Upload 6 files

configuration_omchat.py

@@ -0,0 +1,198 @@
+
+from transformers import LlamaConfig, PretrainedConfig
+from transformers.utils import logging
+from transformers import Qwen2Config, Qwen2Model, Qwen2ForCausalLM, AutoConfig, AutoModelForCausalLM
+
+
+logger = logging.get_logger(__name__)
+
+class InternVisionConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`InternVisionModel`]. It is used to
+    instantiate a vision encoder according to the specified arguments, defining the model architecture.
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+    Args:
+        num_channels (`int`, *optional*, defaults to 3):
+            Number of color channels in the input images (e.g., 3 for RGB).
+        patch_size (`int`, *optional*, defaults to 14):
+            The size (resolution) of each patch.
+        image_size (`int`, *optional*, defaults to 224):
+            The size (resolution) of each image.
+        qkv_bias (`bool`, *optional*, defaults to `False`):
+            Whether to add a bias to the queries and values in the self-attention layers.
+        hidden_size (`int`, *optional*, defaults to 3200):
+            Dimensionality of the encoder layers and the pooler layer.
+        num_attention_heads (`int`, *optional*, defaults to 25):
+            Number of attention heads for each attention layer in the Transformer encoder.
+        intermediate_size (`int`, *optional*, defaults to 12800):
+            Dimensionality of the "intermediate" (i.e., feed-forward) layer in the Transformer encoder.
+        qk_normalization (`bool`, *optional*, defaults to `True`):
+            Whether to normalize the queries and keys in the self-attention layers.
+        num_hidden_layers (`int`, *optional*, defaults to 48):
+            Number of hidden layers in the Transformer encoder.
+        use_flash_attn (`bool`, *optional*, defaults to `True`):
+            Whether to use flash attention mechanism.
+        hidden_act (`str` or `function`, *optional*, defaults to `"gelu"`):
+            The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`,
+            `"relu"`, `"selu"` and `"gelu_new"` ``"gelu"` are supported.
+        layer_norm_eps (`float`, *optional*, defaults to 1e-6):
+            The epsilon used by the layer normalization layers.
+        dropout (`float`, *optional*, defaults to 0.0):
+            The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.
+        drop_path_rate (`float`, *optional*, defaults to 0.0):
+            Dropout rate for stochastic depth.
+        attention_dropout (`float`, *optional*, defaults to 0.0):
+            The dropout ratio for the attention probabilities.
+        initializer_range (`float`, *optional*, defaults to 0.02):
+            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
+        initializer_factor (`float`, *optional*, defaults to 0.1):
+            A factor for layer scale.
+    """
+
+    model_type = 'intern_vit_6b'
+
+    def __init__(
+            self,
+            num_channels=3,
+            patch_size=14,
+            image_size=448,
+            qkv_bias=False,
+            hidden_size=3200,
+            num_attention_heads=25,
+            intermediate_size=12800,
+            qk_normalization=True,
+            num_hidden_layers=45,
+            use_flash_attn=True,
+            hidden_act='gelu',
+            layer_norm_eps=1e-6,
+            dropout=0.0,
+            drop_path_rate=0.0,
+            attention_dropout=0.0,
+            initializer_range=1e-10,
+            initializer_factor=0.1,
+            **kwargs,
+    ):
+        super().__init__(**kwargs)
+
+        self.hidden_size = hidden_size
+        self.intermediate_size = intermediate_size
+        self.dropout = dropout
+        self.drop_path_rate = drop_path_rate
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.num_channels = num_channels
+        self.patch_size = patch_size
+        self.image_size = image_size
+        self.initializer_range = initializer_range
+        self.initializer_factor = initializer_factor
+        self.attention_dropout = attention_dropout
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+        self.qkv_bias = qkv_bias
+        self.qk_normalization = qk_normalization
+        self.use_flash_attn = use_flash_attn
+
+
+class OmChatConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`OmChatForConditionalGeneration`]. It is used to instantiate an
+    Llava-NeXT model according to the specified arguments, defining the model architecture. Instantiating a configuration
+    with the defaults will yield a similar configuration to that of the [llava-hf/llava-v1.6-mistral-7b-hf](https://huggingface.co/llava-hf/llava-v1.6-mistral-7b-hf)
+    model.
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+    Args:
+        vision_config (`Union[AutoConfig, dict]`,  *optional*, defaults to `CLIPVisionConfig`):
+            The config object or dictionary of the vision backbone.
+        text_config (`Union[AutoConfig, dict]`, *optional*, defaults to `LlamaConfig`):
+            The config object or dictionary of the text backbone.
+        ignore_index (`int`, *optional*, defaults to -100):
+            The ignore index for the loss function.
+        image_token_index (`int`, *optional*, defaults to 32000):
+            The image token index to encode the image prompt.
+        projector_hidden_act (`str`, *optional*, defaults to `"gelu"`):
+            The activation function used by the multimodal projector.
+        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.
+        vision_feature_layer (`int`, *optional*, defaults to -2):
+            The index of the layer to select the vision feature.
+        image_grid_pinpoints (`List`, *optional*, defaults to `[[336, 672], [672, 336], [672, 672], [1008, 336], [336, 1008]]`):
+            A list of possible resolutions to use for processing high resolution images. Each item in the list should be a tuple or list
+            of the form `(height, width)`.
+        tie_word_embeddings (`bool`, *optional*, defaults to `False`):
+            Whether the model's input and output word embeddings should be tied.
+
+    Example:
+
+    ```python
+    >>> from transformers import OmChatForConditionalGeneration, OmChatConfig, CLIPVisionConfig, LlamaConfig
+
+    >>> # Initializing a CLIP-vision config
+    >>> vision_config = CLIPVisionConfig()
+
+    >>> # Initializing a Llama config
+    >>> text_config = LlamaConfig()
+
+    >>> # Initializing a Llava-Next llava-hf/llava-v1.6-mistral-7b-hf style configuration
+    >>> configuration = OmChatConfig(vision_config, text_config)
+
+    >>> # Initializing a model from the llava-hf/llava-v1.6-mistral-7b-hf style configuration
+    >>> model = OmChatForConditionalGeneration(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "omchat"
+    is_composition = False
+
+    def __init__(
+        self,
+        vision_config=None,
+        text_config=None,
+        ignore_index=-100,
+        image_token_index=32000,
+        projector_hidden_act="gelu",
+        vision_feature_select_strategy="default",
+        vision_feature_layer=-1,
+        image_grid_pinpoints=None,
+        tie_word_embeddings=False,
+        **kwargs,
+    ):
+        self.ignore_index = ignore_index
+        self.image_token_index = image_token_index
+        self.projector_hidden_act = projector_hidden_act
+
+        if vision_feature_select_strategy not in ["default", "full"]:
+            raise ValueError(
+                "vision_feature_select_strategy should be one of 'default', 'full'."
