adapt to transformers==4.57.6

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,199 @@
+---
+library_name: transformers
+tags: []
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]

added_tokens.json

@@ -0,0 +1,28 @@
+{
+  "</think>": 151668,
+  "</tool_call>": 151658,
+  "</tool_response>": 151666,
+  "<think>": 151667,
+  "<tool_call>": 151657,
+  "<tool_response>": 151665,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

chat_template.jinja

@@ -0,0 +1,120 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {%- if messages[0].content is string %}
+            {{- messages[0].content }}
+        {%- else %}
+            {%- for content in messages[0].content %}
+                {%- if 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' }}
+        {%- if messages[0].content is string %}
+            {{- messages[0].content }}
+        {%- else %}
+            {%- for content in messages[0].content %}
+                {%- if 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set image_count = namespace(value=0) %}
+{%- set video_count = namespace(value=0) %}
+{%- for message in messages %}
+    {%- if message.role == "user" %}
+        {{- '<|im_start|>' + message.role + '\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- else %}
+            {%- for content in message.content %}
+                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
+                    <|vision_start|><|image_pad|><|vision_end|>
+                {%- elif content.type == 'video' or 'video' in content %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
+                    <|vision_start|><|video_pad|><|vision_end|>
+                {%- elif 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role + '\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- else %}
+            {%- for content_item in message.content %}
+                {%- if 'text' in content_item %}
+                    {{- content_item.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and message.content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- else %}
+            {%- for content in message.content %}
+                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
+                    <|vision_start|><|image_pad|><|vision_end|>
+                {%- elif content.type == 'video' or 'video' in content %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
+                    <|vision_start|><|video_pad|><|vision_end|>
+                {%- elif 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

