processing_youtu_vl.py

from typing import List, Union
import numpy
from transformers.feature_extraction_utils import BatchFeature
from transformers.image_utils import ImageInput
from transformers.processing_utils import ProcessingKwargs, ProcessorMixin, Unpack, VideosKwargs
from transformers.tokenization_utils_base import PreTokenizedInput, TextInput

class YoutuVLVideosProcessorKwargs(VideosKwargs, total=False):
    fps: Union[List[float], float]


class YoutuVLProcessorKwargs(ProcessingKwargs, total=False):
    videos_kwargs: YoutuVLVideosProcessorKwargs
    _defaults = {
        "text_kwargs": {
            "padding": False,
        },
        "videos_kwargs": {"fps": 2.0},
    }


class YoutuVLProcessor(ProcessorMixin):
   
    attributes = ["image_processor", "tokenizer"]
    valid_kwargs = ["chat_template"]

    image_processor_class = "AutoImageProcessor"
    tokenizer_class = ("PreTrainedTokenizer", "PreTrainedTokenizerFast")
    
    def __init__(self, image_processor=None, tokenizer=None, chat_template=None, **kwargs):
        self.image_token = "<|image_pad|>" if not hasattr(tokenizer, "image_token") else tokenizer.image_token
        self.video_token = "<|video_pad|>" if not hasattr(tokenizer, "video_token") else tokenizer.video_token
        super().__init__(image_processor, tokenizer, chat_template=chat_template)

    def __call__(
        self,
        text: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None,
        images: ImageInput = None,
        max_image_patches: int=36864,
        **kwargs: Unpack[YoutuVLProcessorKwargs],
    ) -> BatchFeature:
        """
        Args:
            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.
            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).
            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`.
            - **pixel_values_videos** -- Pixel values of videos to be fed to a model.
              Returned when `videos` is not `None`.
            - **image_grid_thw** -- List of image 3D grid in LLM. Returned when `images` is not `None`.
            - **video_grid_thw** -- List of video 3D grid in LLM. Returned when `videos` is not `None`.
            - **second_per_grid_ts** -- List of video seconds per time grid. Returned when `videos` is not `None`.
        """
        output_kwargs = self._merge_kwargs(
            YoutuVLProcessorKwargs,
            tokenizer_init_kwargs=self.tokenizer.init_kwargs,
            **kwargs,
        )
        if images is not None:
            image_inputs = self.image_processor(images=images, max_num_patches=max_image_patches, return_tensors="pt")
        else:
            image_inputs = {}
            image_grid_thw = None

        videos_inputs = {}
        video_grid_thw = None

        if not isinstance(text, list):
            text = [text]

        image_tokens = []
        if images is not None:
            merge_length = 4
            index = 0
            for i in range(len(text)):
                while self.image_token in text[i]:
                    h = image_inputs['spatial_shapes'][index][0]
                    w = image_inputs['spatial_shapes'][index][1]
                    repeats = h* w // merge_length
                    text[i] = text[i].replace(
                        self.image_token,
                        "<|placeholder|>" * repeats,
                        1,
                    )
                    index += 1
                text[i] = text[i].replace("<|placeholder|>", self.image_token)
            assert(index == image_inputs['spatial_shapes'].shape[0])


        if video_grid_thw is not None:
            merge_length = self.image_processor.merge_size ** 2
            index = 0
            for i in range(len(text)):
                while self.video_token in text[i]:
                    text[i] = text[i].replace(
                        self.video_token,
                        "<|placeholder|>" * (video_grid_thw[index].prod() // merge_length),
                        1,
                    )
                    index += 1
                text[i] = text[i].replace("<|placeholder|>", self.video_token)

        text_inputs = self.tokenizer(text, **output_kwargs["text_kwargs"])

        return BatchFeature(data={**text_inputs, **image_inputs, **videos_inputs})
    
    def get_max_image_patches(self, images):
        return self.image_processor.get_max_image_patches(images)

    def batch_decode(self, *args, **kwargs):
        return self.tokenizer.batch_decode(*args, **kwargs)

    def decode(self, *args, **kwargs):
        return self.tokenizer.decode(*args, **kwargs)

    def post_process_image_text_to_text(
        self, generated_outputs, skip_special_tokens=True, clean_up_tokenization_spaces=False, **kwargs
    ):
        """
        Post-process the output of the model to decode the text.

        Args:
            generated_outputs (`torch.Tensor` or `np.ndarray`):
                The output of the model `generate` function. The output is
                expected to be a tensor of shape `(batch_size, sequence_length)`
                or `(sequence_length,)`.
            skip_special_tokens (`bool`, *optional*, defaults to `True`):
                Whether or not to remove special tokens in the output. Argument
                passed to the tokenizer's `batch_decode` method.
            Clean_up_tokenization_spaces (`bool`, *optional*, defaults to `False`):
                Whether or not to clean up the tokenization spaces. Argument
                passed to the tokenizer's `batch_decode` method.
            **kwargs:
                Additional arguments to be passed to the tokenizer's `batch_decode method`.

        Returns:
            `List[str]`: The decoded text.
        """
        return self.tokenizer.batch_decode(
            generated_outputs,
            skip_special_tokens=skip_special_tokens,
            clean_up_tokenization_spaces=clean_up_tokenization_spaces,
            **kwargs,
        )

    @property
    def model_input_names(self):
        tokenizer_input_names = self.tokenizer.model_input_names
        image_processor_input_names = self.image_processor.model_input_names
        names_from_processor = list(dict.fromkeys(tokenizer_input_names + image_processor_input_names))
        return names_from_processor + ["second_per_grid_ts"]


__all__ = ["YoutuVLProcessor"]