Upload 29 files

.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

added_tokens.json

@@ -0,0 +1,25 @@
+{
+  "</box>": 151651,
+  "</image>": 151647,
+  "</image_id>": 151659,
+  "</point>": 151655,
+  "</quad>": 151653,
+  "</ref>": 151649,
+  "</slice>": 151657,
+  "<box>": 151650,
+  "<image>": 151646,
+  "<image_id>": 151658,
+  "<point>": 151654,
+  "<quad>": 151652,
+  "<ref>": 151648,
+  "<slice>": 151656,
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|reserved_special_token_0|>": 151660,
+  "<|reserved_special_token_1|>": 151661,
+  "<|reserved_special_token_2|>": 151662,
+  "<|reserved_special_token_3|>": 151663,
+  "<|reserved_special_token_4|>": 151664,
+  "<|reserved_special_token_5|>": 151665
+}

chat_template.jinja

@@ -0,0 +1,6 @@
+{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system
+You are a helpful assistant.<|im_end|>
+' }}{% endif %}{{'<|im_start|>' + message['role'] + '
+' + message['content'] + '<|im_end|>' + '
+'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant
+' }}{% endif %}

config.json

@@ -0,0 +1,91 @@
+{
+  "architectures": [
+    "MiniCPMV"
+  ],
+  "attention_dropout": 0.0,
+  "auto_map": {
+    "AutoConfig": "configuration_minicpm.MiniCPMVConfig",
+    "AutoModel": "modeling_minicpmv.MiniCPMV",
+    "AutoModelForCausalLM": "modeling_minicpmv.MiniCPMV"
+  },
+  "batch_vision_input": true,
+  "bos_token_id": 151643,
+  "drop_vision_last_layer": false,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "image_size": 448,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "minicpmv",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "patch_size": 14,
+  "query_num": 64,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "slice_config": {
+    "max_slice_nums": 9,
+    "model_type": "minicpmv",
+    "patch_size": 14,
+    "scale_resolution": 448
+  },
+  "slice_mode": true,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.53.3",
+  "use_cache": true,
+  "use_image_id": true,
+  "use_sliding_window": false,
+  "version": 2.6,
+  "vision_batch_size": 16,
+  "vision_config": {
+    "attention_dropout": 0.0,
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 980,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_channels": 3,
+    "num_hidden_layers": 27,
+    "patch_size": 14
+  },
+  "vocab_size": 151666
+}

configuration_minicpm.py

@@ -0,0 +1,100 @@
+# coding=utf-8
+""" MiniCPMV model configuration"""
+
+import os
+from typing import Union
+
+from transformers.utils import logging
+from transformers import Qwen2Config, PretrainedConfig
+from .modeling_navit_siglip import SiglipVisionConfig
+
+logger = logging.get_logger(__name__)
+
+
+class MiniCPMVSliceConfig(PretrainedConfig):
+    model_type = "minicpmv"
+
+    def __init__(
+        self,
+        patch_size=14,
+        max_slice_nums=9,
+        scale_resolution=448,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        self.patch_size = patch_size
+        self.max_slice_nums = max_slice_nums
+        self.scale_resolution = scale_resolution
+
+    @classmethod
+    def from_pretrained(cls, pretrained_model_name_or_path: Union[str, os.PathLike], **kwargs) -> "PretrainedConfig":
+        cls._set_token_in_kwargs(kwargs)
+
+        config_dict, kwargs = cls.get_config_dict(pretrained_model_name_or_path, **kwargs)
+
+        if config_dict.get("model_type") == "minicpmv":
+            config_dict = config_dict["slice_config"]
+
+        if "model_type" in config_dict and hasattr(cls, "model_type") and config_dict["model_type"] != cls.model_type:
+            logger.warning(
+                f"You are using a model of type {config_dict['model_type']} to instantiate a model of type "
+                f"{cls.model_type}. This is not supported for all configurations of models and can yield errors."
+            )
+
+        return cls.from_dict(config_dict, **kwargs)
+
+
+
+class MiniCPMVConfig(Qwen2Config):
+    model_type = "minicpmv"
+    keys_to_ignore_at_inference = ["past_key_values"]
+
+    default_vision_config = {
+        "hidden_size": 1152,
+        "image_size": 980,
+        "intermediate_size": 4304,
+        "model_type": "siglip",
+        "num_attention_heads": 16,
+        "num_hidden_layers": 27,
+        "patch_size": 14,
+    }
+
+    def __init__(
+        self,
+        use_cache=True,
+        query_num=64,
+        image_size=448,
+        drop_vision_last_layer=True,
+        batch_vision_input=True,
+        slice_config=None,
+        vision_config=None,
+        use_image_id=True,
+        vision_batch_size=16,
+        **kwargs,
+    ):
+        self.use_cache = use_cache
+        self.query_num = query_num
+        self.image_size = image_size
+        self.drop_vision_last_layer = drop_vision_last_layer
+        self.batch_vision_input = batch_vision_input
+        self.use_image_id = use_image_id
+        self.vision_batch_size = vision_batch_size
+
+        if slice_config is None:
+            self.slice_config = MiniCPMVSliceConfig(max_slice_nums=1)
+        else:
+            self.slice_config = MiniCPMVSliceConfig(**slice_config)
+        self.slice_mode = True
+
+        # same as HuggingFaceM4/siglip-so400m-14-980-flash-attn2-navit add tgt_sizes
+        if vision_config is None:
+            self.vision_config = SiglipVisionConfig(**self.default_vision_config)
+            logger.info("vision_config is None, using default vision config")
+        elif isinstance(vision_config, dict):
+            self.vision_config = SiglipVisionConfig(**vision_config)
+        elif isinstance(vision_config, SiglipVisionConfig):
+            self.vision_config = vision_config
+
+        self.patch_size = self.vision_config.patch_size
+
+        super().__init__(**kwargs)

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "transformers_version": "4.53.3"
+}

image_processing_minicpmv.py

@@ -0,0 +1,418 @@
+from typing import Optional, Union, Dict, Any, List
+
+import torch
+import math
+import PIL.Image
+import PIL.ImageSequence
+import numpy as np
+import PIL
+from PIL import Image
+
+from transformers.utils import TensorType, requires_backends, is_torch_dtype, is_torch_device
+from transformers.image_processing_utils import BaseImageProcessor, BatchFeature
+from transformers import AutoImageProcessor
+from transformers.image_transforms import to_channel_dimension_format
+from transformers.image_utils import (
+    ImageInput, 
+    make_list_of_images, 
+    valid_images, 
+    is_torch_tensor, 
+    is_batched,
+    to_numpy_array, 
+    infer_channel_dimension_format,
+    ChannelDimension
+)
+
+
+def recursive_converter(converter, value):
+    if isinstance(value, list):
+        new_value = []
+        for v in value:
+            new_value += [recursive_converter(converter, v)]
+        return new_value
+    else:
+        return converter(value)
+
+
+class MiniCPMVBatchFeature(BatchFeature):
+    r"""
+    Extend from BatchFeature for supporting various image size
+    """
+    def __init__(self, data: Optional[Dict[str, Any]] = None, tensor_type: Union[None, str, TensorType] = None):
+        super().__init__(data)
+        self.convert_to_tensors(tensor_type=tensor_type)
+
+    def convert_to_tensors(self, tensor_type: Optional[Union[str, TensorType]] = None):
+        if tensor_type is None:
+            return self
+        
+        is_tensor, as_tensor = self._get_is_as_tensor_fns(tensor_type)
+
+        def converter(value):
+            try:
+                if not is_tensor(value):
+                    tensor = as_tensor(value)
+                    return tensor
+            except:  # noqa E722
+                if key == "overflowing_values":
+                    raise ValueError("Unable to create tensor returning overflowing values of different lengths. ")
+                raise ValueError(
+                    "Unable to create tensor, you should probably activate padding "
+                    "with 'padding=True' to have batched tensors with the same length."
+                )
+
+
+        for key, value in self.items():
+            self[key] = recursive_converter(converter, value)
+        return self
+            
+    def to(self, *args, **kwargs) -> "MiniCPMVBatchFeature":
+        requires_backends(self, ["torch"])
+        import torch
+
+        def cast_tensor(v):
+            # check if v is a floating point
+            if torch.is_floating_point(v):
+                # cast and send to device
+                return v.to(*args, **kwargs)
+            elif device is not None:
+                return v.to(device=device)
+            else:
+                return v
+
+        new_data = {}
+        device = kwargs.get("device")
+        # Check if the args are a device or a dtype
+        if device is None and len(args) > 0:
+            # device should be always the first argument
+            arg = args[0]
+            if is_torch_dtype(arg):
+                # The first argument is a dtype
+                pass
+            elif isinstance(arg, str) or is_torch_device(arg) or isinstance(arg, int):
+                device = arg
+            else:
+                # it's something else
+                raise ValueError(f"Attempting to cast a BatchFeature to type {str(arg)}. This is not supported.")