image_processing_qwen2_vl.py

@@ -0,0 +1,501 @@
+import math
+from typing import Optional, Union
+
+import numpy as np
+
+from transformers.image_processing_utils import BaseImageProcessor, BatchFeature
+from transformers.image_transforms import (
+    convert_to_rgb,
+    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,
+    make_flat_list_of_images,
+    to_numpy_array,
+    valid_images,
+    validate_preprocess_arguments,
+)
+from transformers.utils import TensorType, logging
+from transformers.video_utils import VideoInput, make_batched_videos
+
+
+logger = logging.get_logger(__name__)
+
+
+def smart_resize(
+    height: int, width: int, factor: int = 28, min_pixels: int = 56 * 56, max_pixels: int = 14 * 14 * 4 * 1280
+):
+    """Rescales the image so that the following conditions are met:
+
+    1. Both dimensions (height and width) are divisible by 'factor'.
+
+    2. The total number of pixels is within the range ['min_pixels', 'max_pixels'].
+
+    3. The aspect ratio of the image is maintained as closely as possible.
+
+    """
+    if max(height, width) / min(height, width) > 200:
+        raise ValueError(
+            f"absolute aspect ratio must be smaller than 200, got {max(height, width) / min(height, width)}"
+        )
+    h_bar = round(height / factor) * factor
+    w_bar = round(width / factor) * factor
+    if h_bar * w_bar > max_pixels:
+        beta = math.sqrt((height * width) / max_pixels)
+        h_bar = max(factor, math.floor(height / beta / factor) * factor)
+        w_bar = max(factor, math.floor(width / beta / factor) * factor)
+    elif h_bar * w_bar < min_pixels:
+        beta = math.sqrt(min_pixels / (height * width))
+        h_bar = math.ceil(height * beta / factor) * factor
+        w_bar = math.ceil(width * beta / factor) * factor
+    return h_bar, w_bar
+
+
+class Qwen2VLImageProcessor(BaseImageProcessor):
+    r"""
+    Constructs a Qwen2-VL image processor that dynamically resizes images based on the original images.
+
+    Args:
+        do_resize (`bool`, *optional*, defaults to `True`):
+            Whether to resize the image's (height, width) dimensions.
+        size (`dict[str, int]`, *optional*, defaults to `{"shortest_edge": 56 * 56, "longest_edge": 28 * 28 * 1280}`):
+            Size of the image after resizing. `shortest_edge` and `longest_edge` keys must be present.
+        resample (`PILImageResampling`, *optional*, defaults to `Resampling.BICUBIC`):
+            Resampling filter to use when resizing the image.
+        do_rescale (`bool`, *optional*, defaults to `True`):
+            Whether to rescale the image by the specified scale `rescale_factor`.
+        rescale_factor (`int` or `float`, *optional*, defaults to `1/255`):
+            Scale factor to use if rescaling the image.
+        do_normalize (`bool`, *optional*, defaults to `True`):
+            Whether to normalize the image.
+        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 for each channel in the image.
+        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 for each channel in the image.
+        do_convert_rgb (`bool`, *optional*, defaults to `True`):
+            Whether to convert the image to RGB.
+        min_pixels (`int`, *optional*, defaults to `56 * 56`):
+            The min pixels of the image to resize the image.
+        max_pixels (`int`, *optional*, defaults to `28 * 28 * 1280`):
+            The max pixels of the image to resize the image.
+        patch_size (`int`, *optional*, defaults to 14):
+            The spatial patch size of the vision encoder.
+        temporal_patch_size (`int`, *optional*, defaults to 2):
+            The temporal patch size of the vision encoder.
+        merge_size (`int`, *optional*, defaults to 2):
+            The merge size of the vision encoder to llm encoder.
+    """
+
+    model_input_names = ["pixel_values", "image_grid_thw", "pixel_values_videos", "video_grid_thw"]
+
+    def __init__(
+        self,
+        do_resize: bool = True,
+        size: Optional[dict[str, int]] = None,
+        resample: PILImageResampling = PILImageResampling.BICUBIC,
+        do_rescale: bool = True,
+        rescale_factor: Union[int, float] = 1 / 255,
+        do_normalize: bool = True,
+        image_mean: Optional[Union[float, list[float]]] = None,
+        image_std: Optional[Union[float, list[float]]] = None,
+        do_convert_rgb: bool = True,
+        min_pixels: Optional[int] = None,
+        max_pixels: Optional[int] = None,
+        patch_size: int = 14,
+        temporal_patch_size: int = 2,
+        merge_size: int = 2,
+        focus_size: int = 2,
+        **kwargs,
+    ) -> None:
+        super().__init__(**kwargs)
+        if size is not None and ("shortest_edge" not in size or "longest_edge" not in size):
+            raise ValueError("size must contain 'shortest_edge' and 'longest_edge' keys.")
+        else:
+            size = {"shortest_edge": 56 * 56, "longest_edge": 28 * 28 * 1280}
+        # backward compatibility: override size with min_pixels and max_pixels if they are provided
+        if min_pixels is not None:
+            size["shortest_edge"] = min_pixels
+        if max_pixels is not None:
+            size["longest_edge"] = max_pixels
+        self.min_pixels = size["shortest_edge"]
+        self.max_pixels = size["longest_edge"]
+        self.size = size
+
+        self.do_resize = do_resize
+        self.resample = resample
+        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.patch_size = patch_size
+        self.temporal_patch_size = temporal_patch_size
+        self.merge_size = merge_size
+        self.focus_size = focus_size
+        self.do_convert_rgb = do_convert_rgb
+
+    def _preprocess(
+        self,
+        images: Union[ImageInput, VideoInput],
+        do_resize: Optional[bool] = None,
+        size: Optional[dict[str, int]] = None,
+        resample: Optional[PILImageResampling] = None,
+        do_rescale: Optional[bool] = None,
+        rescale_factor: Optional[float] = None,
+        do_normalize: Optional[bool] = None,
+        image_mean: Optional[Union[float, list[float]]] = None,
+        image_std: Optional[Union[float, list[float]]] = None,
+        patch_size: Optional[int] = None,
+        temporal_patch_size: Optional[int] = None,
+        merge_size: Optional[int] = None,
+        focus_size: Optional[int] = None,
+        do_convert_rgb: Optional[bool] = None,
+        data_format: Optional[ChannelDimension] = ChannelDimension.FIRST,
+        input_data_format: Optional[Union[str, ChannelDimension]] = None,
+    ):
+        """
+        Preprocess an image or batch of images. Copy of the `preprocess` method from `CLIPImageProcessor`.
+
+        Args:
+            images (`ImageInput`):
+                Image or batch of images to preprocess. Expects pixel values ranging from 0 to 255. If pixel values range from 0 to 1, set `do_rescale=False`.
+            vision_info (`list[Dict]`, *optional*):
+                Optional list of dictionaries containing additional information about vision inputs.
+            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` and `longest_edge` keys must be present.
+            resample (`PILImageResampling`, *optional*, defaults to `self.resample`):
+                Resampling filter to use if resizing the image. This can be one of the `PILImageResampling` enums.
+            do_rescale (`bool`, *optional*, defaults to `self.do_rescale`):
+                Whether to rescale the image.
+            rescale_factor (`float`, *optional*, defaults to `self.rescale_factor`):
+                Scale factor to use if rescaling the image.
+            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`):
+                Mean to use if normalizing the image. Can be a float or a list of floats corresponding to the number of channels in the image.
+            image_std (`float` or `list[float]`, *optional*, defaults to `self.image_std`):
+                Standard deviation to use if normalizing the image. Can be a float or a list of floats corresponding to the number of channels in the image.
+            patch_size (`int`, *optional*, defaults to `self.patch_size`):
+                The spatial patch size of the vision encoder.
+            temporal_patch_size (`int`, *optional*, defaults to `self.temporal_patch_size`):
+                The temporal patch size of the vision encoder.
+            merge_size (`int`, *optional*, defaults to `self.merge_size`):
+                The merge size of the vision encoder to llm encoder.
+            do_convert_rgb (`bool`, *optional*, defaults to `self.do_convert_rgb`):
+                Whether to convert the image to RGB.
+            data_format (`ChannelDimension`, *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. 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.   - `"none"` or `ChannelDimension.NONE`: image in (height, width) format.
+        """
+        images = make_flat_list_of_images(images)
+
+        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 do_rescale and is_scaled_image(images[0]):
+            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])
+
+        height, width = get_image_size(images[0], channel_dim=input_data_format)
+        resized_height, resized_width = height, width
+        processed_images = []
+        for image in images:
+            if do_resize:
+                resized_height, resized_width = smart_resize(
+                    height,
+                    width,
+                    factor=patch_size * merge_size * focus_size,
+                    min_pixels=size["shortest_edge"],
+                    max_pixels=size["longest_edge"],
+                )
+                image = resize(