+                f"Got: {vision_feature_select_strategy}"
+            )
+
+        self.vision_feature_select_strategy = vision_feature_select_strategy
+        self.vision_feature_layer = vision_feature_layer
+        image_grid_pinpoints = (
+            image_grid_pinpoints
+            if image_grid_pinpoints is not None
+            else [[336, 672], [672, 336], [672, 672], [1008, 336], [336, 1008]]
+        )
+        self.image_grid_pinpoints = image_grid_pinpoints
+
+        if isinstance(vision_config, dict):
+
+            vision_config = InternVisionConfig(**vision_config)
+        self.vision_config = vision_config
+
+        if isinstance(text_config, dict):
+            text_config = Qwen2Config(**text_config)
+
+        self.text_config = text_config
+
+        super().__init__(tie_word_embeddings=tie_word_embeddings, **kwargs)

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "transformers_version": "4.41.2"
+}

image_processing_omchat.py

@@ -0,0 +1,733 @@
+
+import math
+from typing import Dict, Iterable, List, Optional, Tuple, Union
+
+import numpy as np
+
+from transformers.image_processing_utils import BaseImageProcessor, BatchFeature, get_size_dict, select_best_resolution
+from transformers.image_transforms import (
+    PaddingMode,
+    convert_to_rgb,
+    get_resize_output_image_size,
+    pad,
+    resize,
+    to_channel_dimension_format,
+)
+from transformers.image_utils import (
+    OPENAI_CLIP_MEAN,
+    OPENAI_CLIP_STD,
+    ChannelDimension,
+    ImageInput,
+    PILImageResampling,
+    get_image_size,
+    infer_channel_dimension_format,
+    is_scaled_image,
+    is_valid_image,
+    make_list_of_images,
+    to_numpy_array,
+    valid_images,
+    validate_preprocess_arguments,
+)
+from transformers.utils import TensorType, is_vision_available, logging
+
+
+logger = logging.get_logger(__name__)
+
+
+if is_vision_available():
+    from PIL import Image
+
+
+def make_batched_images(images) -> List[List[ImageInput]]:
+    """
+    Accepts images in list or nested list format, and makes a list of images for preprocessing.
+
+    Args:
+        images (`Union[List[List[ImageInput]], List[ImageInput], ImageInput]`):
+            The input image.
+
+    Returns:
+        list: A list of images.
+    """
+    if isinstance(images, (list, tuple)) and isinstance(images[0], (list, tuple)) and is_valid_image(images[0][0]):
+        return [img for img_list in images for img in img_list]
+
+    elif isinstance(images, (list, tuple)) and is_valid_image(images[0]):
+        return images
+
+    elif is_valid_image(images):
+        return [images]
+
+    raise ValueError(f"Could not make batched video from {images}")
+
+
+def divide_to_patches(image: np.array, patch_size: int, input_data_format) -> List[np.array]:
+    """
+    Divides an image into patches of a specified size.
+
+    Args:
+        image (`np.array`):
+            The input image.
+        patch_size (`int`):
+            The size of each patch.
+        input_data_format (`ChannelDimension` or `str`):
+            The channel dimension format of the input image.
+
+    Returns:
+        list: A list of np.array representing the patches.
+    """
+    patches = []
+    height, width = get_image_size(image, channel_dim=input_data_format)
+    for i in range(0, height, patch_size):
+        for j in range(0, width, patch_size):
+            if input_data_format == ChannelDimension.LAST:
+                patch = image[i : i + patch_size, j : j + patch_size]
+            else:
+                patch = image[:, i : i + patch_size, j : j + patch_size]
+            patches.append(patch)
+
+    return patches
+
+
+def expand_to_square(image: np.array, background_color, input_data_format) -> np.array:
+    """
+    Expands an image to a square by adding a background color.
+    """
+
+    height, width = get_image_size(image, channel_dim=input_data_format)
+    if width == height:
+        return image
+    elif width > height:
+        result = np.ones((width, width, image.shape[2]), dtype=image.dtype) * background_color
+        result[(width - height) // 2 : (width - height) // 2 + height, :] = image
+        return result
+    else:
+        result = np.ones((height, height, image.shape[2]), dtype=image.dtype) * background_color
+        result[:, (height - width) // 2 : (height - width) // 2 + width] = image
+        return result
+
+
+def _get_patch_output_size(image, target_resolution, input_data_format):
+    original_height, original_width = get_image_size(image, channel_dim=input_data_format)
+    target_height, target_width = target_resolution
+
+    scale_w = target_width / original_width
+    scale_h = target_height / original_height
+
+    if scale_w < scale_h:
+        new_width = target_width
+        new_height = min(math.ceil(original_height * scale_w), target_height)
+    else:
+        new_height = target_height
+        new_width = min(math.ceil(original_width * scale_h), target_width)
+
+    return new_height, new_width
+
+
+class OmChatImageProcessor(BaseImageProcessor):
+    r"""
+    Constructs a LLaVa-NeXT image processor. Based on [`CLIPImageProcessor`] with incorporation of additional techniques
+    for processing high resolution images as explained in the [LLaVa paper](https://arxiv.org/abs/2310.03744).
+
+    Args:
+        do_resize (`bool`, *optional*, defaults to `True`):
+            Whether to resize the image's (height, width) dimensions to the specified `size`. Can be overridden by
+            `do_resize` in the `preprocess` method.
+        size (`Dict[str, int]` *optional*, defaults to `{"shortest_edge": 224}`):
+            Size of the image after resizing. The shortest edge of the image is resized to size["shortest_edge"], with
+            the longest edge resized to keep the input aspect ratio. Can be overridden by `size` in the `preprocess`
+            method.