+        # We cast only floating point tensors to avoid issues with tokenizers casting `LongTensor` to `FloatTensor`
+        for k, v in self.items():
+            new_data[k] = recursive_converter(cast_tensor, v)
+        self.data = new_data
+        return self
+
+
+class MiniCPMVImageProcessor(BaseImageProcessor):
+    model_input_names = ["pixel_values"]
+
+    def __init__(
+            self, 
+            max_slice_nums=9,
+            scale_resolution=448,
+            patch_size=14,
+            **kwargs):
+        super().__init__(**kwargs)
+        self.max_slice_nums = max_slice_nums
+        self.scale_resolution = scale_resolution
+        self.patch_size = patch_size
+        self.use_image_id = kwargs.pop("use_image_id", False)
+        self.image_feature_size = kwargs.pop("image_feature_size", 64)
+        self.im_start_token = kwargs.pop("im_start", "<image>")
+        self.im_end_token = kwargs.pop("im_end", "</image>")
+        self.slice_start_token = kwargs.pop("slice_start", "<slice>")
+        self.slice_end_token = kwargs.pop("slice_end", "</slice>")
+        self.unk_token = kwargs.pop("unk", "<unk>")
+        self.im_id_start = kwargs.pop("im_id_start", "<image_id>")
+        self.im_id_end = kwargs.pop("im_id_end", "</image_id>")
+        self.slice_mode = kwargs.pop("slice_mode", True)
+        self.mean = np.array(kwargs.pop("norm_mean", [0.5, 0.5, 0.5]))
+        self.std = np.array(kwargs.pop("norm_std", [0.5, 0.5, 0.5]))
+        self.version = kwargs.pop("version", 2.0)
+
+    def ensure_divide(self, length, patch_size):
+        return max(round(length / patch_size) * patch_size, patch_size)
+
+    def find_best_resize(self,
+                         original_size,
+                         scale_resolution,
+                         patch_size,
+                         allow_upscale=False):
+        width, height = original_size
+        if (width * height >
+                scale_resolution * scale_resolution) or allow_upscale:
+            r = width / height
+            height = int(scale_resolution / math.sqrt(r))
+            width = int(height * r)
+        best_width = self.ensure_divide(width, patch_size)
+        best_height = self.ensure_divide(height, patch_size)
+        return (best_width, best_height)
+
+    def get_refine_size(self,
+                        original_size,
+                        grid,
+                        scale_resolution,
+                        patch_size,
+                        allow_upscale=False):
+        width, height = original_size
+        grid_x, grid_y = grid
+
+        refine_width = self.ensure_divide(width, grid_x)
+        refine_height = self.ensure_divide(height, grid_y)
+
+        grid_width = refine_width / grid_x
+        grid_height = refine_height / grid_y
+
+        best_grid_size = self.find_best_resize((grid_width, grid_height),
+                                               scale_resolution,
+                                               patch_size,
+                                               allow_upscale=allow_upscale)
+        refine_size = (best_grid_size[0] * grid_x, best_grid_size[1] * grid_y)
+        return refine_size
+
+    def split_to_patches(self, image, grid):
+        patches = []
+        width, height = image.size
+        grid_x = int(width / grid[0])
+        grid_y = int(height / grid[1])
+        for i in range(0, height, grid_y):
+            images = []
+            for j in range(0, width, grid_x):
+                box = (j, i, j + grid_x, i + grid_y)
+                patch = image.crop(box)
+                images.append(patch)
+            patches.append(images)
+        return patches
+
+    def slice_image(
+        self, image, max_slice_nums=9, scale_resolution=448, patch_size=14, never_split=False
+    ):
+        original_size = image.size
+        source_image = None
+        best_grid = self.get_sliced_grid(original_size, max_slice_nums, never_split)
+        patches = []
+
+        if best_grid is None:
+            # dont need to slice, upsample
+            best_size = self.find_best_resize(
+                original_size, scale_resolution, patch_size, allow_upscale=True
+            )
+            source_image = image.resize(best_size, resample=Image.Resampling.BICUBIC)
+        else:
+            # source image, down-sampling and ensure divided by patch_size
+            best_resize = self.find_best_resize(original_size, scale_resolution, patch_size)
+            source_image = image.copy().resize(best_resize, resample=Image.Resampling.BICUBIC)
+            refine_size = self.get_refine_size(
+                original_size, best_grid, scale_resolution, patch_size, allow_upscale=True
+            )
+            refine_image = image.resize(refine_size, resample=Image.Resampling.BICUBIC)
+            patches = self.split_to_patches(refine_image, best_grid)
+
+        return source_image, patches, best_grid
+
+    def get_grid_placeholder(self, grid):
+        if grid is None:
+            return ""
+        slice_image_placeholder = (
+            self.slice_start_token 
+            + self.unk_token * self.image_feature_size
+            + self.slice_end_token
+        )
+
+        cols = grid[0]
+        rows = grid[1]
+        slices = []
+        for i in range(rows):
+            lines = []
+            for j in range(cols):
+                lines.append(slice_image_placeholder)
+            slices.append("".join(lines))
+            
+        slice_placeholder = "\n".join(slices)
+        return slice_placeholder
+
+    def get_image_id_placeholder(self, idx=0):
+        return f"{self.im_id_start}{idx}{self.im_id_end}"
+    
+    def get_sliced_images(self, image, max_slice_nums=None):
+        slice_images = []
+
+        if not self.slice_mode:
+            return [image]
+
+        max_slice_nums = self.max_slice_nums if max_slice_nums is None else int(max_slice_nums)
+        assert max_slice_nums > 0 
+        source_image, patches, sliced_grid = self.slice_image(
+            image,
+            max_slice_nums,  # default: 9
+            self.scale_resolution,  # default: 448
+            self.patch_size  # default: 14
+        )
+
+        slice_images.append(source_image)
+        if len(patches) > 0:
+            for i in range(len(patches)):
+                for j in range(len(patches[0])):
+                    slice_images.append(patches[i][j])
+        return slice_images
+
+    def get_sliced_grid(self, image_size, max_slice_nums, nerver_split=False):
+        original_width, original_height = image_size
+        log_ratio = math.log(original_width / original_height)
+        ratio = original_width * original_height / (self.scale_resolution * self.scale_resolution)
+        multiple = min(math.ceil(ratio), max_slice_nums)
+        if multiple <= 1 or nerver_split:
+            return None
+        candidate_split_grids_nums = []
+        for i in [multiple - 1, multiple, multiple + 1]:
+            if i == 1 or i > max_slice_nums:
+                continue
+            candidate_split_grids_nums.append(i)
+        
+        candidate_grids = []
+        for split_grids_nums in candidate_split_grids_nums:
+            m = 1
+            while m <= split_grids_nums:
+                if split_grids_nums % m == 0:
+                    candidate_grids.append([m, split_grids_nums // m])
+                m += 1
+
+        best_grid = [1, 1]
+        min_error = float("inf")
+        for grid in candidate_grids:
+            error = abs(log_ratio - math.log(grid[0] / grid[1]))
+            if error < min_error:
+                best_grid = grid
+                min_error = error
+        
+        return best_grid
+    
+    def get_slice_image_placeholder(self, image_size, image_idx=0, max_slice_nums=None, use_image_id=None):
+        max_slice_nums = self.max_slice_nums if max_slice_nums is None else int(max_slice_nums)
+        assert max_slice_nums > 0        
+        grid = self.get_sliced_grid(image_size=image_size, max_slice_nums=max_slice_nums)
+
+        image_placeholder = (
+            self.im_start_token 
+            + self.unk_token * self.image_feature_size 
+            + self.im_end_token
+        )
+        use_image_id = self.use_image_id if use_image_id is None else bool(use_image_id)
+        if use_image_id:
+            final_placeholder = self.get_image_id_placeholder(image_idx) + image_placeholder
+        else:
+            final_placeholder = image_placeholder
+
+        if self.slice_mode:
+            final_placeholder = final_placeholder + self.get_grid_placeholder(grid=grid)
+        return final_placeholder
+        
+    def to_pil_image(self, image, rescale=None) -> PIL.Image.Image:
+        """
+        Converts `image` to a PIL Image. Optionally rescales it and puts the channel dimension back as the last axis if
+        needed.
+
+        Args:
+            image (`PIL.Image.Image` or `numpy.ndarray` or `torch.Tensor`):
+                The image to convert to the PIL Image format.
+            rescale (`bool`, *optional*):
+                Whether or not to apply the scaling factor (to make pixel values integers between 0 and 255). Will
+                default to `True` if the image type is a floating type, `False` otherwise.
+        """
+        if isinstance(image, PIL.Image.Image):
+            return image
+        if is_torch_tensor(image):
+            image = image.numpy()
+
+        if isinstance(image, np.ndarray):
+            if rescale is None:
+                # rescale default to the array being of floating type.
+                rescale = isinstance(image.flat[0], np.floating)
+            # If the channel as been moved to first dim, we put it back at the end.
+            if image.ndim == 3 and image.shape[0] in [1, 3]:
+                image = image.transpose(1, 2, 0)
+            if rescale:
+                image = image * 255
+            image = image.astype(np.uint8)
+            return PIL.Image.fromarray(image)
+        return image
+
+    def reshape_by_patch(self, image):
+        """
+        :param image: shape [3, H, W]
+        :param patch_size:
+        :return: [3, patch_size, HW/patch_size]
+        """
+        image = torch.from_numpy(image)
+        patch_size = self.patch_size
+        patches = torch.nn.functional.unfold(
+            image,
+            (patch_size, patch_size),
+            stride=(patch_size, patch_size)
+        )
+
+        patches = patches.reshape(image.size(0), patch_size, patch_size, -1)
+        patches = patches.permute(0, 1, 3, 2).reshape(image.size(0), patch_size, -1)
+        return patches.numpy()
+
+    def preprocess(
+            self, 
+            images: Union[Image.Image, List[Image.Image], List[List[Image.Image]]],
+            do_pad: Optional[bool] = True, # TODO: add pad for MiniCPM-Llama3-V-2_5
+            max_slice_nums: int = None,
+            return_tensors: Optional[Union[str, TensorType]] = None,
+            **kwargs
+        ) -> MiniCPMVBatchFeature:
+        if isinstance(images, Image.Image):
+            images_list = [[images]]
+        elif isinstance(images[0], Image.Image):
+            images_list = [images]
+        else:
+            images_list = images
+
+        new_images_list = []
+        image_sizes_list = []
+        tgt_sizes_list = []
+        
+        for _images in images_list:
+            if _images is None or len(_images) == 0:
+                new_images_list.append([])
+                image_sizes_list.append([])
+                tgt_sizes_list.append([])
+                continue             
+            if not valid_images(_images):
+                raise ValueError(
+                    "Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, "
+                    "torch.Tensor, tf.Tensor or jax.ndarray."