+                    image, size=(resized_height, resized_width), resample=resample, input_data_format=input_data_format
+                )
+
+            if do_rescale:
+                image = self.rescale(image, scale=rescale_factor, input_data_format=input_data_format)
+
+            if do_normalize:
+                image = self.normalize(
+                    image=image, mean=image_mean, std=image_std, input_data_format=input_data_format
+                )
+
+            image = to_channel_dimension_format(image, data_format, input_channel_dim=input_data_format)
+            processed_images.append(image)
+
+        patches = np.array(processed_images)
+        if data_format == ChannelDimension.LAST:
+            patches = patches.transpose(0, 3, 1, 2)
+        if patches.shape[0] % temporal_patch_size != 0:
+            repeats = np.repeat(
+                patches[-1][np.newaxis], temporal_patch_size - (patches.shape[0] % temporal_patch_size), axis=0
+            )
+            patches = np.concatenate([patches, repeats], axis=0)
+        channel = patches.shape[1]
+        grid_t = patches.shape[0] // temporal_patch_size
+        grid_h, grid_w = resized_height // patch_size, resized_width // patch_size
+        patches = patches.reshape(
+            grid_t,
+            temporal_patch_size,
+            channel,
+            grid_h // merge_size,
+            merge_size,
+            patch_size,
+            grid_w // merge_size,
+            merge_size,
+            patch_size,
+        )
+        patches = patches.transpose(0, 3, 6, 4, 7, 2, 1, 5, 8)
+        flatten_patches = patches.reshape(
+            grid_t * grid_h * grid_w, channel * temporal_patch_size * patch_size * patch_size
+        )
+
+        return flatten_patches, (grid_t, grid_h, grid_w)
+
+    def preprocess(
+        self,
+        images: ImageInput,
+        videos: Optional[VideoInput] = None,
+        do_resize: Optional[bool] = None,
+        size: Optional[dict[str, int]] = None,
+        min_pixels: Optional[int] = None,
+        max_pixels: Optional[int] = None,
+        resample: Optional[PILImageResampling] = None,
+        do_rescale: Optional[bool] = None,
+        rescale_factor: Optional[float] = None,
+        do_normalize: Optional[bool] = None,
+        image_mean: Optional[Union[float, list[float]]] = None,
+        image_std: Optional[Union[float, list[float]]] = None,
+        patch_size: Optional[int] = None,
+        temporal_patch_size: Optional[int] = None,
+        merge_size: Optional[int] = None,
+        do_convert_rgb: Optional[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`.
+            videos (`VideoInput`):
+                Video to preprocess. Expects a single or batch of videos with pixel values ranging from 0 to 255. If
+                passing in videos 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_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`.
+            min_pixels (`int`, *optional*, defaults to `self.min_pixels`):
+                The min pixels of the image to resize the image.
+            max_pixels (`int`, *optional*, defaults to `self.max_pixels`):
+                The max pixels of the image to resize the image.
+            patch_size (`int`, *optional*, defaults to `self.patch_size`):
+                The spatial patch size of the vision encoder.
+            temporal_patch_size (`int`, *optional*, defaults to `self.temporal_patch_size`):
+                The temporal patch size of the vision encoder.
+            merge_size (`int`, *optional*, defaults to `self.merge_size`):
+                The merge size of the vision encoder to llm encoder.
+            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.
+
+        """
+        min_pixels = min_pixels if min_pixels is not None else self.min_pixels
+        max_pixels = max_pixels if max_pixels is not None else self.max_pixels
+
+        if size is not None:
+            if "shortest_edge" not in size or "longest_edge" not in size:
+                raise ValueError("size must contain 'shortest_edge' and 'longest_edge' keys.")
+            min_pixels = size["shortest_edge"]
+        elif min_pixels is not None and max_pixels is not None:
+            # backward compatibility: override size with min_pixels and max_pixels if they are provided
+            size = {"shortest_edge": min_pixels, "longest_edge": max_pixels}
+        else:
+            size = {**self.size}
+
+        do_resize = do_resize if do_resize is not None else self.do_resize
+
+        resample = resample if resample is not None else self.resample
+        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
+        patch_size = patch_size if patch_size is not None else self.patch_size
+        temporal_patch_size = temporal_patch_size if temporal_patch_size is not None else self.temporal_patch_size
+        merge_size = merge_size if merge_size is not None else self.merge_size
+        do_convert_rgb = do_convert_rgb if do_convert_rgb is not None else self.do_convert_rgb
+
+        if images is not None:
+            images = self.fetch_images(images)
+            images = make_flat_list_of_images(images)
+
+        if images is not None and 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(
+            rescale_factor=rescale_factor,
+            do_normalize=do_normalize,
+            image_mean=image_mean,
+            image_std=image_std,
+            do_resize=do_resize,
+            size=size,
+            resample=resample,
+        )
+
+        data = {}
+        if images is not None:
+            pixel_values, vision_grid_thws = [], []
+            for image in images:
+                patches, image_grid_thw = self._preprocess(
+                    image,
+                    do_resize=do_resize,
+                    size=size,
+                    resample=resample,
+                    do_rescale=do_rescale,
+                    rescale_factor=rescale_factor,
+                    do_normalize=do_normalize,
+                    image_mean=image_mean,
+                    image_std=image_std,
+                    patch_size=patch_size,
+                    temporal_patch_size=temporal_patch_size,
+                    merge_size=merge_size,
+                    data_format=data_format,
+                    do_convert_rgb=do_convert_rgb,
+                    input_data_format=input_data_format,
+                )
+                pixel_values.extend(patches)
+                vision_grid_thws.append(image_grid_thw)
+            pixel_values = np.array(pixel_values)
+            vision_grid_thws = np.array(vision_grid_thws)
+            data.update({"pixel_values": pixel_values, "image_grid_thw": vision_grid_thws})
+
+        # kept for BC only and should be removed after v5.0
+        if videos is not None:
+            logger.warning(
+                "`Qwen2VLImageProcessor` works only with image inputs and doesn't process videos anymore. "
+                "This is a deprecated behavior and will be removed in v5.0. "
+                "Your videos should be forwarded to `Qwen2VLVideoProcessor`. "
+            )
+            videos = make_batched_videos(videos)
+            pixel_values_videos, vision_grid_thws_videos = [], []
+            for images in videos:
+                patches, video_grid_thw = self._preprocess(
+                    images,
+                    do_resize=do_resize,
+                    size=size,
+                    resample=resample,
+                    do_rescale=do_rescale,
+                    rescale_factor=rescale_factor,
+                    do_normalize=do_normalize,
+                    image_mean=image_mean,
+                    image_std=image_std,
+                    patch_size=patch_size,
+                    temporal_patch_size=temporal_patch_size,
+                    merge_size=merge_size,
+                    data_format=data_format,
+                    do_convert_rgb=do_convert_rgb,
+                    input_data_format=input_data_format,
+                )
+                pixel_values_videos.extend(patches)
+                vision_grid_thws_videos.append(video_grid_thw)
+            data.update(
+                {
+                    "pixel_values_videos": np.array(pixel_values_videos),
+                    "video_grid_thw": np.array(vision_grid_thws_videos),
+                }
+            )
+
+        return BatchFeature(data=data, tensor_type=return_tensors)
+
+    def get_number_of_image_patches(self, height: int, width: int, images_kwargs=None):
+        """
+        A utility that returns number of image patches for a given image size.
+
+        Args:
+            height (`int`):
+                Height of the input image.
+            width (`int`):
+                Width of the input image.
+            images_kwargs (`dict`, *optional*)
+                Any kwargs to override defaults of the image processor.
+        Returns:
+            `int`: Number of image patches per image.
+        """
+        min_pixels = images_kwargs["min_pixels"] if "min_pixels" in images_kwargs else self.size["shortest_edge"]
+        max_pixels = images_kwargs["max_pixels"] if "max_pixels" in images_kwargs else self.size["longest_edge"]
+        patch_size = images_kwargs.get("patch_size", self.patch_size)
+        merge_size = images_kwargs.get("merge_size", self.merge_size)
+        focus_size = images_kwargs.get("focus_size", self.focus_size)
+
+        factor = patch_size * merge_size * focus_size
+        resized_height, resized_width = smart_resize(
+            height, width, factor, min_pixels=min_pixels, max_pixels=max_pixels
+        )
+        grid_h, grid_w = resized_height // patch_size, resized_width // patch_size
+        return grid_h * grid_w
+
+
+__all__ = ["Qwen2VLImageProcessor"]