+        image_grid_pinpoints (`List` *optional*, defaults to `[[896, 448], [448, 896], [896, 896], [448, 1344], [1344, 448]]`):
+            A list of possible resolutions to use for processing high resolution images. The best resolution is selected
+            based on the original size of the image. Can be overridden by `image_grid_pinpoints` in the `preprocess`
+            method.
+        resample (`PILImageResampling`, *optional*, defaults to `Resampling.BICUBIC`):
+            Resampling filter to use if resizing the image. Can be overridden by `resample` in the `preprocess` method.
+        do_center_crop (`bool`, *optional*, defaults to `True`):
+            Whether to center crop the image to the specified `crop_size`. Can be overridden by `do_center_crop` in the
+            `preprocess` method.
+        crop_size (`Dict[str, int]` *optional*, defaults to 224):
+            Size of the output image after applying `center_crop`. Can be overridden by `crop_size` in the `preprocess`
+            method.
+        do_rescale (`bool`, *optional*, defaults to `True`):
+            Whether to rescale the image by the specified scale `rescale_factor`. Can be overridden by `do_rescale` in
+            the `preprocess` method.
+        rescale_factor (`int` or `float`, *optional*, defaults to `1/255`):
+            Scale factor to use if rescaling the image. Can be overridden by `rescale_factor` in the `preprocess`
+            method.
+        do_normalize (`bool`, *optional*, defaults to `True`):
+            Whether to normalize the image. Can be overridden by `do_normalize` in the `preprocess` method.
+        image_mean (`float` or `List[float]`, *optional*, defaults to `[0.48145466, 0.4578275, 0.40821073]`):
+            Mean to use if normalizing the image. This is a float or list of floats the length of the number of
+            channels in the image. Can be overridden by the `image_mean` parameter in the `preprocess` method.
+        image_std (`float` or `List[float]`, *optional*, defaults to `[0.26862954, 0.26130258, 0.27577711]`):
+            Standard deviation to use if normalizing the image. This is a float or list of floats the length of the
+            number of channels in the image. Can be overridden by the `image_std` parameter in the `preprocess` method.
+            Can be overridden by the `image_std` parameter in the `preprocess` method.
+        do_pad (`bool`, *optional*, defaults to `True`):
+                Whether to pad the image. If `True`, will pad the patch dimension of the images in the batch to the largest
+                number of patches in the batch. Padding will be applied to the bottom and right with zeros.
+        do_convert_rgb (`bool`, *optional*, defaults to `True`):
+            Whether to convert the image to RGB.
+    """
+
+    model_input_names = ["pixel_values"]
+
+    def __init__(
+        self,
+        do_resize: bool = True,
+        size: Dict[str, int] = None,
+        image_grid_pinpoints: List = None,
+        resample: PILImageResampling = PILImageResampling.BICUBIC,
+        do_center_crop: bool = True,
+        crop_size: Dict[str, int] = None,
+        do_rescale: bool = True,
+        rescale_factor: Union[int, float] = 1 / 255,
+        do_normalize: bool = True,
+        image_mean: Optional[Union[float, List[float]]] = [0.485, 0.456, 0.406],
+        image_std: Optional[Union[float, List[float]]] = [0.229, 0.224, 0.225],
+        do_convert_rgb: bool = True,
+        **kwargs,
+    ) -> None:
+        super().__init__(**kwargs)
+        size = size if size is not None else {"shortest_edge": 448}
+        size = get_size_dict(size, default_to_square=False)
+        image_grid_pinpoints = (
+            image_grid_pinpoints
+            if image_grid_pinpoints is not None
+            else [[448, 896], [896, 448], [896, 896], [1344, 448], [448, 1344],[1344, 1344]]
+        )
+        crop_size = crop_size if crop_size is not None else {"height": 448, "width": 448}
+        crop_size = get_size_dict(crop_size, default_to_square=True, param_name="crop_size")
+
+        self.do_resize = do_resize
+        self.size = size
+        self.image_grid_pinpoints = image_grid_pinpoints
+        self.resample = resample
+        self.do_center_crop = do_center_crop
+        self.crop_size = crop_size
+        self.do_rescale = do_rescale
+        self.rescale_factor = rescale_factor
+        self.do_normalize = do_normalize
+        self.image_mean = image_mean if image_mean is not None else OPENAI_CLIP_MEAN
+        self.image_std = image_std if image_std is not None else OPENAI_CLIP_STD
+        self.do_convert_rgb = do_convert_rgb
+
+    # Copied from transformers.models.clip.image_processing_clip.CLIPImageProcessor.resize with CLIP->LLaVa
+    def resize(
+        self,
+        image: np.ndarray,
+        size: Dict[str, int],
+        resample: PILImageResampling = PILImageResampling.BICUBIC,
+        data_format: Optional[Union[str, ChannelDimension]] = None,
+        input_data_format: Optional[Union[str, ChannelDimension]] = None,
+        **kwargs,
+    ) -> np.ndarray:
+        """
+        Resize an image. The shortest edge of the image is resized to size["shortest_edge"], with the longest edge
+        resized to keep the input aspect ratio.
+
+        Args:
+            image (`np.ndarray`):
+                Image to resize.
+            size (`Dict[str, int]`):
+                Size of the output image.
+            resample (`PILImageResampling`, *optional*, defaults to `PILImageResampling.BICUBIC`):
+                Resampling filter to use when resiizing the image.
+            data_format (`str` or `ChannelDimension`, *optional*):
+                The channel dimension format of the image. If not provided, it will be the same as the input image.
+            input_data_format (`ChannelDimension` or `str`, *optional*):
+                The channel dimension format of the input image. If not provided, it will be inferred.
+        """
+        default_to_square = True
+        if "shortest_edge" in size:
+            size = size["shortest_edge"]
+            default_to_square = False
+        elif "height" in size and "width" in size:
+            size = (size["height"], size["width"])
+        else:
+            raise ValueError("Size must contain either 'shortest_edge' or 'height' and 'width'.")