+                )
+            
+            _images = [self.to_pil_image(image).convert("RGB") for image in _images]
+            input_data_format = infer_channel_dimension_format(np.array(_images[0]))
+
+            new_images = []
+            image_sizes = [image.size for image in _images]
+            tgt_sizes = []
+            for image in _images:
+                image_patches = self.get_sliced_images(image, max_slice_nums)
+                image_patches = [to_numpy_array(image).astype(np.float32) / 255 for image in image_patches]
+                image_patches = [
+                    self.normalize(image=image, mean=self.mean, std=self.std, input_data_format=input_data_format)
+                        for image in image_patches
+                ]
+                image_patches = [
+                    to_channel_dimension_format(image, ChannelDimension.FIRST, input_channel_dim=input_data_format) 
+                        for image in image_patches
+                ]
+                for slice_image in image_patches:
+                    new_images.append(self.reshape_by_patch(slice_image))
+                    tgt_sizes.append(np.array((slice_image.shape[1] // self.patch_size, slice_image.shape[2] // self.patch_size)))
+
+            if tgt_sizes:
+                tgt_sizes = np.vstack(tgt_sizes)
+            
+            new_images_list.append(new_images)
+            image_sizes_list.append(image_sizes)
+            tgt_sizes_list.append(tgt_sizes)
+        return MiniCPMVBatchFeature(
+            data={"pixel_values": new_images_list, "image_sizes": image_sizes_list, "tgt_sizes": tgt_sizes_list}, tensor_type=return_tensors
+        )
+
+AutoImageProcessor.register("MiniCPMVImageProcessor", MiniCPMVImageProcessor)

modeling_navit_siglip.py

@@ -0,0 +1,938 @@
+# coding=utf-8
+# Copyright 2024 Google AI and The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+""" PyTorch Siglip model. """
+# Copied from  HuggingFaceM4/siglip-so400m-14-980-flash-attn2-navit and add tgt_sizes
+
+
+import os
+import math
+import warnings
+from dataclasses import dataclass
+from typing import Any, Optional, Tuple, Union
+
+import numpy as np
+import torch
+import torch.nn.functional as F
+import torch.utils.checkpoint
+from torch import nn
+from torch.nn.init import _calculate_fan_in_and_fan_out
+
+from transformers.activations import ACT2FN
+from transformers.modeling_attn_mask_utils import _prepare_4d_attention_mask
+from transformers.modeling_outputs import BaseModelOutput, BaseModelOutputWithPooling
+from transformers.modeling_utils import PreTrainedModel
+from transformers.configuration_utils import PretrainedConfig
+from transformers.utils import (
+    ModelOutput,
+    add_start_docstrings,
+    add_start_docstrings_to_model_forward,
+    is_flash_attn_2_available,
+    logging,
+    replace_return_docstrings,
+)
+from transformers.utils import logging
+
+logger = logging.get_logger(__name__)
+
+class SiglipVisionConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`SiglipVisionModel`]. It is used to instantiate a
+    Siglip vision encoder according to the specified arguments, defining the model architecture. Instantiating a
+    configuration with the defaults will yield a similar configuration to that of the vision encoder of the Siglip
+    [google/siglip-base-patch16-224](https://huggingface.co/google/siglip-base-patch16-224) architecture.
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+    Args:
+        hidden_size (`int`, *optional*, defaults to 768):
+            Dimensionality of the encoder layers and the pooler layer.
+        intermediate_size (`int`, *optional*, defaults to 3072):
+            Dimensionality of the "intermediate" (i.e., feed-forward) layer in the Transformer encoder.
+        num_hidden_layers (`int`, *optional*, defaults to 12):
+            Number of hidden layers in the Transformer encoder.
+        num_attention_heads (`int`, *optional*, defaults to 12):
+            Number of attention heads for each attention layer in the Transformer encoder.
+        num_channels (`int`, *optional*, defaults to 3):
+            Number of channels in the input images.
+        image_size (`int`, *optional*, defaults to 224):
+            The size (resolution) of each image.
+        patch_size (`int`, *optional*, defaults to 16):
+            The size (resolution) of each patch.
+        hidden_act (`str` or `function`, *optional*, defaults to `"gelu_pytorch_tanh"`):
+            The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`,
+            `"relu"`, `"selu"` and `"gelu_new"` ``"quick_gelu"` are supported.
+        layer_norm_eps (`float`, *optional*, defaults to 1e-06):
+            The epsilon used by the layer normalization layers.
+        attention_dropout (`float`, *optional*, defaults to 0.0):
+            The dropout ratio for the attention probabilities.
+    Example:
+    ```python
+    >>> from transformers import SiglipVisionConfig, SiglipVisionModel
+    >>> # Initializing a SiglipVisionConfig with google/siglip-base-patch16-224 style configuration
+    >>> configuration = SiglipVisionConfig()
+    >>> # Initializing a SiglipVisionModel (with random weights) from the google/siglip-base-patch16-224 style configuration
+    >>> model = SiglipVisionModel(configuration)
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip_vision_model"
+
+    def __init__(
+        self,
+        hidden_size=768,
+        intermediate_size=3072,
+        num_hidden_layers=12,
+        num_attention_heads=12,
+        num_channels=3,
+        image_size=224,
+        patch_size=16,
+        hidden_act="gelu_pytorch_tanh",
+        layer_norm_eps=1e-6,
+        attention_dropout=0.0,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+
+        self.hidden_size = hidden_size
+        self.intermediate_size = intermediate_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.num_channels = num_channels
+        self.patch_size = patch_size
+        self.image_size = image_size
+        self.attention_dropout = attention_dropout
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+
+    @classmethod
+    def from_pretrained(cls, pretrained_model_name_or_path: Union[str, os.PathLike], **kwargs) -> "PretrainedConfig":
+        cls._set_token_in_kwargs(kwargs)
+
+        config_dict, kwargs = cls.get_config_dict(pretrained_model_name_or_path, **kwargs)
+
+        # get the vision config dict if we are loading from SiglipConfig
+        if config_dict.get("model_type") == "siglip":
+            config_dict = config_dict["vision_config"]
+
+        if "model_type" in config_dict and hasattr(cls, "model_type") and config_dict["model_type"] != cls.model_type:
+            logger.warning(
+                f"You are using a model of type {config_dict['model_type']} to instantiate a model of type "
+                f"{cls.model_type}. This is not supported for all configurations of models and can yield errors."
+            )
+
+        return cls.from_dict(config_dict, **kwargs)
+        
+
+_CHECKPOINT_FOR_DOC = "google/siglip-base-patch16-224"
+
+SIGLIP_PRETRAINED_MODEL_ARCHIVE_LIST = [
+    "google/siglip-base-patch16-224",
+    # See all SigLIP models at https://huggingface.co/models?filter=siglip
+]
+
+if is_flash_attn_2_available():
+    from flash_attn import flash_attn_func, flash_attn_varlen_func
+    from flash_attn.bert_padding import index_first_axis, pad_input, unpad_input  # noqa
+
+
+# Copied from transformers.models.llama.modeling_llama._get_unpad_data
+def _get_unpad_data(attention_mask):
+    seqlens_in_batch = attention_mask.sum(dim=-1, dtype=torch.int32)
+    indices = torch.nonzero(attention_mask.flatten(), as_tuple=False).flatten()
+    max_seqlen_in_batch = seqlens_in_batch.max().item()
+    cu_seqlens = F.pad(torch.cumsum(seqlens_in_batch, dim=0, dtype=torch.torch.int32), (1, 0))
+    return (
+        indices,
+        cu_seqlens,
+        max_seqlen_in_batch,
+    )
+
+
+def _trunc_normal_(tensor, mean, std, a, b):
+    # Cut & paste from PyTorch official master until it's in a few official releases - RW
+    # Method based on https://people.sc.fsu.edu/~jburkardt/presentations/truncated_normal.pdf
+    def norm_cdf(x):
+        # Computes standard normal cumulative distribution function
+        return (1.0 + math.erf(x / math.sqrt(2.0))) / 2.0
+
+    if (mean < a - 2 * std) or (mean > b + 2 * std):
+        warnings.warn(
+            "mean is more than 2 std from [a, b] in nn.init.trunc_normal_. "
+            "The distribution of values may be incorrect.",
+            stacklevel=2,
+        )
+
+    # Values are generated by using a truncated uniform distribution and
+    # then using the inverse CDF for the normal distribution.
+    # Get upper and lower cdf values
+    l = norm_cdf((a - mean) / std)
+    u = norm_cdf((b - mean) / std)
+
+    # Uniformly fill tensor with values from [l, u], then translate to
+    # [2l-1, 2u-1].
+    tensor.uniform_(2 * l - 1, 2 * u - 1)
+
+    # Use inverse cdf transform for normal distribution to get truncated
+    # standard normal
+    if tensor.dtype in [torch.float16, torch.bfloat16]:
+        # The `erfinv_` op is not (yet?) defined in float16+cpu, bfloat16+gpu
+        og_dtype = tensor.dtype
+        tensor = tensor.to(torch.float32)
+        tensor.erfinv_()
+        tensor = tensor.to(og_dtype)
+    else:
+        tensor.erfinv_()
+
+    # Transform to proper mean, std
+    tensor.mul_(std * math.sqrt(2.0))
+    tensor.add_(mean)
+
+    # Clamp to ensure it's in the proper range
+    if tensor.dtype == torch.float16:
+        # The `clamp_` op is not (yet?) defined in float16+cpu
+        tensor = tensor.to(torch.float32)
+        tensor.clamp_(min=a, max=b)
+        tensor = tensor.to(torch.float16)
+    else:
+        tensor.clamp_(min=a, max=b)
+
+
+def trunc_normal_tf_(
+    tensor: torch.Tensor, mean: float = 0.0, std: float = 1.0, a: float = -2.0, b: float = 2.0
+) -> torch.Tensor:
+    """Fills the input Tensor with values drawn from a truncated
+    normal distribution. The values are effectively drawn from the
+    normal distribution :math:`\\mathcal{N}(\text{mean}, \text{std}^2)`
+    with values outside :math:`[a, b]` redrawn until they are within
+    the bounds. The method used for generating the random values works
+    best when :math:`a \\leq \text{mean} \\leq b`.