image_processing_qwen2_vl_fast.py

@@ -0,0 +1,288 @@
+from typing import Optional, Union
+
+import torch
+from torchvision.transforms.v2 import functional as F
+
+from transformers.image_processing_utils import BatchFeature
+from transformers.image_processing_utils_fast import (
+    BaseImageProcessorFast,
+    DefaultFastImageProcessorKwargs,
+    group_images_by_shape,
+    reorder_images,
+)
+from transformers.image_utils import (
+    OPENAI_CLIP_MEAN,
+    OPENAI_CLIP_STD,
+    ChannelDimension,
+    ImageInput,
+    PILImageResampling,
+    SizeDict,
+)
+from transformers.processing_utils import Unpack
+from transformers.utils import (
+    TensorType,
+    auto_docstring,
+    logging,
+)
+from transformers.video_utils import VideoInput, make_batched_videos
+from .image_processing_qwen2_vl import smart_resize
+
+
+logger = logging.get_logger(__name__)
+
+
+class Qwen2VLFastImageProcessorKwargs(DefaultFastImageProcessorKwargs):
+    """
+    min_pixels (`int`, *optional*, defaults to `56 * 56`):
+        The min pixels of the image to resize the image.
+    max_pixels (`int`, *optional*, defaults to `28 * 28 * 1280`):
+        The max pixels of the image to resize the image.
+    patch_size (`int`, *optional*, defaults to 14):
+        The spatial patch size of the vision encoder.
+    temporal_patch_size (`int`, *optional*, defaults to 2):
+        The temporal patch size of the vision encoder.
+    merge_size (`int`, *optional*, defaults to 2):
+        The merge size of the vision encoder to llm encoder.
+    """
+
+    min_pixels: Optional[int]
+    max_pixels: Optional[int]
+    patch_size: Optional[int]
+    temporal_patch_size: Optional[int]
+    merge_size: Optional[int]
+    focus_size: Optional[int]
+
+
+@auto_docstring
+class ZFQwen2VLImageProcessorFast(BaseImageProcessorFast):
+    do_resize = True
+    resample = PILImageResampling.BICUBIC
+    size = {"shortest_edge": 56 * 56, "longest_edge": 28 * 28 * 1280}
+    do_rescale = True
+    do_normalize = True
+    image_mean = OPENAI_CLIP_MEAN
+    image_std = OPENAI_CLIP_STD
+    do_convert_rgb = True
+    patch_size = 14
+    temporal_patch_size = 2
+    merge_size = 2
+    focus_size = 2
+    min_pixels = None
+    max_pixels = None
+    valid_kwargs = Qwen2VLFastImageProcessorKwargs
+    model_input_names = ["pixel_values", "image_grid_thw", "pixel_values_videos", "video_grid_thw"]
+
+    def __init__(self, **kwargs: Unpack[Qwen2VLFastImageProcessorKwargs]):
+        size = kwargs.pop("size", None)
+        min_pixels = kwargs.pop("min_pixels", None)
+        max_pixels = kwargs.pop("max_pixels", None)
+        # backward compatibility: override size with min_pixels and max_pixels if they are provided
+        size = self.size if size is None else size
+        if min_pixels is not None:
+            size["shortest_edge"] = min_pixels
+            size.pop("min_pixels", None)
+        if max_pixels is not None:
+            size["longest_edge"] = max_pixels
+            size.pop("max_pixels", None)
+        if "shortest_edge" not in size or "longest_edge" not in size:
+            raise ValueError("size must contain 'shortest_edge' and 'longest_edge' keys.")
+
+        super().__init__(size=size, min_pixels=min_pixels, max_pixels=max_pixels, **kwargs)
+
+    def _further_process_kwargs(
+        self,
+        size: Optional[SizeDict] = None,
+        min_pixels: Optional[int] = None,
+        max_pixels: Optional[int] = None,
+        **kwargs,
+    ) -> dict:
+        """
+        Update kwargs that need further processing before being validated
+        Can be overridden by subclasses to customize the processing of kwargs.
+        """
+        if min_pixels is not None and max_pixels is not None:
+            size = {"shortest_edge": min_pixels, "longest_edge": max_pixels}
+        elif size is not None:
+            if "shortest_edge" not in size or "longest_edge" not in size:
+                raise ValueError("size must contain 'shortest_edge' and 'longest_edge' keys.")
+            min_pixels = size["shortest_edge"]
+            max_pixels = size["longest_edge"]
+        else:
+            size = {**self.size}
+
+        return super()._further_process_kwargs(size=size, min_pixels=min_pixels, max_pixels=max_pixels, **kwargs)
+
+    @auto_docstring
+    def preprocess(
+        self,
+        images: ImageInput,
+        videos: Optional[VideoInput] = None,
+        **kwargs: Unpack[Qwen2VLFastImageProcessorKwargs],
+    ) -> BatchFeature:
+        return super().preprocess(images, videos, **kwargs)
+
+    def _preprocess_image_like_inputs(
+        self,
+        images: ImageInput,
+        videos: VideoInput,
+        do_convert_rgb: bool,
+        input_data_format: ChannelDimension,
+        device: Optional[Union[str, "torch.device"]] = None,
+        **kwargs: Unpack[DefaultFastImageProcessorKwargs],
+    ) -> BatchFeature:
+        """
+        Preprocess image-like inputs.
+        To be overridden by subclasses when image-like inputs other than images should be processed.
+        It can be used for segmentation maps, depth maps, etc.
+        """
+        # Prepare input images
+        batch_feature = BatchFeature()
+        if images is not None:
+            images = self._prepare_image_like_inputs(
+                images=images, do_convert_rgb=do_convert_rgb, input_data_format=input_data_format, device=device
+            )
+            batch_feature = self._preprocess(images, **kwargs)
+        if videos is not None:
+            logger.warning(
+                "`Qwen2VLImageProcessorFast` works only with image inputs and doesn't process videos anymore. "
+                "This is a deprecated behavior and will be removed in v5.0. "
+                "Your videos should be forwarded to `Qwen2VLVideoProcessor`. "
+            )
+            # Can't change _prepare_images_structure to work with videos because it also needs to work with images.
+            videos = make_batched_videos(videos)
+            videos = [
+                torch.stack(self._prepare_image_like_inputs(video, do_convert_rgb, input_data_format, device))
+                for video in videos
+            ]
+            video_outputs = self._preprocess(videos, **kwargs)
+            batch_feature.update(
+                {"pixel_values_videos": video_outputs.pixel_values, "video_grid_thw": video_outputs.image_grid_thw}
+            )
+        return batch_feature
+
+    def _preprocess(
+        self,
+        images: list["torch.Tensor"],
+        do_resize: bool,
+        size: SizeDict,
+        interpolation: Optional["F.InterpolationMode"],
+        do_rescale: bool,
+        rescale_factor: float,
+        do_normalize: bool,
+        image_mean: Optional[Union[float, list[float]]],
+        image_std: Optional[Union[float, list[float]]],
+        patch_size: int,
+        temporal_patch_size: int,
+        merge_size: int,
+        focus_size: int,
+        disable_grouping: Optional[bool],
+        return_tensors: Optional[Union[str, TensorType]],
+        **kwargs,
+    ):
+        # Group images by size for batched resizing
+        grouped_images, grouped_images_index = group_images_by_shape(images, disable_grouping=disable_grouping)
+        resized_images_grouped = {}
+        for shape, stacked_images in grouped_images.items():
+            height, width = stacked_images.shape[-2:]
+            if do_resize:
+                resized_height, resized_width = smart_resize(
+                    height,
+                    width,
+                    factor=patch_size * merge_size * focus_size,
+                    min_pixels=size["shortest_edge"],
+                    max_pixels=size["longest_edge"],
+                )
+                stacked_images = self.resize(
+                    image=stacked_images,
+                    size=SizeDict(height=resized_height, width=resized_width),
+                    interpolation=interpolation,
+                )
+            resized_images_grouped[shape] = stacked_images
+        resized_images = reorder_images(resized_images_grouped, grouped_images_index)
+
+        # Group images by size for further processing
+        # Needed in case do_resize is False, or resize returns images with different sizes
+        grouped_images, grouped_images_index = group_images_by_shape(resized_images, disable_grouping=disable_grouping)
+        processed_images_grouped = {}
+        processed_grids = {}
+        for shape, stacked_images in grouped_images.items():
+            resized_height, resized_width = stacked_images.shape[-2:]
+            # Fused rescale and normalize
+            patches = self.rescale_and_normalize(
+                stacked_images, do_rescale, rescale_factor, do_normalize, image_mean, image_std
+            )
+            if patches.ndim == 4:
+                # add a temporal dimension if we have images
+                patches = patches.unsqueeze(1)
+            if patches.shape[1] % temporal_patch_size != 0:
+                repeats = patches[:, -1:].repeat(1, temporal_patch_size - 1, 1, 1, 1)
+                patches = torch.cat([patches, repeats], dim=1)
+            batch_size, grid_t, channel = patches.shape[:3]
+            grid_t = grid_t // temporal_patch_size
+            grid_h, grid_w = resized_height // patch_size, resized_width // patch_size
+
+            patches = patches.view(
+                batch_size,
+                grid_t,
+                temporal_patch_size,
+                channel,
+                grid_h // merge_size,
+                merge_size,
+                patch_size,
+                grid_w // merge_size,
+                merge_size,
+                patch_size,
+            )
+            # Reorder dimensions to group grid and patch information for subsequent flattening.
+            # (batch, grid_t, grid_h, grid_w, merge_h, merge_w, channel, temp_patch_size, patch_h, patch_w)
+            patches = patches.permute(0, 1, 4, 7, 5, 8, 3, 2, 6, 9)
+            flatten_patches = patches.reshape(
+                batch_size,
+                grid_t * grid_h * grid_w,
+                channel * temporal_patch_size * patch_size * patch_size,
+            )
+
+            processed_images_grouped[shape] = flatten_patches
+            processed_grids[shape] = [[grid_t, grid_h, grid_w]] * batch_size
+
+        processed_images = reorder_images(processed_images_grouped, grouped_images_index)
+        processed_grids = reorder_images(processed_grids, grouped_images_index)
+        pixel_values = torch.cat(processed_images, dim=0)
+        image_grid_thw = torch.tensor(processed_grids)
+
+        return BatchFeature(
+            data={"pixel_values": pixel_values, "image_grid_thw": image_grid_thw}, tensor_type=return_tensors
+        )
+
+    def get_number_of_image_patches(self, height: int, width: int, images_kwargs=None):
+        """
+        A utility that returns number of image patches for a given image size.
+
+        Note: Do not remove this method! It is used by vLLM to infer the number of patches and placeholders
+        without an image input.
+
+        Args:
+            height (`int`):
+                Height of the input image.
+            width (`int`):
+                Width of the input image.
+            images_kwargs (`dict`, *optional*)
+                Any kwargs to override defaults of the image processor.
+        Returns:
+            `int`: Number of image patches per image.
+        """
+        min_pixels = images_kwargs["min_pixels"] if "min_pixels" in images_kwargs else self.size["shortest_edge"]
+        max_pixels = images_kwargs["max_pixels"] if "max_pixels" in images_kwargs else self.size["longest_edge"]
+        patch_size = images_kwargs.get("patch_size", self.patch_size)
+        merge_size = images_kwargs.get("merge_size", self.merge_size)
+        focus_size = images_kwargs.get("focus_size", self.focus_size)
+
+        factor = patch_size * merge_size * focus_size
+        resized_height, resized_width = smart_resize(
+            height, width, factor, min_pixels=min_pixels, max_pixels=max_pixels
+        )
+        grid_h, grid_w = resized_height // patch_size, resized_width // patch_size
+        return grid_h * grid_w
+
+
+__all__ = ["ZFQwen2VLImageProcessorFast"]

