIL import Image
        >>> import requests
        >>> from transformers import AutoProcessor, LlavaNextForConditionalGeneration

        >>> model = LlavaNextForConditionalGeneration.from_pretrained("llava-hf/llava-v1.6-mistral-7b-hf")
        >>> processor = AutoProcessor.from_pretrained("llava-hf/llava-v1.6-mistral-7b-hf")

        >>> prompt = "[INST] <image>\nWhat is shown in this image? [/INST]"
        >>> url = "https://www.ilankelman.org/stopsigns/australia.jpg"
        >>> image = Image.open(requests.get(url, stream=True).raw)

        >>> inputs = processor(text=prompt, images=image, return_tensors="pt")

        >>> # Generate
        >>> generate_ids = model.generate(**inputs, max_length=30)
        >>> processor.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
        "[INST]  \nWhat is shown in this image? [/INST] The image appears to be a radar chart, which is a type of multi-dimensional plot (...)"
        ```"""

        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
        output_hidden_states = (
            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
        )
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
        vision_feature_layer = (
            vision_feature_layer if vision_feature_layer is not None else self.config.vision_feature_layer
        )
        vision_feature_select_strategy = (
            vision_feature_select_strategy
            if vision_feature_select_strategy is not None
            else self.config.vision_feature_select_strategy
        )

        if inputs_embeds is None:
            self.prefill_time = -1 # reset prefill time
            # 1. Extract the input embeddings
            # In case image_token_index is not in the embeddings (extra token but embedding don't have it)
            for_inputs_embeds_ids = input_ids.clone()
            for_inputs_embeds_ids[(input_ids == self.config.image_token_index)] = 0
            inputs_embeds = self.get_input_embeddings()(for_inputs_embeds_ids)

            # 2. Merge text and images
            if pixel_values is not None and input_ids.shape[1] != 1 and pixel_values.size(0) > 0:
                # ! infer image_num_patches from image_sizes
                image_num_patches = [
                    image_size_to_num_patches(
                        image_size=imsize,
                        grid_pinpoints=self.config.image_grid_pinpoints,
                        patch_size=self.config.vision_config.image_size,
                    )
                    for imsize in image_sizes
                ]
                # figure out if pixel_values is concatenated or stacked
                if pixel_values.dim() == 5:
                    # stacking when input is (batch_size, num_patches, num_channels, height, width)
                    _pixel_values_list = [
                        pix_val[:num_patch] for pix_val, num_patch in zip(pixel_values, image_num_patches)
                    ]
                    pixel_values = torch.cat(_pixel_values_list, dim=0)
                elif pixel_values.dim() != 4:
                    # otherwise has to be stacked from list of (num_patches, num_channels, height, width)
                    raise ValueError(f"pixel_values of shape {pixel_values.shape}, expect to be of 4 or 5 dimensions")

                ##################### PruMerge and PruMerge+ #####################

                use_prumerge_plus = True

                image_features = []
                feature_lens = []
                for img in pixel_values:
                    if use_prumerge_plus:
                        img_feature = self.token_prune_merge_advanced_plus(img.unsqueeze(0), if_adaptive=True)
                    else:
                        img_feature = self.token_prune_merge_advanced(img.unsqueeze(0), if_adaptive=True)
                    img_feature = self.multi_modal_projector(img_feature)
                    image_features.append(img_feature)
                image_features = torch.cat(image_features, dim=1)[0]
                feature_lens.append(image_features.shape[0])
                feature_lens = torch.tensor(feature_lens, dtype=torch.long, device=image_features.device)
                
                if use_prumerge_plus:
                    print(f"PruMerge+ number of visual tokens: {image_features.shape}")
                else:
                    print(f"PruMerge number of visual tokens: {image_features.shape}")

                inputs_embeds = inputs_embeds.to(image_features.dtype)
                inputs_embeds, attention_mask, position_ids, labels, self.vit_to_llm_mapping = self._merge_input_ids_with_image_features(
                    image_features,
                    feature_lens,
                    inputs_embeds,
                    input_ids,
                    attention_mask,
                    position_ids,
                    labels=labels,
                )

            # pixel_values is not None but is empty ---> text only cases
            elif pixel_values is not None and input_ids.shape[1] != 1 and pixel_values.size(0) == 0:
                # there are no images
                pass

            # In case input_ids.shape[1] == 1 & pixel_values==None & past_key_values != None, we are in the case of
            # generation with cache
            elif past_key_values is not None and pixel_values is not None and input_ids.shape[1] == 1:
                if self.prefill_time < 0:
                    self.prefill_time = time.time()
                first_layer_past_key_value = past_key_values[0][0][:, :, :, 0]

                # Sum all dimensions of head_dim (-2) to avoid random errors such as: https://github.com/huggingface/transformers/pull/28032#issuecomment-1863691941
                batch_index, non_attended_tokens = torch.where(first_layer_past_key_value.float().sum(-2) == 0)