+    NOTE: this 'tf' variant behaves closer to Tensorflow / JAX impl where the
+    bounds [a, b] are applied when sampling the normal distribution with mean=0, std=1.0
+    and the result is subsquently scaled and shifted by the mean and std args.
+    Args:
+        tensor: an n-dimensional `torch.Tensor`
+        mean: the mean of the normal distribution
+        std: the standard deviation of the normal distribution
+        a: the minimum cutoff value
+        b: the maximum cutoff value
+    """
+    with torch.no_grad():
+        _trunc_normal_(tensor, 0, 1.0, a, b)
+        tensor.mul_(std).add_(mean)
+
+
+def variance_scaling_(tensor, scale=1.0, mode="fan_in", distribution="normal"):
+    fan_in, fan_out = _calculate_fan_in_and_fan_out(tensor)
+    if mode == "fan_in":
+        denom = fan_in
+    elif mode == "fan_out":
+        denom = fan_out
+    elif mode == "fan_avg":
+        denom = (fan_in + fan_out) / 2
+
+    variance = scale / denom
+
+    if distribution == "truncated_normal":
+        # constant is stddev of standard normal truncated to (-2, 2)
+        trunc_normal_tf_(tensor, std=math.sqrt(variance) / 0.87962566103423978)
+    elif distribution == "normal":
+        with torch.no_grad():
+            tensor.normal_(std=math.sqrt(variance))
+    elif distribution == "uniform":
+        bound = math.sqrt(3 * variance)
+        with torch.no_grad():
+            tensor.uniform_(-bound, bound)
+    else:
+        raise ValueError(f"invalid distribution {distribution}")
+
+
+def lecun_normal_(tensor):
+    variance_scaling_(tensor, mode="fan_in", distribution="truncated_normal")
+
+
+def default_flax_embed_init(tensor):
+    variance_scaling_(tensor, mode="fan_in", distribution="normal")
+
+
+@dataclass
+# Copied from transformers.models.clip.modeling_clip.CLIPVisionModelOutput with CLIP->Siglip
+class SiglipVisionModelOutput(ModelOutput):
+    """
+    Base class for vision model's outputs that also contains image embeddings of the pooling of the last hidden states.
+    Args:
+        image_embeds (`torch.FloatTensor` of shape `(batch_size, output_dim)` *optional* returned when model is initialized with `with_projection=True`):
+            The image embeddings obtained by applying the projection layer to the pooler_output.
+        last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
+            Sequence of hidden-states at the output of the last layer of the model.
+        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
+            Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
+            one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.
+            Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
+        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
+            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
+            sequence_length)`.
+            Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
+            heads.
+    """
+
+    image_embeds: Optional[torch.FloatTensor] = None
+    last_hidden_state: torch.FloatTensor = None
+    hidden_states: Optional[Tuple[torch.FloatTensor]] = None
+    attentions: Optional[Tuple[torch.FloatTensor]] = None
+
+
+class SiglipVisionEmbeddings(nn.Module):
+    def __init__(self, config: SiglipVisionConfig):
+        super().__init__()
+        self.config = config
+        self.embed_dim = config.hidden_size
+        self.image_size = config.image_size
+        self.patch_size = config.patch_size
+
+        self.patch_embedding = nn.Conv2d(
+            in_channels=config.num_channels,
+            out_channels=self.embed_dim,
+            kernel_size=self.patch_size,
+            stride=self.patch_size,
+            padding="valid",
+        )
+
+        self.num_patches_per_side = self.image_size // self.patch_size
+        self.num_patches = self.num_patches_per_side**2
+        self.num_positions = self.num_patches
+        self.position_embedding = nn.Embedding(self.num_positions, self.embed_dim)
+
+    def forward(self, pixel_values: torch.FloatTensor, patch_attention_mask: torch.BoolTensor, tgt_sizes: Optional[torch.IntTensor]=None) -> torch.Tensor:
+        batch_size = pixel_values.size(0)
+
+        patch_embeds = self.patch_embedding(pixel_values)
+        embeddings = patch_embeds.flatten(2).transpose(1, 2)
+
+        max_im_h, max_im_w = pixel_values.size(2), pixel_values.size(3)
+        max_nb_patches_h, max_nb_patches_w = max_im_h // self.patch_size, max_im_w // self.patch_size
+        boundaries = torch.arange(1 / self.num_patches_per_side, 1.0, 1 / self.num_patches_per_side)
+        position_ids = torch.full(
+            size=(
+                batch_size,
+                max_nb_patches_h * max_nb_patches_w,
+            ),
+            fill_value=0,
+        )
+
+        for batch_idx, p_attn_mask in enumerate(patch_attention_mask):
+            if tgt_sizes is not None:
+                nb_patches_h = tgt_sizes[batch_idx][0]
+                nb_patches_w = tgt_sizes[batch_idx][1]
+            else:
+                nb_patches_h = p_attn_mask[:, 0].sum()
+                nb_patches_w = p_attn_mask[0].sum()
+
+            fractional_coords_h = torch.arange(0, 1 - 1e-6, 1 / nb_patches_h)
+            fractional_coords_w = torch.arange(0, 1 - 1e-6, 1 / nb_patches_w)
+
+            bucket_coords_h = torch.bucketize(fractional_coords_h, boundaries, right=True)
+            bucket_coords_w = torch.bucketize(fractional_coords_w, boundaries, right=True)
+
+            pos_ids = (bucket_coords_h[:, None] * self.num_patches_per_side + bucket_coords_w).flatten()
+            position_ids[batch_idx][p_attn_mask.view(-1).cpu()] = pos_ids
+
+        position_ids = position_ids.to(self.position_embedding.weight.device)
+
+        embeddings = embeddings + self.position_embedding(position_ids)
+        return embeddings
+
+
+class SiglipAttention(nn.Module):
+    """Multi-headed attention from 'Attention Is All You Need' paper"""
+
+    # Copied from transformers.models.clip.modeling_clip.CLIPAttention.__init__
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.embed_dim = config.hidden_size
+        self.num_heads = config.num_attention_heads
+        self.head_dim = self.embed_dim // self.num_heads
+        if self.head_dim * self.num_heads != self.embed_dim:
+            raise ValueError(
+                f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim} and `num_heads`:"
+                f" {self.num_heads})."
+            )
+        self.scale = self.head_dim**-0.5
+        self.dropout = config.attention_dropout
+
+        self.k_proj = nn.Linear(self.embed_dim, self.embed_dim)
+        self.v_proj = nn.Linear(self.embed_dim, self.embed_dim)
+        self.q_proj = nn.Linear(self.embed_dim, self.embed_dim)
+        self.out_proj = nn.Linear(self.embed_dim, self.embed_dim)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = False,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        """Input shape: Batch x Time x Channel"""
+
+        batch_size, q_len, _ = hidden_states.size()
+
+        query_states = self.q_proj(hidden_states)
+        key_states = self.k_proj(hidden_states)
+        value_states = self.v_proj(hidden_states)
+
+        query_states = query_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        key_states = key_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        value_states = value_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+
+        k_v_seq_len = key_states.shape[-2]
+        attn_weights = torch.matmul(query_states, key_states.transpose(2, 3)) * self.scale
+
+        if attn_weights.size() != (batch_size, self.num_heads, q_len, k_v_seq_len):
+            raise ValueError(
+                f"Attention weights should be of size {(batch_size, self.num_heads, q_len, k_v_seq_len)}, but is"
+                f" {attn_weights.size()}"
+            )
+
+        if attention_mask is not None:
+            if attention_mask.size() != (batch_size, 1, q_len, k_v_seq_len):
+                raise ValueError(
+                    f"Attention mask should be of size {(batch_size, 1, q_len, k_v_seq_len)}, but is {attention_mask.size()}"
+                )
+            attn_weights = attn_weights + attention_mask
+
+        # upcast attention to fp32
+        attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_states.dtype)
+        attn_weights = nn.functional.dropout(attn_weights, p=self.dropout, training=self.training)
+        attn_output = torch.matmul(attn_weights, value_states)
+
+        if attn_output.size() != (batch_size, self.num_heads, q_len, self.head_dim):
+            raise ValueError(
+                f"`attn_output` should be of size {(batch_size, self.num_heads, q_len, self.head_dim)}, but is"
+                f" {attn_output.size()}"
+            )
+
+        attn_output = attn_output.transpose(1, 2).contiguous()
+        attn_output = attn_output.reshape(batch_size, q_len, self.embed_dim)
+
+        attn_output = self.out_proj(attn_output)
+
+        return attn_output, attn_weights
+
+
+class SiglipFlashAttention2(SiglipAttention):
+    """
+    Llama flash attention module. This module inherits from `LlamaAttention` as the weights of the module stays
+    untouched. The only required change would be on the forward pass where it needs to correctly call the public API of
+    flash attention and deal with padding tokens in case the input contains any of them.
+    """
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.is_causal = False  # Hack to make sure we don't use a causal mask
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_value: Optional[Tuple[torch.Tensor]] = None,
+        output_attentions: bool = False,
+        use_cache: bool = False,
+        **kwargs,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        output_attentions = False
+
+        bsz, q_len, _ = hidden_states.size()
+
+        query_states = self.q_proj(hidden_states)
+        key_states = self.k_proj(hidden_states)
+        value_states = self.v_proj(hidden_states)
+
+        # Flash attention requires the input to have the shape
+        # batch_size x seq_length x head_dim x hidden_dim
+        # therefore we just need to keep the original shape
+        query_states = query_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        key_states = key_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        value_states = value_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+
+        kv_seq_len = key_states.shape[-2]
+        if past_key_value is not None:
+            kv_seq_len += past_key_value.get_usable_length(kv_seq_len, self.layer_idx)
+        # cos, sin = self.rotary_emb(value_states, seq_len=kv_seq_len)
+        # query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, position_ids)
+
+        # if past_key_value is not None:
+        #     cache_kwargs = {"sin": sin, "cos": cos}  # Specific to RoPE models
+        #     key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)
+
+        # TODO: These transpose are quite inefficient but Flash Attention requires the layout [batch_size, sequence_length, num_heads, head_dim]. We would need to refactor the KV cache
+        # to be able to avoid many of these transpose/reshape/view.