preprocessor_config.json

@@ -0,0 +1,44 @@
+{
+  "auto_map": {
+    "AutoImageProcessor": "image_processing_qwen2_vl_fast.ZFQwen2VLImageProcessorFast",
+    "AutoProcessor": "processing_qwen3_vl.ZFQwen3VLProcessor"
+  },
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "disable_grouping": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": null,
+  "do_rescale": true,
+  "do_resize": true,
+  "focus_size": 2,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ZFQwen2VLImageProcessorFast",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "input_data_format": null,
+  "max_pixels": null,
+  "merge_size": 2,
+  "min_pixels": null,
+  "pad_size": null,
+  "patch_size": 16,
+  "processor_class": "ZFQwen3VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_tensors": null,
+  "size": {
+    "longest_edge": 16777216,
+    "shortest_edge": 65536
+  },
+  "temporal_patch_size": 2
+}

processing_qwen3_vl.py

@@ -0,0 +1,320 @@
+from typing import Optional, Union
+
+import numpy as np
+
+from transformers.feature_extraction_utils import BatchFeature
+from transformers.image_utils import ImageInput
+from transformers.processing_utils import ImagesKwargs, MultiModalData, ProcessingKwargs, ProcessorMixin, Unpack, VideosKwargs
+from transformers.tokenization_utils_base import PreTokenizedInput, TextInput
+from transformers.utils import logging
+from transformers.video_utils import VideoInput
+
+
+logger = logging.get_logger(__name__)
+
+
+class Qwen3VLVideosProcessorKwargs(VideosKwargs, total=False):
+    focus_size: Optional[int]
+
+
+class Qwen3VLImagesKwargs(ImagesKwargs):
+    min_pixels: Optional[int]
+    max_pixels: Optional[int]
+    patch_size: Optional[int]
+    temporal_patch_size: Optional[int]
+    merge_size: Optional[int]
+    focus_size: Optional[int]
+
+
+class Qwen3VLProcessorKwargs(ProcessingKwargs, total=False):
+    images_kwargs: Qwen3VLImagesKwargs
+    videos_kwargs: Qwen3VLVideosProcessorKwargs
+    _defaults = {
+        "text_kwargs": {
+            "padding": False,
+            "return_token_type_ids": False,
+            "return_mm_token_type_ids": False,
+        },
+        "videos_kwargs": {"return_metadata": True},
+    }
+
+
+class ZFQwen3VLProcessor(ProcessorMixin):
+    r"""
+    Constructs a Qwen3VL processor which wraps a Qwen3VL image processor and a Qwen2 tokenizer into a single processor.
+    [`Qwen3VLProcessor`] offers all the functionalities of [`Qwen2VLImageProcessor`] and [`Qwen2TokenizerFast`]. See the
+    [`~Qwen3VLProcessor.__call__`] and [`~Qwen3VLProcessor.decode`] for more information.
+    Args:
+        image_processor ([`Qwen2VLImageProcessor`], *optional*):
+            The image processor is a required input.
+        tokenizer ([`Qwen2TokenizerFast`], *optional*):
+            The tokenizer is a required input.
+        video_processor ([`Qwen3VLVideoProcessor`], *optional*):
+            The video processor 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", "video_processor"]
+    image_processor_class = "AutoImageProcessor"
+    video_processor_class = "AutoVideoProcessor"
+    tokenizer_class = ("Qwen2Tokenizer", "Qwen2TokenizerFast")
+
+    def __init__(self, image_processor=None, tokenizer=None, video_processor=None, chat_template=None, **kwargs):
+        super().__init__(image_processor, tokenizer, video_processor, chat_template=chat_template)
+        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
+        self.image_token_id = (
+            tokenizer.image_token_id
+            if getattr(tokenizer, "image_token_id", None)
+            else tokenizer.convert_tokens_to_ids(self.image_token)
+        )
+        self.video_token_id = (
+            tokenizer.video_token_id
+            if getattr(tokenizer, "video_token_id", None)
+            else tokenizer.convert_tokens_to_ids(self.video_token)
+        )
+        self.vision_start_token = (
+            "<|vision_start|>" if not hasattr(tokenizer, "vision_start_token") else tokenizer.vision_start_token
+        )
+        self.vision_end_token = (
+            "<|vision_end|>" if not hasattr(tokenizer, "vision_end_token") else tokenizer.vision_end_token
+        )
+        self.vision_start_token_id = (
+            tokenizer.vision_start_token_id
+            if getattr(tokenizer, "vision_start_token_id", None)
+            else tokenizer.convert_tokens_to_ids(self.vision_start_token)
+        )
+        self.vision_end_token_id = (
+            tokenizer.vision_end_token_id
+            if getattr(tokenizer, "vision_end_token_id", None)
+            else tokenizer.convert_tokens_to_ids(self.vision_end_token)
+        )
+
+    def __call__(
+        self,
+        images: ImageInput = None,
+        text: Union[TextInput, PreTokenizedInput, list[TextInput], list[PreTokenizedInput]] = None,
+        videos: VideoInput = None,
+        **kwargs: Unpack[Qwen3VLProcessorKwargs],
+    ) -> 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 Qwen2TokenizerFast's [`~Qwen2TokenizerFast.__call__`] if `text` is not `None` to encode
+        the text. To prepare the vision inputs, this method forwards the `vision_infos` and `kwrags` arguments to
+        Qwen2VLImageProcessor's [`~Qwen2VLImageProcessor.__call__`] if `vision_infos` is not `None`.
+
+        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).
+            videos (`np.ndarray`, `torch.Tensor`, `list[np.ndarray]`, `list[torch.Tensor]`):
+                The image or batch of videos to be prepared. Each video can be a 4D NumPy array or PyTorch
+                tensor, or a nested list of 3D frames. Both channels-first and channels-last formats are supported.
+            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`.
+        """
+        output_kwargs = self._merge_kwargs(
+            Qwen3VLProcessorKwargs,
+            tokenizer_init_kwargs=self.tokenizer.init_kwargs,
+            **kwargs,
+        )
+        if images is not None:
+            image_inputs = self.image_processor(images=images, **output_kwargs["images_kwargs"])
+            image_grid_thw = image_inputs["image_grid_thw"]
+        else:
+            image_inputs = {}
+            image_grid_thw = None
+
+        if videos is not None:
+            videos_inputs = self.video_processor(videos=videos, **output_kwargs["videos_kwargs"])
+            video_grid_thw = videos_inputs["video_grid_thw"]
+            # If user has not requested video metadata, pop it
+            if "return_metadata" not in kwargs:
+                video_metadata = videos_inputs.pop("video_metadata")
+            else:
+                video_metadata = videos_inputs["video_metadata"]
+            video_grid_thw = videos_inputs["video_grid_thw"]
+        else:
+            videos_inputs = {}
+            video_grid_thw = None
+
+        if not isinstance(text, list):