+
+        output_size = get_resize_output_image_size(
+            image,
+            size=size,
+            default_to_square=default_to_square,
+            input_data_format=input_data_format,
+        )
+
+        return resize(
+            image,
+            size=output_size,
+            resample=resample,
+            data_format=data_format,
+            input_data_format=input_data_format,
+            **kwargs,
+        )
+
+    def pad(
+        self,
+        image: np.ndarray,
+        padding: Union[int, Tuple[int, int], Iterable[Tuple[int, int]]],
+        mode: PaddingMode = PaddingMode.CONSTANT,
+        constant_values: Union[float, Iterable[float]] = 0.0,
+        data_format: Optional[Union[str, ChannelDimension]] = None,
+        input_data_format: Optional[Union[str, ChannelDimension]] = None,
+    ) -> np.ndarray:
+        """
+        Pads the `image` with the specified `padding` and `mode`. Padding can be in the (`height`, `width`)
+        dimension of in the (`num_patches`) dimension. In the second case an iterable if tuples is expected
+        as input.
+
+        Args:
+            image (`np.ndarray`):
+                The image to pad.
+            padding (`int` or `Tuple[int, int]` or `Iterable[Tuple[int, int]]`):
+                Padding to apply to the edges of the height, width axes. Can be one of three formats:
+                - `((before_height, after_height), (before_width, after_width))` unique pad widths for each axis.
+                - `((before, after),)` yields same before and after pad for height and width.
+                - `(pad,)` or int is a shortcut for before = after = pad width for all axes.
+            mode (`PaddingMode`):
+                The padding mode to use. Can be one of:
+                    - `"constant"`: pads with a constant value.
+                    - `"reflect"`: pads with the reflection of the vector mirrored on the first and last values of the
+                    vector along each axis.
+                    - `"replicate"`: pads with the replication of the last value on the edge of the array along each axis.
+                    - `"symmetric"`: pads with the reflection of the vector mirrored along the edge of the array.
+            constant_values (`float` or `Iterable[float]`, *optional*):
+                The value to use for the padding if `mode` is `"constant"`.
+            data_format (`str` or `ChannelDimension`, *optional*):
+                The channel dimension format for the output image. Can be one of:
+                    - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
+                    - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
+                If unset, will use same as the input image.
+            input_data_format (`str` or `ChannelDimension`, *optional*):
+                The channel dimension format for the input image. Can be one of:
+                    - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
+                    - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
+                If unset, will use the inferred format of the input image.
+
+        Returns:
+            `np.ndarray`: The padded image.
+
+        """
+
+        # call the general `pad` if padding on `height/width`, otherwise it's the `num_patched` dim
+        if isinstance(padding, int) or len(padding) != 4:
+            return pad(image, padding, mode, constant_values, data_format, input_data_format)
+
+        if input_data_format is None:
+            input_data_format = infer_channel_dimension_format(image)
+        if mode == PaddingMode.CONSTANT:
+            image = np.pad(image, padding, mode="constant", constant_values=constant_values)
+        elif mode == PaddingMode.REFLECT:
+            image = np.pad(image, padding, mode="reflect")
+        elif mode == PaddingMode.REPLICATE:
+            image = np.pad(image, padding, mode="edge")
+        elif mode == PaddingMode.SYMMETRIC:
+            image = np.pad(image, padding, mode="symmetric")
+        else:
+            raise ValueError(f"Invalid padding mode: {mode}")
+        image = (
+            to_channel_dimension_format(image, data_format, input_data_format) if data_format is not None else image
+        )
+        return image
+
+    def _preprocess(
+        self,
+        images: ImageInput,
+        do_resize: bool = None,
+        size: Dict[str, int] = None,
+        resample: PILImageResampling = None,
+        do_center_crop: bool = None,
+        crop_size: int = None,
+        do_rescale: bool = None,
+        rescale_factor: float = None,
+        do_normalize: bool = None,
+        image_mean: Optional[Union[float, List[float]]] = None,
+        image_std: Optional[Union[float, List[float]]] = None,
+        data_format: Optional[ChannelDimension] = ChannelDimension.FIRST,
+        input_data_format: Optional[Union[str, ChannelDimension]] = None,
+    ) -> Image.Image:
+        """
+        Preprocess an image or batch of images. Copy of the `preprocess` method from `CLIPImageProcessor`.
+
+        Args:
+            images (`ImageInput`):
+                Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If
+                passing in images with pixel values between 0 and 1, set `do_rescale=False`.
+            do_resize (`bool`, *optional*, defaults to `self.do_resize`):
+                Whether to resize the image.
+            size (`Dict[str, int]`, *optional*, defaults to `self.size`):
+                Size of the image after resizing. Shortest edge of the image is resized to size["shortest_edge"], with
+                the longest edge resized to keep the input aspect ratio.
+            resample (`int`, *optional*, defaults to `self.resample`):
+                Resampling filter to use if resizing the image. This can be one of the enum `PILImageResampling`. Only
+                has an effect if `do_resize` is set to `True`.
+            do_center_crop (`bool`, *optional*, defaults to `self.do_center_crop`):
+                Whether to center crop the image.
+            crop_size (`Dict[str, int]`, *optional*, defaults to `self.crop_size`):
+                Size of the center crop. Only has an effect if `do_center_crop` is set to `True`.
+            do_rescale (`bool`, *optional*, defaults to `self.do_rescale`):
+                Whether to rescale the image.
+            rescale_factor (`float`, *optional*, defaults to `self.rescale_factor`):
+                Rescale factor to rescale the image by if `do_rescale` is set to `True`.
+            do_normalize (`bool`, *optional*, defaults to `self.do_normalize`):
+                Whether to normalize the image.
+            image_mean (`float` or `List[float]`, *optional*, defaults to `self.image_mean`):
+                Image mean to use for normalization. Only has an effect if `do_normalize` is set to `True`.
+            image_std (`float` or `List[float]`, *optional*, defaults to `self.image_std`):
+                Image standard deviation to use for normalization. Only has an effect if `do_normalize` is set to
+                `True`.
+            data_format (`ChannelDimension` or `str`, *optional*, defaults to `ChannelDimension.FIRST`):
+                The channel dimension format for the output image. Can be one of:
+                - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
+                - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
+                - Unset: Use the channel dimension format of the input image.
+            input_data_format (`ChannelDimension` or `str`, *optional*):
+                The channel dimension format for the input image. If unset, the channel dimension format is inferred
+                from the input image. Can be one of:
+                - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
+                - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
+                - `"none"` or `ChannelDimension.NONE`: image in (height, width) format.