+        query_states = query_states.transpose(1, 2)
+        key_states = key_states.transpose(1, 2)
+        value_states = value_states.transpose(1, 2)
+
+        dropout_rate = self.dropout if self.training else 0.0
+
+        # In PEFT, usually we cast the layer norms in float32 for training stability reasons
+        # therefore the input hidden states gets silently casted in float32. Hence, we need
+        # cast them back in the correct dtype just to be sure everything works as expected.
+        # This might slowdown training & inference so it is recommended to not cast the LayerNorms
+        # in fp32. (LlamaRMSNorm handles it correctly)
+
+        input_dtype = query_states.dtype
+        if input_dtype == torch.float32:
+            if torch.is_autocast_enabled():
+                target_dtype = torch.get_autocast_gpu_dtype()
+            # Handle the case where the model is quantized
+            elif hasattr(self.config, "_pre_quantization_dtype"):
+                target_dtype = self.config._pre_quantization_dtype
+            else:
+                target_dtype = self.q_proj.weight.dtype
+
+            logger.warning_once(
+                "The input hidden states seems to be silently casted in float32, this might be related to the fact"
+                " you have upcasted embedding or layer norm layers in float32. We will cast back the input in"
+                f" {target_dtype}."
+            )
+
+            query_states = query_states.to(target_dtype)
+            key_states = key_states.to(target_dtype)
+            value_states = value_states.to(target_dtype)
+
+        attn_output = self._flash_attention_forward(
+            query_states, key_states, value_states, attention_mask, q_len, dropout=dropout_rate
+        )
+
+        attn_output = attn_output.reshape(bsz, q_len, self.embed_dim).contiguous()
+        attn_output = self.out_proj(attn_output)
+
+        if not output_attentions:
+            attn_weights = None
+
+        return attn_output, attn_weights
+
+    def _flash_attention_forward(
+        self, query_states, key_states, value_states, attention_mask, query_length, dropout=0.0, softmax_scale=None
+    ):
+        """
+        Calls the forward method of Flash Attention - if the input hidden states contain at least one padding token
+        first unpad the input, then computes the attention scores and pad the final attention scores.
+        Args:
+            query_states (`torch.Tensor`):
+                Input query states to be passed to Flash Attention API
+            key_states (`torch.Tensor`):
+                Input key states to be passed to Flash Attention API
+            value_states (`torch.Tensor`):
+                Input value states to be passed to Flash Attention API
+            attention_mask (`torch.Tensor`):
+                The padding mask - corresponds to a tensor of size `(batch_size, seq_len)` where 0 stands for the
+                position of padding tokens and 1 for the position of non-padding tokens.
+            dropout (`int`, *optional*):
+                Attention dropout
+            softmax_scale (`float`, *optional*):
+                The scaling of QK^T before applying softmax. Default to 1 / sqrt(head_dim)
+        """
+
+        # TODO: Remove the `query_length != 1` check once Flash Attention for RoCm is bumped to 2.1. For details, please see the comment in LlamaFlashAttention2 __init__.
+        causal = self.is_causal and query_length != 1
+
+        # Contains at least one padding token in the sequence
+        if attention_mask is not None:
+            batch_size = query_states.shape[0]
+            query_states, key_states, value_states, indices_q, cu_seq_lens, max_seq_lens = self._upad_input(
+                query_states, key_states, value_states, attention_mask, query_length
+            )
+
+            cu_seqlens_q, cu_seqlens_k = cu_seq_lens
+            max_seqlen_in_batch_q, max_seqlen_in_batch_k = max_seq_lens
+
+            attn_output_unpad = flash_attn_varlen_func(
+                query_states,
+                key_states,
+                value_states,
+                cu_seqlens_q=cu_seqlens_q,
+                cu_seqlens_k=cu_seqlens_k,
+                max_seqlen_q=max_seqlen_in_batch_q,
+                max_seqlen_k=max_seqlen_in_batch_k,
+                dropout_p=dropout,
+                softmax_scale=softmax_scale,
+                causal=causal,
+            )
+
+            attn_output = pad_input(attn_output_unpad, indices_q, batch_size, query_length)
+        else:
+            attn_output = flash_attn_func(
+                query_states, key_states, value_states, dropout, softmax_scale=softmax_scale, causal=causal
+            )
+
+        return attn_output
+
+    def _upad_input(self, query_layer, key_layer, value_layer, attention_mask, query_length):
+        indices_k, cu_seqlens_k, max_seqlen_in_batch_k = _get_unpad_data(attention_mask)
+        batch_size, kv_seq_len, num_key_value_heads, head_dim = key_layer.shape
+
+        key_layer = index_first_axis(
+            key_layer.reshape(batch_size * kv_seq_len, num_key_value_heads, head_dim), indices_k
+        )
+        value_layer = index_first_axis(
+            value_layer.reshape(batch_size * kv_seq_len, num_key_value_heads, head_dim), indices_k
+        )
+        if query_length == kv_seq_len:
+            query_layer = index_first_axis(
+                query_layer.reshape(batch_size * kv_seq_len, self.num_heads, head_dim), indices_k
+            )
+            cu_seqlens_q = cu_seqlens_k
+            max_seqlen_in_batch_q = max_seqlen_in_batch_k
+            indices_q = indices_k
+        elif query_length == 1:
+            max_seqlen_in_batch_q = 1
+            cu_seqlens_q = torch.arange(
+                batch_size + 1, dtype=torch.int32, device=query_layer.device
+            )  # There is a memcpy here, that is very bad.
+            indices_q = cu_seqlens_q[:-1]
+            query_layer = query_layer.squeeze(1)
+        else:
+            # The -q_len: slice assumes left padding.
+            attention_mask = attention_mask[:, -query_length:]
+            query_layer, indices_q, cu_seqlens_q, max_seqlen_in_batch_q = unpad_input(query_layer, attention_mask)
+
+        return (
+            query_layer,
+            key_layer,
+            value_layer,
+            indices_q,
+            (cu_seqlens_q, cu_seqlens_k),
+            (max_seqlen_in_batch_q, max_seqlen_in_batch_k),
+        )
+
+
+# Copied from transformers.models.clip.modeling_clip.CLIPMLP with CLIP->Siglip
+class SiglipMLP(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.activation_fn = ACT2FN[config.hidden_act]
+        self.fc1 = nn.Linear(config.hidden_size, config.intermediate_size)
+        self.fc2 = nn.Linear(config.intermediate_size, config.hidden_size)
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        hidden_states = self.fc1(hidden_states)
+        hidden_states = self.activation_fn(hidden_states)
+        hidden_states = self.fc2(hidden_states)
+        return hidden_states
+
+
+# Copied from transformers.models.clip.modeling_clip.CLIPEncoderLayer with CLIP->Siglip
+class SiglipEncoderLayer(nn.Module):
+    def __init__(self, config: SiglipVisionConfig):
+        super().__init__()
+        self.embed_dim = config.hidden_size
+        self._use_flash_attention_2 = config._attn_implementation == "flash_attention_2"
+        self.self_attn = (
+            SiglipAttention(config)
+            if not self._use_flash_attention_2
+            else SiglipFlashAttention2(config)
+        )
+        self.layer_norm1 = nn.LayerNorm(self.embed_dim, eps=config.layer_norm_eps)
+        self.mlp = SiglipMLP(config)
+        self.layer_norm2 = nn.LayerNorm(self.embed_dim, eps=config.layer_norm_eps)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: torch.Tensor,
+        output_attentions: Optional[bool] = False,
+    ) -> Tuple[torch.FloatTensor]:
+        """
+        Args:
+            hidden_states (`torch.FloatTensor`):
+                Input to the layer of shape `(batch, seq_len, embed_dim)`.
+            attention_mask (`torch.FloatTensor`):
+                Attention mask of shape `(batch, 1, q_len, k_v_seq_len)` where padding elements are indicated by very large negative values.
+            output_attentions (`bool`, *optional*, defaults to `False`):
+                Whether or not to return the attentions tensors of all attention layers. See `attentions` under
+                returned tensors for more detail.
+        """
+        residual = hidden_states
+
+        hidden_states = self.layer_norm1(hidden_states)
+        hidden_states, attn_weights = self.self_attn(
+            hidden_states=hidden_states,
+            attention_mask=attention_mask,
+            output_attentions=output_attentions,
+        )
+        hidden_states = residual + hidden_states
+
+        residual = hidden_states
+        hidden_states = self.layer_norm2(hidden_states)
+        hidden_states = self.mlp(hidden_states)
+        hidden_states = residual + hidden_states
+
+        outputs = (hidden_states,)
+
+        if output_attentions:
+            outputs += (attn_weights,)
+
+        return outputs
+
+
+class SiglipPreTrainedModel(PreTrainedModel):
+    """
+    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
+    models.
+    """
+
+    config_class = SiglipVisionConfig
+    base_model_prefix = "siglip"
+    supports_gradient_checkpointing = True
+
+    def _init_weights(self, module):
+        """Initialize the weights"""
+
+        if isinstance(module, SiglipVisionEmbeddings):
+            width = self.config.hidden_size
+            nn.init.normal_(module.position_embedding.weight, std=1 / np.sqrt(width))
+        elif isinstance(module, nn.Embedding):
+            default_flax_embed_init(module.weight)
+        elif isinstance(module, SiglipAttention):
+            nn.init.normal_(module.q_proj.weight)
+            nn.init.normal_(module.k_proj.weight)
+            nn.init.normal_(module.v_proj.weight)
+            nn.init.normal_(module.out_proj.weight)
+            nn.init.zeros_(module.q_proj.bias)
+            nn.init.zeros_(module.k_proj.bias)
+            nn.init.zeros_(module.v_proj.bias)
+            nn.init.zeros_(module.out_proj.bias)
+        elif isinstance(module, SiglipMLP):
+            nn.init.normal_(module.fc1.weight)
+            nn.init.normal_(module.fc2.weight)
+            nn.init.normal_(module.fc1.bias, std=1e-6)
+            nn.init.normal_(module.fc2.bias, std=1e-6)
+        elif isinstance(module, (nn.Linear, nn.Conv2d)):
+            lecun_normal_(module.weight)
+            if module.bias is not None:
+                nn.init.zeros_(module.bias)
+        elif isinstance(module, nn.LayerNorm):
+            module.bias.data.zero_()
+            module.weight.data.fill_(1.0)
+
+
+SIGLIP_START_DOCSTRING = r"""
+    This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the
+    library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
+    etc.)