+            text = [text]
+
+        text = text.copy()  # below lines change text in-place
+        if image_grid_thw is not None:
+            merge_length = self.image_processor.merge_size**2
+            index = 0
+            for i in range(len(text)):
+                while self.image_token in text[i]:
+                    num_image_tokens = image_grid_thw[index].prod() // merge_length
+                    text[i] = text[i].replace(self.image_token, "<|placeholder|>" * num_image_tokens, 1)
+                    index += 1
+                text[i] = text[i].replace("<|placeholder|>", self.image_token)
+
+        if video_grid_thw is not None:
+            merge_length = self.video_processor.merge_size**2
+            index = 0
+            for i in range(len(text)):
+                while self.video_token in text[i]:
+                    metadata = video_metadata[index]
+                    if metadata.fps is None:
+                        logger.warning_once(
+                            "Qwen3VL requires frame timestamps to construct prompts, but the `fps` of the input video could not be inferred. "
+                            "Probably `video_metadata` was missing from inputs and you passed pre-sampled frames. "
+                            "Defaulting to `fps=24`. Please provide `video_metadata` for more accurate results."
+                        )
+                        metadata.fps = 24 if metadata.fps is None else metadata.fps
+
+                    # if timestamps are not provided, calculate them
+                    curr_timestamp = self._calculate_timestamps(
+                        metadata.frames_indices,
+                        metadata.fps,
+                        self.video_processor.merge_size,
+                        self.video_processor.focus_size,
+                    )
+
+                    print(len(curr_timestamp), curr_timestamp)
+                    video_placeholder = ""
+                    frame_seqlen = video_grid_thw[index][1:].prod() // merge_length
+                    print(video_grid_thw)
+                    for frame_idx in range(video_grid_thw[index][0]):
+                        curr_time = curr_timestamp[frame_idx]
+                        video_placeholder += f"<{curr_time:.1f} seconds>"
+                        video_placeholder += (
+                            self.vision_start_token + "<|placeholder|>" * frame_seqlen + self.vision_end_token
+                        )
+                    if f"{self.vision_start_token}{self.video_token}{self.vision_end_token}" in text[i]:
+                        text[i] = text[i].replace(
+                            f"{self.vision_start_token}{self.video_token}{self.vision_end_token}", video_placeholder, 1
+                        )
+                    else:
+                        # vllm may input video token directly
+                        text[i] = text[i].replace(self.video_token, video_placeholder, 1)
+                    index += 1
+
+                text[i] = text[i].replace("<|placeholder|>", self.video_token)
+
+        return_tensors = output_kwargs["text_kwargs"].pop("return_tensors", None)
+        return_mm_token_type_ids = output_kwargs["text_kwargs"].pop("return_mm_token_type_ids", None)
+        text_inputs = self.tokenizer(text, **output_kwargs["text_kwargs"])
+        self._check_special_mm_tokens(text, text_inputs, modalities=["image", "video"])
+
+        if return_mm_token_type_ids:
+            array_ids = np.array(text_inputs["input_ids"])
+            mm_token_type_ids = np.zeros_like(text_inputs["input_ids"])
+            mm_token_type_ids[array_ids == self.image_token_id] = 1
+            text_inputs["mm_token_type_ids"] = mm_token_type_ids.tolist()
+
+        return BatchFeature(data={**text_inputs, **image_inputs, **videos_inputs}, tensor_type=return_tensors)
+
+    def _get_num_multimodal_tokens(self, image_sizes=None, video_sizes=None, **kwargs):
+        """
+        Computes the number of placeholder tokens needed for multimodal inputs with the given sizes.
+        Args:
+            image_sizes (`list[list[int]]`, *optional*):
+                The input sizes formatted as (height, width) per each image.
+            video_sizes (`list[list[int]]`, *optional*):
+                The input sizes formatted as (num_frames, height, width) per each video.
+        Returns:
+            `MultiModalData`: A `MultiModalData` object holding number of tokens per each of the provided
+            input modalities, along with other useful data.
+        """
+
+        vision_data = {}
+        if image_sizes is not None:
+            images_kwargs = Qwen3VLProcessorKwargs._defaults.get("images_kwargs", {})
+            images_kwargs.update(kwargs)
+            merge_size = images_kwargs.get("merge_size", None) or self.image_processor.merge_size
+
+            num_image_patches = [
+                self.image_processor.get_number_of_image_patches(*image_size, images_kwargs)
+                for image_size in image_sizes
+            ]
+            num_image_tokens = [(num_patches // merge_size**2) for num_patches in num_image_patches]
+            vision_data.update({"num_image_tokens": num_image_tokens, "num_image_patches": num_image_patches})
+
+        if video_sizes is not None:
+            videos_kwargs = Qwen3VLProcessorKwargs._defaults.get("videos_kwargs", {})
+            videos_kwargs.update(kwargs)
+            num_video_patches = [
+                self.video_processor.get_number_of_video_patches(*video_size, videos_kwargs)
+                for video_size in video_sizes
+            ]
+            num_video_tokens = [(num_patches // merge_size**2) for num_patches in num_video_patches]
+            vision_data["num_video_tokens"] = num_video_tokens
+
+        return MultiModalData(**vision_data)
+
+    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,
+        )
+
+    def _calculate_timestamps(
+        self,
+        indices: Union[list[int], np.ndarray],
+        video_fps: float,
+        merge_size: int = 2,
+        focus_size: int = 2,
+    ):
+        if not isinstance(indices, list):
+            indices = indices.tolist()
+        print(len(indices), indices)
+        b_size = merge_size * focus_size
+        if len(indices) % b_size != 0:
+            indices.extend(indices[-1] for _ in range(b_size - len(indices) % b_size))
+        print(len(indices), indices)
+        timestamps = [idx / video_fps for idx in indices]
+        # @JJJYmmm frames are merged by self.merge_size, \
+        # so we need to average the timestamps between the first/last frame within the temporal patch
+        timestamps = [
+            (timestamps[i] + timestamps[i + merge_size - 1]) / 2 for i in range(0, len(timestamps), merge_size)
+        ]
+        return timestamps
+
+
+__all__ = ["ZFQwen3VLProcessor"]