+        """
+        images = make_list_of_images(images)
+
+        if do_resize:
+            images = [
+                self.resize(image=image, size=size, resample=resample, input_data_format=input_data_format)
+                for image in images
+            ]
+
+        if do_center_crop:
+            images = [
+                self.center_crop(image=image, size=crop_size, input_data_format=input_data_format) for image in images
+            ]
+
+        if do_rescale:
+            images = [
+                self.rescale(image=image, scale=rescale_factor, input_data_format=input_data_format)
+                for image in images
+            ]
+
+        if do_normalize:
+            images = [
+                self.normalize(image=image, mean=image_mean, std=image_std, input_data_format=input_data_format)
+                for image in images
+            ]
+
+        images = [
+            to_channel_dimension_format(image, data_format, input_channel_dim=input_data_format) for image in images
+        ]
+
+        return images
+
+    def _resize_for_patching(
+        self, image: np.array, target_resolution: tuple, resample, input_data_format: ChannelDimension
+    ) -> np.array:
+        """
+        Resizes an image to a target resolution while maintaining aspect ratio.
+
+        Args:
+            image (np.array):
+                The input image.
+            target_resolution (tuple):
+                The target resolution (height, width) of the image.
+            resample (`PILImageResampling`):
+                Resampling filter to use if resizing the image.
+            input_data_format (`ChannelDimension` or `str`):
+                The channel dimension format of the input image.
+
+        Returns:
+            np.array: The resized and padded image.
+        """
+        new_height, new_width = _get_patch_output_size(image, target_resolution, input_data_format)
+
+        # Resize the image
+        resized_image = resize(image, (new_height, new_width), resample=resample, input_data_format=input_data_format)
+
+        return resized_image
+
+    def _pad_for_patching(
+        self, image: np.array, target_resolution: tuple, input_data_format: ChannelDimension
+    ) -> np.array:
+        """
+        Pad an image to a target resolution while maintaining aspect ratio.
+        """
+        target_height, target_width = target_resolution
+        new_height, new_width = _get_patch_output_size(image, target_resolution, input_data_format)
+
+        paste_x = (target_width - new_width) // 2
+        paste_y = (target_height - new_height) // 2
+
+        padded_image = self.pad(image, padding=((paste_y, paste_y), (paste_x, paste_x)))
+
+        return padded_image
+
+    def get_image_patches(
+        self,
+        image: np.array,
+        grid_pinpoints,
+        size: tuple,
+        patch_size: int,
+        resample: PILImageResampling,
+        data_format: ChannelDimension,
+        input_data_format: ChannelDimension,
+    ) -> List[np.array]:
+        """
+        Process an image with variable resolutions by dividing it into patches.
+
+        Args:
+            image (np.array):
+                The input image to be processed.
+            grid_pinpoints (List):
+                A string representation of a list of possible resolutions.
+            size (`tuple`):
+                Size to resize the original image to.
+            patch_size (`int`):
+                Size of the patches to divide the image into.
+            resample (`PILImageResampling`):
+                Resampling filter to use if resizing the image.
+            data_format (`ChannelDimension` or `str`):
+                The channel dimension format for the output image.
+            input_data_format (`ChannelDimension` or `str`):
+                The channel dimension format of the input image.
+
+        Returns:
+            List[np.array]: A list of NumPy arrays containing the processed image patches.
+        """
+        if not isinstance(grid_pinpoints, list):
+            raise TypeError("grid_pinpoints must be a list of possible resolutions.")
+
+        possible_resolutions = grid_pinpoints
+
+        image_size = get_image_size(image, channel_dim=input_data_format)
+        best_resolution = select_best_resolution(image_size, possible_resolutions)
+        resized_image = self._resize_for_patching(
+            image, best_resolution, resample=resample, input_data_format=input_data_format
+        )
+        padded_image = self._pad_for_patching(resized_image, best_resolution, input_data_format=input_data_format)
+
+        patches = divide_to_patches(padded_image, patch_size=patch_size, input_data_format=input_data_format)
+
+        # make sure that all patches are in the input data format
+        patches = [
+            to_channel_dimension_format(patch, channel_dim=data_format, input_channel_dim=input_data_format)
+            for patch in patches
+        ]
+
+        resized_original_image = resize(
+            image,
+            size=size,
+            resample=resample,
+            data_format=data_format,
+            input_data_format=input_data_format,
+        )
+
+        image_patches = [resized_original_image] + patches
+
+        return image_patches
+
+    def _pad_for_batching(
+        self,
+        pixel_values: List[np.ndarray],
+        data_format: Optional[Union[str, ChannelDimension]] = None,
+        input_data_format: Optional[Union[str, ChannelDimension]] = None,
+    ):
+        """
+        Pads images on the `num_of_patches` dimension with zeros to form a batch of same number of patches.
+
+        Args:
+            pixel_values (`List[np.ndarray]`):
+                An array of pixel values of each images of shape (`batch_size`, `num_patches`, `image_in_3D`)
+            data_format (`str` or `ChannelDimension`, *optional*):
+                The channel dimension format for the output image. Can be one of:
+                    - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
+                    - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
+                If unset, will use same as the input image.
+            input_data_format (`str` or `ChannelDimension`, *optional*):
+                The channel dimension format for the input image. Can be one of:
+                    - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
+                    - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
+                If unset, will use the inferred format of the input image.
+
+        Returns:
+            List[`np.ndarray`]: The padded images.
+        """
+        max_patch = max(len(x) for x in pixel_values)
+        pixel_values = [
+            self.pad(
+                image,
+                padding=((0, max_patch - image.shape[0]), (0, 0), (0, 0), (0, 0)),
+                data_format=data_format,
+                input_data_format=input_data_format,
+            )
+            for image in pixel_values
+        ]
+
+        return pixel_values
+
+    def preprocess(
+        self,
+        images: ImageInput,
+        do_resize: bool = None,
+        size: Dict[str, int] = None,
+        image_grid_pinpoints: List = None,
+        resample: PILImageResampling = None,
+        do_center_crop: bool = None,
+        crop_size: int = None,
+        do_rescale: bool = None,
+        rescale_factor: float = None,
+        do_normalize: bool = None,
+        image_mean: Optional[Union[float, List[float]]] = None,
+        image_std: Optional[Union[float, List[float]]] = None,
+        do_convert_rgb: bool = None,
+        return_tensors: Optional[Union[str, TensorType]] = None,
+        data_format: Optional[ChannelDimension] = ChannelDimension.FIRST,
+        input_data_format: Optional[Union[str, ChannelDimension]] = None,
+    ):
+        """
+        Args:
+            images (`ImageInput`):
+                Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If
+                passing in images with pixel values between 0 and 1, set `do_rescale=False`.