+    This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
+    Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
+    and behavior.
+    Parameters:
+        config ([`SiglipVisionConfig`]): Model configuration class with all the parameters of the model.
+            Initializing with a config file does not load the weights associated with the model, only the
+            configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights.
+"""
+
+
+SIGLIP_VISION_INPUTS_DOCSTRING = r"""
+    Args:
+        pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)`):
+            Pixel values. Padding will be ignored by default should you provide it. Pixel values can be obtained using
+            [`AutoImageProcessor`]. See [`CLIPImageProcessor.__call__`] for details.
+        output_attentions (`bool`, *optional*):
+            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
+            tensors for more detail.
+        output_hidden_states (`bool`, *optional*):
+            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
+            more detail.
+        return_dict (`bool`, *optional*):
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+"""
+
+
+# Copied from transformers.models.clip.modeling_clip.CLIPEncoder with CLIP->Siglip
+class SiglipEncoder(nn.Module):
+    """
+    Transformer encoder consisting of `config.num_hidden_layers` self attention layers. Each layer is a
+    [`SiglipEncoderLayer`].
+    Args:
+        config: SiglipConfig
+    """
+
+    def __init__(self, config: SiglipVisionConfig):
+        super().__init__()
+        self.config = config
+        self.layers = nn.ModuleList([SiglipEncoderLayer(config) for _ in range(config.num_hidden_layers)])
+        self.gradient_checkpointing = False
+
+    # Ignore copy
+    def forward(
+        self,
+        inputs_embeds,
+        attention_mask: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutput]:
+        r"""
+        Args:
+            inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
+                Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation.
+                This is useful if you want more control over how to convert `input_ids` indices into associated vectors
+                than the model's internal embedding lookup matrix.
+            attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
+                Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
+                - 1 for tokens that are **not masked**,
+                - 0 for tokens that are **masked**.
+                [What are attention masks?](../glossary#attention-mask)
+            output_attentions (`bool`, *optional*):
+                Whether or not to return the attentions tensors of all attention layers. See `attentions` under
+                returned tensors for more detail.
+            output_hidden_states (`bool`, *optional*):
+                Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
+                for more detail.
+            return_dict (`bool`, *optional*):
+                Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+        """
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        encoder_states = () if output_hidden_states else None
+        all_attentions = () if output_attentions else None
+
+        hidden_states = inputs_embeds
+        for encoder_layer in self.layers:
+            if output_hidden_states:
+                encoder_states = encoder_states + (hidden_states,)
+            if self.gradient_checkpointing and self.training:
+                layer_outputs = self._gradient_checkpointing_func(
+                    encoder_layer.__call__,
+                    hidden_states,
+                    attention_mask,
+                    output_attentions,
+                )
+            else:
+                layer_outputs = encoder_layer(
+                    hidden_states,
+                    attention_mask,
+                    output_attentions=output_attentions,
+                )
+
+            hidden_states = layer_outputs[0]
+
+            if output_attentions:
+                all_attentions = all_attentions + (layer_outputs[1],)
+
+        if output_hidden_states:
+            encoder_states = encoder_states + (hidden_states,)
+
+        if not return_dict:
+            return tuple(v for v in [hidden_states, encoder_states, all_attentions] if v is not None)
+        return BaseModelOutput(
+            last_hidden_state=hidden_states, hidden_states=encoder_states, attentions=all_attentions
+        )
+
+@add_start_docstrings(
+    """The vision model from SigLIP without any head or projection on top.""",
+    SIGLIP_START_DOCSTRING
+)
+class SiglipVisionTransformer(SiglipPreTrainedModel):
+    config_class = SiglipVisionConfig
+    main_input_name = "pixel_values"
+    _supports_flash_attn_2 = True
+    _no_split_modules = []
+
+    def __init__(self, config: SiglipVisionConfig):
+        super().__init__(config)
+        self.config = config
+        embed_dim = config.hidden_size
+
+        self.embeddings = SiglipVisionEmbeddings(config)
+        self.encoder = SiglipEncoder(config)
+        self.post_layernorm = nn.LayerNorm(embed_dim, eps=config.layer_norm_eps)
+        self._use_flash_attention_2 = config._attn_implementation == "flash_attention_2"
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_input_embeddings(self) -> nn.Module:
+        return self.embeddings.patch_embedding
+
+    @add_start_docstrings_to_model_forward(SIGLIP_VISION_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=SiglipVisionConfig)
+    def forward(
+        self,
+        pixel_values,
+        patch_attention_mask: Optional[torch.BoolTensor] = None,
+        tgt_sizes: Optional[torch.IntTensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        Returns:
+        """
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        batch_size = pixel_values.size(0)
+        if patch_attention_mask is None:
+            patch_attention_mask = torch.ones(
+                size=(
+                    batch_size,
+                    pixel_values.size(2) // self.config.patch_size,
+                    pixel_values.size(3) // self.config.patch_size,
+                ),
+                dtype=torch.bool,
+                device=pixel_values.device,
+            )
+
+        hidden_states = self.embeddings(pixel_values=pixel_values, patch_attention_mask=patch_attention_mask, tgt_sizes=tgt_sizes)
+
+        patch_attention_mask = patch_attention_mask.view(batch_size, -1)
+        # The call to `_upad_input` in `_flash_attention_forward` is expensive
+        # So when the `patch_attention_mask` is full of 1s (i.e. attending to the whole sequence),
+        # avoiding passing the attention_mask, which is equivalent to attending to the full sequence
+        if not torch.any(~patch_attention_mask):
+            attention_mask=None
+        else:
+            attention_mask = (
+                _prepare_4d_attention_mask(patch_attention_mask, hidden_states.dtype)
+                if not self._use_flash_attention_2
+                else patch_attention_mask
+            )
+
+        encoder_outputs = self.encoder(
+            inputs_embeds=hidden_states,
+            attention_mask=attention_mask,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        last_hidden_state = encoder_outputs[0]
+        last_hidden_state = self.post_layernorm(last_hidden_state)
+
+        if not return_dict:
+            return (last_hidden_state, None) + encoder_outputs[1:]
+
+        return BaseModelOutputWithPooling(
+            last_hidden_state=last_hidden_state,
+            pooler_output=None,
+            hidden_states=encoder_outputs.hidden_states,
+            attentions=encoder_outputs.attentions,
+        )

openvino_config.json

@@ -0,0 +1,41 @@
+{
+  "dtype": "int4",
+  "input_info": null,
+  "optimum_version": "2.1.0",
+  "output_attentions": false,
+  "quantization_config": {
+    "dataset": null,
+    "default_config": {
+      "quant_method": "default"
+    },
+    "ignored_scope": null,
+    "num_samples": null,
+    "processor": "openbmb/MiniCPM-V-2_6",
+    "quantization_configs": {
+      "lm_model": {
+        "all_layers": null,
+        "backup_precision": null,
+        "bits": 4,
+        "dataset": null,
+        "dtype": "int4",
+        "gptq": null,
+        "group_size": 128,
+        "group_size_fallback": null,
+        "ignored_scope": null,
+        "lora_correction": null,
+        "num_samples": null,
+        "processor": "openbmb/MiniCPM-V-2_6",
+        "quant_method": "default",
+        "ratio": 1.0,
+        "scale_estimation": null,
+        "sensitivity_metric": null,
+        "statistics_path": null,
+        "sym": false,
+        "tokenizer": "openbmb/MiniCPM-V-2_6"
+      }
+    },
+    "tokenizer": "openbmb/MiniCPM-V-2_6"
+  },
+  "save_onnx_model": false,
+  "transformers_version": "4.53.3"
+}

openvino_detokenizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac385c728a001ae9702edc5792ba172cce1a77baa7831a501a822e4b152188a7
+size 2189697

openvino_detokenizer.xml

@@ -0,0 +1,220 @@
+<?xml version="1.0"?>
+<net name="detokenizer" version="11">
+	<layers>
+		<layer id="0" name="Parameter_126966" type="Parameter" version="opset1">
+			<data shape="?,?" element_type="i64" />
+			<output>
+				<port id="0" precision="I64" names="Parameter_126966">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="1" name="Convert_127165" type="Convert" version="opset1">
+			<data destination_type="i32" />
+			<input>
+				<port id="0" precision="I64">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="2" name="Constant_126968" type="Const" version="opset1">
+			<data element_type="i32" shape="151666" offset="0" size="606664" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151666</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="3" name="Constant_126970" type="Const" version="opset1">
+			<data element_type="i32" shape="151666" offset="606664" size="606664" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151666</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="4" name="Constant_126972" type="Const" version="opset1">
+			<data element_type="u8" shape="976273" offset="1213328" size="976273" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>976273</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="5" name="Slice_126977" type="Const" version="opset1">
+			<data element_type="i32" shape="24" offset="2189601" size="96" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>24</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="6" name="VocabDecoder_126979" type="VocabDecoder" version="extension">
+			<data skip_tokens="" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>151666</dim>
+				</port>
+				<port id="2" precision="I32">
+					<dim>151666</dim>
+				</port>
+				<port id="3" precision="U8">
+					<dim>976273</dim>
+				</port>
+				<port id="4" precision="I32">
+					<dim>24</dim>
+				</port>
+			</input>
+			<output>
+				<port id="5" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="6" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="7" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="8" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="9" precision="U8">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="7" name="FuzeRagged_126980" type="FuzeRagged" version="extension">
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="3" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="4" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="5" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="8" name="UTF8Validate_126981" type="UTF8Validate" version="extension">
+			<data replace_mode="true" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="U8">
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="3" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="4" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="5" precision="U8">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="9" name="StringTensorPack_126982" type="StringTensorPack" version="opset15">
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="U8">
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="3" precision="STRING" names="Result_126983,string_output">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="10" name="Result_126983" type="Result" version="opset1" output_names="Result_126983,string_output">
+			<input>
+				<port id="0" precision="STRING">
+					<dim>-1</dim>
+				</port>
+			</input>
+		</layer>
+	</layers>
+	<edges>
+		<edge from-layer="0" from-port="0" to-layer="1" to-port="0" />
+		<edge from-layer="1" from-port="1" to-layer="6" to-port="0" />
+		<edge from-layer="2" from-port="0" to-layer="6" to-port="1" />
+		<edge from-layer="3" from-port="0" to-layer="6" to-port="2" />
+		<edge from-layer="4" from-port="0" to-layer="6" to-port="3" />
+		<edge from-layer="5" from-port="0" to-layer="6" to-port="4" />
+		<edge from-layer="6" from-port="5" to-layer="7" to-port="0" />
+		<edge from-layer="6" from-port="6" to-layer="7" to-port="1" />