processor_config.json

@@ -0,0 +1,6 @@
+{
+  "auto_map": {
+    "AutoProcessor": "processing_qwen3_vl.ZFQwen3VLProcessor"
+  },
+  "processor_class": "ZFQwen3VLProcessor"
+}

special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
+size 11422654

tokenizer_config.json

@@ -0,0 +1,243 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "auto_map": {
+    "AutoProcessor": "processing_qwen3_vl.ZFQwen3VLProcessor"
+  },
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 262144,
+  "pad_token": "<|endoftext|>",
+  "processor_class": "ZFQwen3VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

video_preprocessor_config.json

@@ -0,0 +1,46 @@
+{
+  "auto_map": {
+    "AutoProcessor": "processing_qwen3_vl.ZFQwen3VLProcessor",
+    "AutoVideoProcessor": "video_processing_qwen3_vl.ZFQwen3VLVideoProcessor"
+  },
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "do_sample_frames": true,
+  "focus_size": 2,
+  "fps": 2,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "input_data_format": null,
+  "max_frames": 3600,
+  "merge_size": 2,
+  "min_frames": 4,
+  "num_frames": null,
+  "pad_size": null,
+  "patch_size": 16,
+  "processor_class": "ZFQwen3VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_metadata": false,
+  "size": {
+    "longest_edge": 235929600,
+    "shortest_edge": 4096
+  },
+  "temporal_patch_size": 2,
+  "video_metadata": null,
+  "video_processor_type": "ZFQwen3VLVideoProcessor"
+}