+            do_resize (`bool`, *optional*, defaults to `self.do_resize`):
+                Whether to resize the image.
+            size (`Dict[str, int]`, *optional*, defaults to `self.size`):
+                Size of the image after resizing. Shortest edge of the image is resized to size["shortest_edge"], with
+                the longest edge resized to keep the input aspect ratio.
+            image_grid_pinpoints (`List` *optional*, defaults to `self.image_grid_pinpoints`):
+                A list of possible resolutions to use for processing high resolution images. The best resolution is
+                selected based on the original size of the image.
+            resample (`int`, *optional*, defaults to `self.resample`):
+                Resampling filter to use if resizing the image. This can be one of the enum `PILImageResampling`. Only
+                has an effect if `do_resize` is set to `True`.
+            do_center_crop (`bool`, *optional*, defaults to `self.do_center_crop`):
+                Whether to center crop the image.
+            crop_size (`Dict[str, int]`, *optional*, defaults to `self.crop_size`):
+                Size of the center crop. Only has an effect if `do_center_crop` is set to `True`.
+            do_rescale (`bool`, *optional*, defaults to `self.do_rescale`):
+                Whether to rescale the image.
+            rescale_factor (`float`, *optional*, defaults to `self.rescale_factor`):
+                Rescale factor to rescale the image by if `do_rescale` is set to `True`.
+            do_normalize (`bool`, *optional*, defaults to `self.do_normalize`):
+                Whether to normalize the image.
+            image_mean (`float` or `List[float]`, *optional*, defaults to `self.image_mean`):
+                Image mean to use for normalization. Only has an effect if `do_normalize` is set to `True`.
+            image_std (`float` or `List[float]`, *optional*, defaults to `self.image_std`):
+                Image standard deviation to use for normalization. Only has an effect if `do_normalize` is set to
+                `True`.
+            do_convert_rgb (`bool`, *optional*, defaults to `self.do_convert_rgb`):
+                Whether to convert the image to RGB.
+            return_tensors (`str` or `TensorType`, *optional*):
+                The type of tensors to return. Can be one of:
+                - Unset: Return a list of `np.ndarray`.
+                - `TensorType.TENSORFLOW` or `'tf'`: Return a batch of type `tf.Tensor`.
+                - `TensorType.PYTORCH` or `'pt'`: Return a batch of type `torch.Tensor`.
+                - `TensorType.NUMPY` or `'np'`: Return a batch of type `np.ndarray`.
+                - `TensorType.JAX` or `'jax'`: Return a batch of type `jax.numpy.ndarray`.
+            data_format (`ChannelDimension` or `str`, *optional*, defaults to `ChannelDimension.FIRST`):
+                The channel dimension format for the output image. Can be one of:
+                - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
+                - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
+                - Unset: Use the channel dimension format of the input image.
+            input_data_format (`ChannelDimension` or `str`, *optional*):
+                The channel dimension format for the input image. If unset, the channel dimension format is inferred
+                from the input image. Can be one of:
+                - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
+                - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
+                - `"none"` or `ChannelDimension.NONE`: image in (height, width) format.
+
+        """
+        do_resize = do_resize if do_resize is not None else self.do_resize
+        size = size if size is not None else self.size
+        size = get_size_dict(size, param_name="size", default_to_square=False)
+        image_grid_pinpoints = image_grid_pinpoints if image_grid_pinpoints is not None else self.image_grid_pinpoints
+        resample = resample if resample is not None else self.resample
+        do_center_crop = do_center_crop if do_center_crop is not None else self.do_center_crop
+        crop_size = crop_size if crop_size is not None else self.crop_size
+        crop_size = get_size_dict(crop_size, param_name="crop_size", default_to_square=True)
+        do_rescale = do_rescale if do_rescale is not None else self.do_rescale
+        rescale_factor = rescale_factor if rescale_factor is not None else self.rescale_factor
+        do_normalize = do_normalize if do_normalize is not None else self.do_normalize
+        image_mean = image_mean if image_mean is not None else self.image_mean
+        image_std = image_std if image_std is not None else self.image_std
+        do_convert_rgb = do_convert_rgb if do_convert_rgb is not None else self.do_convert_rgb
+
+        images = make_batched_images(images)
+
+        if not valid_images(images):
+            raise ValueError(
+                "Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, "
+                "torch.Tensor, tf.Tensor or jax.ndarray."
+            )
+
+        validate_preprocess_arguments(
+            do_rescale=do_rescale,
+            rescale_factor=rescale_factor,
+            do_normalize=do_normalize,
+            image_mean=image_mean,
+            image_std=image_std,
+            do_center_crop=do_center_crop,
+            crop_size=crop_size,
+            do_resize=do_resize,
+            size=size,
+            resample=resample,
+        )
+
+        if do_convert_rgb:
+            images = [convert_to_rgb(image) for image in images]
+
+        # All transformations expect numpy arrays.
+        images = [to_numpy_array(image) for image in images]
+
+        if is_scaled_image(images[0]) and do_rescale:
+            logger.warning_once(
+                "It looks like you are trying to rescale already rescaled images. If the input"
+                " images have pixel values between 0 and 1, set `do_rescale=False` to avoid rescaling them again."
+            )
+
+        if input_data_format is None:
+            # We assume that all images have the same channel dimension format.