+		<edge from-layer="6" from-port="7" to-layer="7" to-port="2" />
+		<edge from-layer="6" from-port="8" to-layer="7" to-port="3" />
+		<edge from-layer="6" from-port="9" to-layer="8" to-port="2" />
+		<edge from-layer="7" from-port="4" to-layer="8" to-port="0" />
+		<edge from-layer="7" from-port="5" to-layer="8" to-port="1" />
+		<edge from-layer="8" from-port="3" to-layer="9" to-port="0" />
+		<edge from-layer="8" from-port="4" to-layer="9" to-port="1" />
+		<edge from-layer="8" from-port="5" to-layer="9" to-port="2" />
+		<edge from-layer="9" from-port="3" to-layer="10" to-port="0" />
+	</edges>
+	<rt_info>
+		<add_attention_mask value="True" />
+		<add_prefix_space />
+		<add_special_tokens value="True" />
+		<bos_token_id value="151644" />
+		<chat_template value="{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '&lt;|im_start|>system&#10;You are a helpful assistant.&lt;|im_end|>&#10;' }}{% endif %}{{'&lt;|im_start|>' + message['role'] + '&#10;' + message['content'] + '&lt;|im_end|>' + '&#10;'}}{% endfor %}{% if add_generation_prompt %}{{ '&lt;|im_start|>assistant&#10;' }}{% endif %}" />
+		<clean_up_tokenization_spaces />
+		<detokenizer_input_type value="i64" />
+		<eos_token_id value="151645" />
+		<handle_special_tokens_with_re />
+		<max_length />
+		<number_of_inputs value="1" />
+		<openvino_tokenizers_version value="2026.0.0.0-632-47cea02a2d4" />
+		<openvino_version value="2026.0.0-20965-c6d6a13a886-releases/2026/0" />
+		<original_post_processor_template value="{&quot;type&quot;: &quot;ByteLevel&quot;, &quot;add_prefix_space&quot;: false, &quot;trim_offsets&quot;: false, &quot;use_regex&quot;: false}" />
+		<original_tokenizer_class value="&lt;class 'transformers_modules.openbmb.MiniCPM-V-2_6.6c04d9e3022bcff6e6738dfb1fc19a5cfd2a855f.tokenization_minicpmv_fast.MiniCPMVTokenizerFast'>" />
+		<pad_token_id value="151643" />
+		<processed_post_processor_template value="{&quot;single&quot;: {&quot;ids&quot;: [-1], &quot;type_ids&quot;: [0]}, &quot;pair&quot;: {&quot;ids&quot;: [-1, -2], &quot;type_ids&quot;: [0, 0]}}" />
+		<skip_special_tokens value="True" />
+		<streaming_detokenizer value="False" />
+		<tokenizer_output_type value="i64" />
+		<tokenizers_version value="0.21.4" />
+		<transformers_version value="4.53.3" />
+		<use_max_padding value="False" />
+		<use_sentencepiece_backend value="False" />
+		<utf8_replace_mode value="replace" />
+		<with_detokenizer value="True" />
+	</rt_info>
+</net>

openvino_language_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97b0235d92ba01ba2935ba9ef4309b3cc4a32e7f557050a9285a9b7d8931dff7
+size 3935193003

openvino_resampler_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b98176a4e02e50478e4924b245b0d754a2000fc78b1677d11c40e4626e0afdc8
+size 69429468

openvino_resampler_model.xml

@@ -0,0 +1,1733 @@
+<?xml version="1.0"?>
+<net name="Model0" version="11">
+	<layers>
+		<layer id="0" name="image_feature" type="Parameter" version="opset1">
+			<data shape="?,?,?" element_type="f32" />
+			<output>
+				<port id="0" precision="FP32" names="image_feature">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="1" name="pos_embed" type="Parameter" version="opset1">
+			<data shape="?,?,?" element_type="f32" />
+			<output>
+				<port id="0" precision="FP32" names="pos_embed">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="2" name="key_padding_mask" type="Parameter" version="opset1">
+			<data shape="?,?" element_type="f32" />
+			<output>
+				<port id="0" precision="FP32" names="key_padding_mask">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="3" name="Constant_9686" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="0" size="32" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="4" name="__module.attn/aten::view/Reshape_1" type="Reshape" version="opset1">
+			<data special_zero="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>4</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="111">
+					<dim>-1</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="5" name="__module.attn/aten::expand/Abs" type="Const" version="opset1">
+			<data element_type="i64" shape="4" offset="32" size="32" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="6" name="__module.attn/aten::expand/Broadcast" type="Broadcast" version="opset3">
+			<data mode="bidirectional" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>4</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="113">
+					<dim>-1</dim>
+					<dim>28</dim>
+					<dim>1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="7" name="ShapeOf_9654" type="ShapeOf" version="opset3">
+			<data output_type="i64" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="I64">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="8" name="Constant_9655" type="Const" version="opset1">
+			<data element_type="i64" shape="1" offset="64" size="8" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="9" name="Constant_9656" type="Const" version="opset1">
+			<data element_type="i64" shape="" offset="64" size="8" />
+			<output>
+				<port id="0" precision="I64" />
+			</output>
+		</layer>
+		<layer id="10" name="Gather_9657" type="Gather" version="opset8">
+			<data batch_dims="0" />
+			<input>
+				<port id="0" precision="I64">
+					<dim>3</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>1</dim>
+				</port>
+				<port id="2" precision="I64" />
+			</input>
+			<output>
+				<port id="3" precision="I64" names="12,14,73">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="11" name="Constant_7566" type="Const" version="opset1">
+			<data element_type="i64" shape="1" offset="72" size="8" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="12" name="__module.attn/aten::mul/Multiply" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="I64" names="104,105,114,115,92,93,98,99">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="13" name="Constant_326" type="Const" version="opset1">
+			<data element_type="i64" shape="1" offset="80" size="8" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="14" name="Constant_9665" type="Const" version="opset1">
+			<data element_type="i64" shape="1" offset="80" size="8" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="15" name="Constant_9666" type="Const" version="opset1">
+			<data element_type="i64" shape="" offset="64" size="8" />
+			<output>
+				<port id="0" precision="I64" />
+			</output>
+		</layer>
+		<layer id="16" name="Gather_9667" type="Gather" version="opset8">
+			<data batch_dims="0" />
+			<input>
+				<port id="0" precision="I64">
+					<dim>3</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>1</dim>
+				</port>
+				<port id="2" precision="I64" />
+			</input>
+			<output>
+				<port id="3" precision="I64" names="103,109,97">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="17" name="__module.attn/prim::ListConstruct_2" type="Concat" version="opset1">
+			<data axis="0" />
+			<input>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>1</dim>
+				</port>
+				<port id="2" precision="I64">
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="3" precision="I64">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="18" name="__module.attn/aten::reshape/Reshape" type="Reshape" version="opset1">
+			<data special_zero="false" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>28</dim>
+					<dim>1</dim>
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="117,key_padding_mask_1">
+					<dim>-1</dim>
+					<dim>1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="19" name="aten::unsqueeze/Unsqueeze" type="Const" version="opset1">
+			<data element_type="f32" shape="64, 1, 3584" offset="88" size="917504" />
+			<output>
+				<port id="0" precision="FP32" names="24">
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="20" name="Constant_891" type="Const" version="opset1">
+			<data element_type="i64" shape="1" offset="80" size="8" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="21" name="Constant_898" type="Const" version="opset1">
+			<data element_type="i64" shape="1" offset="80" size="8" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="22" name="prim::ListConstruct/Concat" type="Concat" version="opset1">
+			<data axis="0" />
+			<input>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>1</dim>
+				</port>
+				<port id="2" precision="I64">
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="3" precision="I64" names="27">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="23" name="aten::repeat/Tile" type="Tile" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="28,query.1">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="24" name="__module.attn/prim::ListUnpack/VariadicSplit.0" type="Const" version="opset1">
+			<data element_type="i8" shape="3584, 3584" offset="917592" size="12845056" />
+			<output>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="25" name="Convert_1993077" type="Convert" version="opset1">
+			<data destination_type="f16" />
+			<input>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="26" name="__module.attn/prim::ListUnpack/VariadicSplit.0/scale" type="Const" version="opset1">
+			<data element_type="f16" shape="3584, 1" offset="13762648" size="7168" />
+			<output>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="27" name="__module.attn/prim::ListUnpack/VariadicSplit.0/fq_weights_1" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="28" name="__module.attn/prim::ListUnpack/VariadicSplit.0/fq_weights_1/convert" type="Convert" version="opset1">
+			<data destination_type="f32" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="29" name="__module.attn/aten::linear/MatMul_1" type="MatMul" version="opset1">
+			<data transpose_a="false" transpose_b="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="30" name="Constant_9232" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 3584" offset="13769816" size="14336" />
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="31" name="__module.attn/aten::linear/Add_1" type="Add" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="89,q.3">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="32" name="Constant_9687" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="13784152" size="24" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="33" name="__module.attn/aten::view/Reshape_2" type="Reshape" version="opset1">
+			<data special_zero="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="95">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="34" name="__module.attn/aten::transpose/Constant_1" type="Const" version="opset1">
+			<data element_type="i32" shape="3" offset="13784176" size="12" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="35" name="__module.attn/aten::transpose/Transpose_1" type="Transpose" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="96,q">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>128</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="36" name="Constant_9233" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 1" offset="13784188" size="4" />
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="37" name="__module.attn/aten::div/Divide_1" type="Divide" version="opset1">
+			<data auto_broadcast="numpy" m_pythondiv="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>128</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="118,q_scaled">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>128</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="38" name="self.kv_proj.weight" type="Const" version="opset1">
+			<data element_type="i8" shape="3584, 1152" offset="13784192" size="4128768" />
+			<output>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>1152</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="39" name="Convert_2008158" type="Convert" version="opset1">
+			<data destination_type="f16" />
+			<input>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>1152</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>1152</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="40" name="self.kv_proj.weight/scale" type="Const" version="opset1">
+			<data element_type="f16" shape="3584, 1" offset="17912960" size="7168" />
+			<output>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="41" name="self.kv_proj.weight/fq_weights_1" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>1152</dim>
+				</port>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP16">
+					<dim>3584</dim>
+					<dim>1152</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="42" name="self.kv_proj.weight/fq_weights_1/convert" type="Convert" version="opset1">
+			<data destination_type="f32" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>1152</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>1152</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="43" name="__module.kv_proj/ov_ext::linear/MatMul" type="MatMul" version="opset1">