video_processing_qwen3_vl.py

@@ -0,0 +1,264 @@
+import math
+from typing import Optional, Union
+
+import numpy as np
+import torch
+
+from transformers.feature_extraction_utils import BatchFeature
+from transformers.image_utils import ChannelDimension, PILImageResampling, SizeDict, get_image_size
+from transformers.processing_utils import Unpack, VideosKwargs
+from transformers.utils import TensorType, add_start_docstrings, logging
+from transformers.video_processing_utils import BASE_VIDEO_PROCESSOR_DOCSTRING, BaseVideoProcessor
+from transformers.video_utils import VideoMetadata, group_videos_by_shape, reorder_videos
+
+
+logger = logging.get_logger(__name__)
+
+
+def smart_resize(
+    num_frames: int,
+    height: int,
+    width: int,
+    temporal_factor: int = 2,
+    factor: int = 32,
+    min_pixels: int = 128 * 128,
+    max_pixels: int = 16 * 16 * 2 * 2 * 2 * 6144,
+):
+    if num_frames < temporal_factor:
+        raise ValueError(f"t:{num_frames} must be larger than temporal_factor:{temporal_factor}")
+    if height < factor or width < factor:
+        raise ValueError(f"height:{height} or width:{width} must be larger than factor:{factor}")
+    elif max(height, width) / min(height, width) > 200:
+        raise ValueError(
+            f"absolute aspect ratio must be smaller than 200, got {max(height, width) / min(height, width)}"
+        )
+    h_bar = round(height / factor) * factor
+    w_bar = round(width / factor) * factor
+    t_bar = round(num_frames / temporal_factor) * temporal_factor
+
+    if t_bar * h_bar * w_bar > max_pixels:
+        beta = math.sqrt((num_frames * height * width) / max_pixels)
+        h_bar = max(factor, math.floor(height / beta / factor) * factor)
+        w_bar = max(factor, math.floor(width / beta / factor) * factor)
+    elif t_bar * h_bar * w_bar < min_pixels:
+        beta = math.sqrt(min_pixels / (num_frames * height * width))
+        h_bar = math.ceil(height * beta / factor) * factor
+        w_bar = math.ceil(width * beta / factor) * factor
+
+    return h_bar, w_bar
+
+
+class Qwen3VLVideoProcessorInitKwargs(VideosKwargs):
+    patch_size: Optional[int]
+    temporal_patch_size: Optional[int]
+    merge_size: Optional[int]
+    focus_size: Optional[int]
+    min_frames: Optional[int]
+    max_frames: Optional[int]
+
+
+@add_start_docstrings(
+    "Constructs a fast Qwen3-VL image processor that dynamically resizes videos based on the original videos.",
+    BASE_VIDEO_PROCESSOR_DOCSTRING,
+    """
+        patch_size (`int`, *optional*, defaults to 16):
+            The spacial patch size of the vision encoder.
+        temporal_patch_size (`int`, *optional*, defaults to 2):
+            The temporal patch size of the vision encoder.
+        merge_size (`int`, *optional*, defaults to 2):
+            The merge size of the vision encoder to llm encoder.
+    """,
+)
+class ZFQwen3VLVideoProcessor(BaseVideoProcessor):
+    resample = PILImageResampling.BICUBIC
+    size = {"shortest_edge": 128 * 32 * 32, "longest_edge": 32 * 32 * 768}
+    image_mean = [0.5, 0.5, 0.5]
+    image_std = [0.5, 0.5, 0.5]
+    do_resize = True
+    do_rescale = True
+    do_normalize = True
+    do_convert_rgb = True
+    patch_size = 16
+    temporal_patch_size = 2
+    merge_size = 2
+    focus_size = 2
+    fps = 2
+    min_frames = 4
+    max_frames = 768
+    do_sample_frames = True
+    valid_kwargs = Qwen3VLVideoProcessorInitKwargs
+    model_input_names = ["pixel_values_videos", "video_grid_thw"]
+
+    def __init__(self, **kwargs: Unpack[Qwen3VLVideoProcessorInitKwargs]):
+        super().__init__(**kwargs)
+        if self.size is not None and (
+            self.size.get("shortest_edge", None) is None or self.size.get("longest_edge", None) is None
+        ):
+            raise ValueError("size must contain 'shortest_edge' and 'longest_edge' keys.")
+
+    def _further_process_kwargs(
+        self,
+        size: Optional[SizeDict] = None,
+        **kwargs,
+    ) -> dict:
+        """
+        Update kwargs that need further processing before being validated
+        Can be overridden by subclasses to customize the processing of kwargs.
+        """
+        if size is not None and ("shortest_edge" not in size or "longest_edge" not in size):
+            raise ValueError("size must contain 'shortest_edge' and 'longest_edge' keys.")
+
+        return super()._further_process_kwargs(size=size, **kwargs)
+
+    def sample_frames(
+        self,
+        metadata: VideoMetadata,
+        num_frames: Optional[int] = None,
+        fps: Optional[Union[int, float]] = None,
+        **kwargs,
+    ):
+        """
+        Default sampling function which uniformly samples the desired number of frames between 0 and total number of frames.
+        If `fps` is passed along with metadata, `fps` frames per second are sampled uniformty. Arguments `num_frames`
+        and `fps` are mutually exclusive.
+
+        Args:
+            video (`torch.Tensor`):
+                Video that need to be sampled.
+            metadata (`VideoMetadata`):
+                Metadata of the video containing information about total duration, fps and total number of frames.
+            num_frames (`int`, *optional*):
+                Maximum number of frames to sample. Defaults to `self.num_frames`.
+            fps (`int` or `float`, *optional*):
+                Target frames to sample per second. Defaults to `self.fps`.
+        Returns:
+            torch.Tensor:
+                Sampled video frames.
+        """
+        if fps is not None and num_frames is not None:
+            raise ValueError("`num_frames` and `fps` are mutually exclusive arguments, please use only one!")
+
+        total_num_frames = metadata.total_num_frames
+        fps = fps if fps is not None else self.fps
+
+        # If num_frames is not given but fps is, calculate num_frames from fps
+        if num_frames is None and fps is not None:
+            if metadata.fps is None:
+                metadata.fps = 24
+                logger.warning_once(
+                    "Asked to sample `fps` frames per second but no video metadata was provided which is required when sampling with `fps`. "
+                    "Defaulting to `fps=24`. Please provide `video_metadata` for more accurate results."
+                )
+            num_frames = int(total_num_frames / metadata.fps * fps)
+            num_frames = min(min(max(num_frames, self.min_frames), self.max_frames), total_num_frames)
+
+        if num_frames is None:
+            num_frames = min(max(total_num_frames, self.min_frames), self.max_frames)
+
+        indices = np.linspace(0, total_num_frames - 1, num_frames).round().astype(int)
+
+        return indices
+
+    def _preprocess(
+        self,
+        videos: list[torch.Tensor],
+        do_convert_rgb: bool = True,
+        do_resize: bool = True,
+        size: Optional[SizeDict] = None,
+        interpolation: PILImageResampling = PILImageResampling.BICUBIC,
+        do_rescale: bool = True,
+        rescale_factor: float = 1 / 255.0,
+        do_normalize: bool = True,
+        image_mean: Optional[Union[float, list[float]]] = None,
+        image_std: Optional[Union[float, list[float]]] = None,
+        patch_size: Optional[int] = None,
+        temporal_patch_size: Optional[int] = None,
+        merge_size: Optional[int] = None,
+        focus_size: Optional[int] = None,
+        return_tensors: Optional[Union[str, TensorType]] = None,
+        **kwargs,
+    ):
+        grouped_videos, grouped_videos_index = group_videos_by_shape(videos)
+        resized_videos_grouped = {}
+
+        for shape, stacked_videos in grouped_videos.items():
+            B, T, C, H, W = stacked_videos.shape
+            num_frames, height, width = T, H, W
+            if do_resize:
+                resized_height, resized_width = smart_resize(
+                    num_frames=num_frames,
+                    height=height,
+                    width=width,
+                    temporal_factor=temporal_patch_size,
+                    factor=patch_size * merge_size * focus_size,
+                    min_pixels=size.shortest_edge,
+                    max_pixels=size.longest_edge,
+                )
+                stacked_videos = stacked_videos.view(B * T, C, H, W)
+                stacked_videos = self.resize(
+                    stacked_videos,
+                    size=SizeDict(height=resized_height, width=resized_width),
+                    interpolation=interpolation,
+                )
+                stacked_videos = stacked_videos.view(B, T, C, resized_height, resized_width)
+            resized_videos_grouped[shape] = stacked_videos
+        resized_videos = reorder_videos(resized_videos_grouped, grouped_videos_index)
+
+        # Group videos by size for further processing
+        # Needed in case do_resize is False, or resize returns videos with different sizes
+        grouped_videos, grouped_videos_index = group_videos_by_shape(resized_videos)
+        processed_videos_grouped = {}
+        processed_grids = {}
+        for shape, stacked_videos in grouped_videos.items():
+            resized_height, resized_width = get_image_size(stacked_videos[0], channel_dim=ChannelDimension.FIRST)
+
+            # Fused rescale and normalize
+            stacked_videos = self.rescale_and_normalize(
+                stacked_videos, do_rescale, rescale_factor, do_normalize, image_mean, image_std
+            )
+            patches = stacked_videos
+
+            temporal_focus_size = temporal_patch_size * focus_size
+            # Check that videos have `num_frames` divisible by `temporal_patch_size`
+            if res := patches.shape[1] % temporal_focus_size:
+                repeats = patches[:, -1:].repeat(1, temporal_focus_size - res, 1, 1, 1)
+                patches = torch.cat([patches, repeats], dim=1)
+            batch_size, grid_t, channel = patches.shape[:3]
+            grid_t = grid_t // temporal_patch_size
+            grid_h, grid_w = resized_height // patch_size, resized_width // patch_size
+
+            patches = patches.view(
+                batch_size,
+                grid_t,
+                temporal_patch_size,
+                channel,
+                grid_h // merge_size,
+                merge_size,
+                patch_size,
+                grid_w // merge_size,
+                merge_size,
+                patch_size,
+            )
+            patches = patches.permute(0, 1, 4, 7, 5, 8, 3, 2, 6, 9)
+            flatten_patches = patches.reshape(
+                batch_size,
+                grid_t * grid_h * grid_w,
+                channel * temporal_patch_size * patch_size * patch_size,
+            )
+
+            processed_videos_grouped[shape] = flatten_patches
+            processed_grids[shape] = [[grid_t, grid_h, grid_w]] * batch_size
+
+        processed_videos = reorder_videos(processed_videos_grouped, grouped_videos_index)
+        processed_grids = reorder_videos(processed_grids, grouped_videos_index)
+        pixel_values_videos = torch.cat(processed_videos, dim=0)
+        video_grid_thw = torch.tensor(processed_grids)
+        data = {
+            "pixel_values_videos": pixel_values_videos,
+            "video_grid_thw": video_grid_thw,
+        }
+
+        return BatchFeature(data=data, tensor_type=return_tensors)
+
+
+__all__ = ["ZFQwen3VLVideoProcessor"]