+            input_data_format = infer_channel_dimension_format(images[0])
+
+        new_images = []
+        image_sizes = [get_image_size(image, channel_dim=input_data_format) for image in images]
+        num_patches = []
+        for image in images:
+            # convert image into a list of patches
+            # we intentially use the same data format as the input data format
+            image_patches = self.get_image_patches(
+                image,
+                image_grid_pinpoints,
+                size=(size["shortest_edge"], size["shortest_edge"]),
+                patch_size=crop_size["height"],
+                resample=resample,
+                data_format=input_data_format,
+                input_data_format=input_data_format,
+            )
+
+            # preprocess patches
+            pixel_values = self._preprocess(
+                image_patches,
+                do_resize=do_resize,
+                size=size,
+                resample=resample,
+                do_center_crop=do_center_crop,
+                crop_size=crop_size,
+                do_rescale=do_rescale,
+                rescale_factor=rescale_factor,
+                do_normalize=do_normalize,
+                image_mean=image_mean,
+                image_std=image_std,
+                data_format=data_format,
+                input_data_format=input_data_format,
+            )
+            num_patches.append(len(pixel_values))
+            pixel_values = np.array(pixel_values)
+            new_images.append(pixel_values)
+        processed_images = self._pad_for_batching(new_images)
+
+        return BatchFeature(
+            #data={"pixel_values": new_images}, tensor_type=return_tensors
+            data={"pixel_values": processed_images, "num_patches":num_patches}, tensor_type=return_tensors
+        )

model-00003-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:faaf2890af5673988fe2ea5507af18d7d7c3d9402ffb74e142f81c96af2c82b7
+size 4946773672

processing_omchat.py

@@ -0,0 +1,285 @@
+from typing import List, Optional, Union
+
+from transformers import PreTrainedTokenizer
+from typing import List, Tuple
+
+from transformers.feature_extraction_utils import BatchFeature
+from transformers.image_utils import ImageInput
+from transformers.processing_utils import ProcessorMixin
+from transformers.tokenization_utils_base import PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy
+from transformers.utils import TensorType
+import torch
+
+def tokenizer_image_token(prompt, tokenizer, image_token_index=-200, return_tensors=None):
+    if "<image_0>" in prompt:
+        image_token_pattern = re.compile(r"<image_(\d+)>")
+        prompt_chunks = re.split(r'<image_[0-9]+>',prompt)
+        # Identify all the image tags
+        image_tags = image_token_pattern.findall(prompt)
+
+        input_ids = []
+        for i, chunk in enumerate(prompt_chunks):
+            input_ids.extend(tokenizer(chunk).input_ids)
+            if i < len(image_tags):
+                #input_ids.append(-100 * (int(image_tags[i]) + 3))
+                input_ids.append(-200)
+    else:
+        prompt_chunks = [tokenizer(chunk).input_ids for chunk in prompt.split('<image>')]
+        def insert_separator(X, sep):
+            return [ele for sublist in zip(X, [sep]*len(X)) for ele in sublist][:-1]
+
+        input_ids = []
+        offset = 0
+        if len(prompt_chunks) > 0 and len(prompt_chunks[0]) > 0 and prompt_chunks[0][0] == tokenizer.bos_token_id:
+            offset = 1
+            input_ids.append(prompt_chunks[0][0])
+
+        for x in insert_separator(prompt_chunks, [image_token_index] * (offset + 1)):
+            input_ids.extend(x[offset:])
+    # Convert to tensor if required
+    if return_tensors is not None:
+        if return_tensors == 'pt':
+            return torch.tensor(input_ids, dtype=torch.long)
+        else:
+            raise ValueError(f'Unsupported tensor type: {return_tensors}')
+
+    return input_ids
+
+
+def make_context(
+    tokenizer: PreTrainedTokenizer,
+    query: str,
+    history: List[Tuple[str, str]] = None,
+    system: str = "",
+    max_window_size: int = 6144,
+    chat_format: str = "chatml",
+):
+    if history is None:
+        history = []
+
+    if chat_format == "chatml":
+        im_start, im_end = "<|im_start|>", "<|im_end|>"
+        im_start_tokens = [151644]
+        im_end_tokens = [151645]
+        nl_tokens = tokenizer.encode("\n")
+
+        def _tokenize_str(role, content):
+            if "<image>" in content:
+                return f"{role}\n{content}", tokenizer.encode(
+                    role
+                ) + nl_tokens + tokenizer_image_token(
+                    content, tokenizer, -200
+                )
+            else:
+                return f"{role}\n{content}", tokenizer.encode(
+                    role
+                ) + nl_tokens + tokenizer.encode(content)
+
+        def _tokenize_str2(role, content):
+            return f"{role}\n{content}", tokenizer.encode(
+                role,
+            ) + nl_tokens + tokenizer.encode(content)
+
+        system_text, system_tokens_part = _tokenize_str("system", system)
+        system_tokens = im_start_tokens + system_tokens_part + im_end_tokens
+
+        raw_text = ""
+        context_tokens = []
+
+        for turn_query, turn_response in reversed(history):
+            query_text, query_tokens_part = _tokenize_str("user", turn_query)
+            query_tokens = im_start_tokens + query_tokens_part + im_end_tokens
+            response_text, response_tokens_part = _tokenize_str(
+                "assistant", turn_response
+            )
+            response_tokens = im_start_tokens + response_tokens_part + im_end_tokens
+
+            next_context_tokens = nl_tokens + query_tokens + nl_tokens + response_tokens
+            prev_chat = (
+                f"\n{im_start}{query_text}{im_end}\n{im_start}{response_text}{im_end}"
+            )
+
+            current_context_size = (
+                len(system_tokens) + len(next_context_tokens) + len(context_tokens)
+            )
+            if current_context_size < max_window_size:
+                context_tokens = next_context_tokens + context_tokens
+                raw_text = prev_chat + raw_text
+            else:
+                break
+
+        context_tokens = system_tokens + context_tokens
+        raw_text = f"{im_start}{system_text}{im_end}" + raw_text
+        context_tokens += (
+            nl_tokens
+            + im_start_tokens
+            + _tokenize_str("user", query)[1]
+            + im_end_tokens
+            + nl_tokens
+            + im_start_tokens
+            + tokenizer.encode("assistant")
+            + nl_tokens
+        )
+        raw_text += f"\n{im_start}user\n{query}{im_end}\n{im_start}assistant\n"
+
+    elif chat_format == "raw":
+        raw_text = query
+        context_tokens = tokenizer.encode(raw_text)
+    else:
+        raise NotImplementedError(f"Unknown chat format {chat_format!r}")
+
+    return raw_text, context_tokens
+
+def split_tensor(A, B):
+    split_tensors = []
+    start_idx = 0
+
+    for i, size in enumerate(B.tolist()):
+        split_tensor = A[i, :size, :, :, :]
+        split_tensors.append(split_tensor)  # Take the first element from the batch dimension
+
+    return split_tensors
+
+class OmChatProcessor(ProcessorMixin):
+    r"""
+    Constructs a OmChat processor which wraps a OmChat image processor and a LLaMa tokenizer into a single processor.