                # Get the target length
                target_length = input_ids.shape[1]
                past_length = first_layer_past_key_value.shape[-1]

                extended_attention_mask = torch.ones(
                    (attention_mask.shape[0], past_length),
                    dtype=attention_mask.dtype,
                    device=attention_mask.device,
                )

                # Filter out only the tokens that can be un-attended, this can happen
                # if one uses Llava + Fused modules where the cache on the
                # first iteration is already big enough, or if one passes custom cache
                valid_indices = non_attended_tokens < extended_attention_mask.size(-1)
                new_batch_index = batch_index[valid_indices]
                new_non_attended_tokens = non_attended_tokens[valid_indices]

                # Zero-out the places where we don't need to attend
                extended_attention_mask[new_batch_index, new_non_attended_tokens] = 0

                attention_mask = torch.cat((extended_attention_mask, attention_mask[:, -target_length:]), dim=1)

                position_ids = torch.sum(attention_mask, dim=1).unsqueeze(-1) - 1

        outputs = self.language_model(
            attention_mask=attention_mask,
            position_ids=position_ids,
            past_key_values=past_key_values,
            inputs_embeds=inputs_embeds,
            use_cache=use_cache,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
        )

        logits = outputs[0]

        loss = None
        if labels is not None:
            # Shift so that tokens < n predict n
            if attention_mask is not None:
                shift_attention_mask = attention_mask[..., 1:]
                shift_logits = logits[..., :-1, :][shift_attention_mask.to(logits.device) != 0].contiguous()
                shift_labels = labels[..., 1:][shift_attention_mask.to(labels.device) != 0].contiguous()
            else:
                shift_logits = logits[..., :-1, :].contiguous()
                shift_labels = labels[..., 1:].contiguous()
            # Flatten the tokens
            loss_fct = nn.CrossEntropyLoss()
            loss = loss_fct(
                shift_logits.view(-1, shift_logits.size(-1)), shift_labels.view(-1).to(shift_logits.device)
            )

        if not return_dict:
            output = (logits,) + outputs[1:]
            return (loss,) + output if loss is not None else output

        return LlavaNextCausalLMOutputWithPast(
            loss=loss,
            logits=logits,
            past_key_values=outputs.past_key_values,
            hidden_states=outputs.hidden_states,
            attentions=outputs.attentions,
        )

    def prepare_inputs_for_generation(
        self,
        input_ids,
        past_key_values=None,
        inputs_embeds=None,
        pixel_values=None,
        image_sizes=None,
        attention_mask=None,
        **kwargs,
    ):
        if past_key_values is not None:
            if isinstance(past_key_values, Cache):
                cache_length = past_key_values.get_seq_length()
                past_length = past_key_values.seen_tokens
            else:
                cache_length = past_length = past_key_values[0][0].shape[2]

            # Keep only the unprocessed tokens:
            # 1 - If the length of the attention_mask exceeds the length of input_ids, then we are in a setting where
            # some of the inputs are exclusively passed as part of the cache (e.g. when passing input_embeds as
            # input)
            if attention_mask is not None and attention_mask.shape[1] > input_ids.shape[1]:
                input_ids = input_ids[:, -(attention_mask.shape[1] - past_length) :]
            # 2 - If the past_length is smaller than input_ids', then input_ids holds all input tokens. We can discard
            # input_ids based on the past_length.
            elif past_length < input_ids.shape[1]:
                input_ids = input_ids[:, past_length:]
            # 3 - Otherwise (past_length >= input_ids.shape[1]), let's assume input_ids only has unprocessed tokens.
            elif self.config.image_token_index in input_ids:
                input_ids = input_ids[:, input_ids.shape[1] - 1 :]
            # If the cache has seen more tokens than it can hold, then the cache has a size limit. Let's discard the
            # older attention values, as their corresponding values are not part of the input.
            if cache_length < past_length and attention_mask is not None:
                attention_mask = attention_mask[:, -(cache_length + input_ids.shape[1]) :]

        position_ids = kwargs.get("position_ids", None)
        if attention_mask is not None and position_ids is None:
            # create position_ids on the fly for batch generation
            position_ids = attention_mask.long().cumsum(-1) - 1
            position_ids.masked_fill_(attention_mask == 0, 1)
            if past_key_values:
                position_ids = position_ids[:, -input_ids.shape[1] :]

        # if `inputs_embeds` are passed, we only want to use them in the 1st generation step
        if inputs_embeds is not None and past_key_values is None:
            model_inputs = {"inputs_embeds": inputs_embeds}
        else:
            model_inputs = {"input_ids": input_ids}

        model_inputs.update(
            {
                "position_ids": position_ids,
                "past_key_values": past_key_values,
                "use_cache": kwargs.get("use_cache"),
                "attention_mask": attention_mask,
                "pixel_values": pixel_values,
                "image_sizes": image_sizes,
            }
        )
        return model_inputs

    # Copied from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration._reorder_cache
    def _reorder_cache(self, *args, **kwargs):
        return self.language_model._reorder_cache(*args, **kwargs)