+
+    [`OmChatProcessor`] offers all the functionalities of [`OmChatImageProcessor`] and [`LlamaTokenizerFast`]. See the
+    [`~OmChatProcessor.__call__`] and [`~OmChatProcessor.decode`] for more information.
+
+    Args:
+        image_processor ([`OmChatImageProcessor`], *optional*):
+            The image processor is a required input.
+        tokenizer ([`LlamaTokenizerFast`], *optional*):
+            The tokenizer is a required input.
+        chat_template (`str`, *optional*): A Jinja template which will be used to convert lists of messages
+            in a chat into a tokenizable string.
+    """
+
+    attributes = ["image_processor", "tokenizer"]
+    valid_kwargs = ["chat_template"]
+    image_processor_class = "AutoImageProcessor"
+    tokenizer_class = "AutoTokenizer"
+
+    def __init__(self, image_processor=None, tokenizer=None,  **kwargs):
+        super().__init__(image_processor, tokenizer)
+
+    def __call__(
+        self,
+        text: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]],
+        system_prompt: str = "You are a helpful assistant.",
+        images: ImageInput = None,
+        padding: Union[bool, str, PaddingStrategy] = False,
+        truncation: Union[bool, str, TruncationStrategy] = None,
+        max_length: Optional[int] = None,
+        return_tensors: Optional[Union[str, TensorType]] = TensorType.PYTORCH,
+    ) -> BatchFeature:
+        """
+        Main method to prepare for the model one or several sequences(s) and image(s). This method forwards the `text`
+        and `kwargs` arguments to LlamaTokenizerFast's [`~LlamaTokenizerFast.__call__`] if `text` is not `None` to encode
+        the text. To prepare the image(s), this method forwards the `images` and `kwrags` arguments to
+        OmChatImageProcessor's [`~OmChatImageProcessor.__call__`] if `images` is not `None`. Please refer to the doctsring
+        of the above two methods for more information.
+
+        Args:
+            text (`str`, `List[str]`, `List[List[str]]`):
+                The sequence or batch of sequences to be encoded. Each sequence can be a string or a list of strings
+                (pretokenized string). If the sequences are provided as list of strings (pretokenized), you must set
+                `is_split_into_words=True` (to lift the ambiguity with a batch of sequences).
+            system_prompt ('str'):
+                the initial system prompt (i.e., You are a helpful assistant.)
+            images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`):
+                The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch
+                tensor. Both channels-first and channels-last formats are supported.
+            padding (`bool`, `str` or [`~utils.PaddingStrategy`], *optional*, defaults to `False`):
+                Select a strategy to pad the returned sequences (according to the model's padding side and padding
+                index) among:
+                - `True` or `'longest'`: Pad to the longest sequence in the batch (or no padding if only a single
+                  sequence if provided).
+                - `'max_length'`: Pad to a maximum length specified with the argument `max_length` or to the maximum
+                  acceptable input length for the model if that argument is not provided.
+                - `False` or `'do_not_pad'` (default): No padding (i.e., can output a batch with sequences of different
+                  lengths).
+            max_length (`int`, *optional*):
+                Maximum length of the returned list and optionally padding length (see above).
+            truncation (`bool`, *optional*):
+                Activates truncation to cut input sequences longer than `max_length` to `max_length`.
+            return_tensors (`str` or [`~utils.TensorType`], *optional*):
+                If set, will return tensors of a particular framework. Acceptable values are:
+
+                - `'tf'`: Return TensorFlow `tf.constant` objects.
+                - `'pt'`: Return PyTorch `torch.Tensor` objects.
+                - `'np'`: Return NumPy `np.ndarray` objects.
+                - `'jax'`: Return JAX `jnp.ndarray` objects.
+
+        Returns:
+            [`BatchFeature`]: A [`BatchFeature`] with the following fields:
+
+            - **input_ids** -- List of token ids to be fed to a model. Returned when `text` is not `None`.
+            - **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when
+              `return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not
+              `None`).
+            - **pixel_values** -- Pixel values to be fed to a model. Returned when `images` is not `None`.
+        """
+        #system_prompt="A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions."
+        if images is not None:
+            image_inputs = self.image_processor(images, return_tensors=return_tensors)
+            new_images = []
+            new_texts = []
+            img = image_inputs["pixel_values"]
+            num_patches = image_inputs["num_patches"]
+            img = split_tensor(img, num_patches)
+            if len(img) == 1:
+                n = num_patches.tolist()[0]
+                inp, context_tokens = make_context(
+                    self.tokenizer,
+                    "<image>\n"+"\n".join(["patch:<image>"]*(n-1)) +"\n"+ text.replace("<image>", ""),
+                    None,
+                    system_prompt,
+                )
+            
+            else:
+                texts = text.split("<image>")
+                final =texts[0]
+                for i, n in enumerate(num_patches.tolist()):
+                    final+= "\n<image>\n"+"\n".join(["patch:<image>"]*(n-1))+"\n"
+                    if i+1 < len(texts):
+                        final += texts[i+1]
+                inp, context_tokens = make_context(self.tokenizer, final, None, system_prompt)
+            text_inputs = {"input_ids": torch.tensor([context_tokens])}
+            image_inputs = {"images":torch.cat(img, dim=0)}
+            return BatchFeature(data={**text_inputs, **image_inputs})
+        else:
+            image_inputs = {"images":None}
+            inp, context_tokens = make_context(
+                    self.tokenizer,
+                    text.replace("<image>", "").strip(),
+                    None,
+                    "You are a helpful assistant.",
+                )
+            text_inputs = {"input_ids": torch.tensor([context_tokens])}
+
+            return BatchFeature(data={**text_inputs})
+
+    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.batch_decode with CLIP->Llama
+    def batch_decode(self, *args, **kwargs):
+        """
+        This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please
+        refer to the docstring of this method for more information.
+        """
+        return self.tokenizer.batch_decode(*args, **kwargs)
+
+    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.decode with CLIP->Llama
+    def decode(self, *args, **kwargs):
+        """
+        This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to
+        the docstring of this method for more information.
+        """
+        return self.tokenizer.decode(*args, **kwargs)
+
+    @property
+    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.model_input_names
+    def model_input_names(self):
+        tokenizer_input_names = self.tokenizer.model_input_names
+        image_processor_input_names = self.image_processor.model_input_names
+        return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names))