+			<data transpose_a="false" transpose_b="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>1152</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="41,input.1">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="44" name="__module.ln_kv/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="17920128" size="4" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="45" name="__module.ln_kv/aten::layer_norm/MVN" type="MVN" version="opset6">
+			<data eps="9.9999999747524271e-07" normalize_variance="true" eps_mode="INSIDE_SQRT" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="46" name="Constant_9229" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 3584" offset="17920132" size="14336" />
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="47" name="__module.ln_kv/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="48" name="Constant_9230" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 3584" offset="17934468" size="14336" />
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="49" name="__module.ln_kv/aten::layer_norm/Add" type="Add" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="48">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="50" name="Constant_53" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="17948804" size="24" />
+			<output>
+				<port id="0" precision="I64" names="20">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="51" name="aten::permute/Transpose" type="Transpose" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="21,image_feature_1">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="52" name="aten::add/Add" type="Add" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="30,key">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="53" name="__module.attn/prim::ListUnpack/VariadicSplit.1" type="Const" version="opset1">
+			<data element_type="i8" shape="3584, 3584" offset="17948828" size="12845056" />
+			<output>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="54" name="Convert_1988050" type="Convert" version="opset1">
+			<data destination_type="f16" />
+			<input>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="55" name="__module.attn/prim::ListUnpack/VariadicSplit.1/scale" type="Const" version="opset1">
+			<data element_type="f16" shape="3584, 1" offset="30793884" size="7168" />
+			<output>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="56" name="__module.attn/prim::ListUnpack/VariadicSplit.1/fq_weights_1" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="57" name="__module.attn/prim::ListUnpack/VariadicSplit.1/fq_weights_1/convert" type="Convert" version="opset1">
+			<data destination_type="f32" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="58" name="__module.attn/aten::linear/MatMul" type="MatMul" version="opset1">
+			<data transpose_a="false" transpose_b="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="59" name="Constant_9231" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 3584" offset="30801052" size="14336" />
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="60" name="__module.attn/aten::linear/Add" type="Add" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="90,k.1">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="61" name="Constant_9688" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="30815388" size="24" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="62" name="__module.attn/aten::view/Reshape" type="Reshape" version="opset1">
+			<data special_zero="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="101">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="63" name="__module.attn/aten::transpose/Constant" type="Const" version="opset1">
+			<data element_type="i32" shape="3" offset="13784176" size="12" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="64" name="__module.attn/aten::transpose/Transpose" type="Transpose" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="102,k">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="65" name="__module.attn/aten::baddbmm/MatMul" type="MatMul" version="opset1">
+			<data transpose_a="false" transpose_b="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>128</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="66" name="__module.attn/aten::baddbmm/Add" type="Add" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>1</dim>
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="120,input.5">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="67" name="__module.attn/aten::softmax/Softmax" type="SoftMax" version="opset8">
+			<data axis="-1" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP32" names="121,attn_output_weights.1">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="68" name="__module.attn/prim::ListUnpack/VariadicSplit.2" type="Const" version="opset1">
+			<data element_type="i8" shape="3584, 3584" offset="30815412" size="12845056" />
+			<output>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="69" name="Convert_1983023" type="Convert" version="opset1">
+			<data destination_type="f16" />
+			<input>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="70" name="__module.attn/prim::ListUnpack/VariadicSplit.2/scale" type="Const" version="opset1">
+			<data element_type="f16" shape="3584, 1" offset="43660468" size="7168" />
+			<output>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="71" name="__module.attn/prim::ListUnpack/VariadicSplit.2/fq_weights_1" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="72" name="__module.attn/prim::ListUnpack/VariadicSplit.2/fq_weights_1/convert" type="Convert" version="opset1">
+			<data destination_type="f32" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="73" name="__module.attn/aten::linear/MatMul_2" type="MatMul" version="opset1">
+			<data transpose_a="false" transpose_b="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="74" name="Constant_9234" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 3584" offset="43667636" size="14336" />
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="75" name="__module.attn/aten::linear/Add_2" type="Add" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="91,v.1">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="76" name="Constant_9689" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="30815388" size="24" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="77" name="__module.attn/aten::view/Reshape_3" type="Reshape" version="opset1">
+			<data special_zero="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="107">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="78" name="__module.attn/aten::transpose/Constant_3" type="Const" version="opset1">
+			<data element_type="i32" shape="3" offset="13784176" size="12" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="79" name="__module.attn/aten::transpose/Transpose_3" type="Transpose" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="108,v">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="80" name="__module.attn/aten::bmm/MatMul" type="MatMul" version="opset1">
+			<data transpose_a="false" transpose_b="false" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="122,attn_output.1">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>128</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="81" name="__module.attn/aten::transpose/Constant_4" type="Const" version="opset1">
+			<data element_type="i32" shape="3" offset="13784176" size="12" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="82" name="__module.attn/aten::transpose/Transpose_4" type="Transpose" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>128</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="123">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="83" name="Concat_4375" type="Const" version="opset1">
+			<data element_type="i64" shape="2" offset="43681972" size="16" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="84" name="__module.attn/aten::view/Reshape_4" type="Reshape" version="opset1">
+			<data special_zero="false" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>128</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>2</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="128">
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="85" name="self.attn.out_proj.weight" type="Const" version="opset1">
+			<data element_type="i8" shape="3584, 3584" offset="43681988" size="12845056" />
+			<output>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="86" name="Convert_1998104" type="Convert" version="opset1">
+			<data destination_type="f16" />
+			<input>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="87" name="self.attn.out_proj.weight/scale" type="Const" version="opset1">
+			<data element_type="f16" shape="3584, 1" offset="56527044" size="7168" />
+			<output>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="88" name="self.attn.out_proj.weight/fq_weights_1" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="89" name="self.attn.out_proj.weight/fq_weights_1/convert" type="Convert" version="opset1">
+			<data destination_type="f32" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="90" name="__module.attn.out_proj/ov_ext::linear/MatMul" type="MatMul" version="opset1">
+			<data transpose_a="false" transpose_b="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="91" name="Constant_9235" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 3584" offset="56534212" size="14336" />
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="92" name="__module.attn.out_proj/ov_ext::linear/Add" type="Add" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="131,attn_output">
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="93" name="Constant_9690" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="56548548" size="24" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="94" name="__module.attn/aten::view/Reshape_5" type="Reshape" version="opset1">
+			<data special_zero="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="134,out">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="95" name="Constant_486" type="Const" version="opset1">
+			<data element_type="i64" shape="3" offset="17948804" size="24" />
+			<output>
+				<port id="0" precision="I64" names="35">
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="96" name="aten::permute/Transpose_1" type="Transpose" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>64</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="I64">
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="36,input">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="97" name="__module.ln_post/aten::layer_norm/Multiply" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="17920128" size="4" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="98" name="__module.ln_post/aten::layer_norm/MVN" type="MVN" version="opset6">
+			<data eps="9.9999999747524271e-07" normalize_variance="true" eps_mode="INSIDE_SQRT" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="99" name="Constant_9236" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 3584" offset="56548572" size="14336" />
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="100" name="__module.ln_post/aten::layer_norm/Multiply_1" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="101" name="Constant_9237" type="Const" version="opset1">
+			<data element_type="f32" shape="1, 1, 3584" offset="56562908" size="14336" />
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="102" name="__module.ln_post/aten::layer_norm/Add" type="Add" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>1</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="141,x">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="103" name="Transpose_8047" type="Const" version="opset1">
+			<data element_type="i8" shape="3584, 3584" offset="56577244" size="12845056" />
+			<output>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="104" name="Convert_2003131" type="Convert" version="opset1">
+			<data destination_type="f16" />
+			<input>
+				<port id="0" precision="I8">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="105" name="Transpose_8047/scale" type="Const" version="opset1">
+			<data element_type="f16" shape="3584, 1" offset="69422300" size="7168" />
+			<output>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="106" name="Transpose_8047/fq_weights_1" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP16">
+					<dim>3584</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="107" name="Transpose_8047/fq_weights_1/convert" type="Convert" version="opset1">
+			<data destination_type="f32" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="108" name="aten::matmul/MatMul" type="MatMul" version="opset1">
+			<data transpose_a="false" transpose_b="true" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>3584</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32" names="last_hidden_state">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="109" name="Result_545" type="Result" version="opset1" output_names="last_hidden_state">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>64</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+		</layer>
+	</layers>
+	<edges>
+		<edge from-layer="0" from-port="0" to-layer="7" to-port="0" />
+		<edge from-layer="0" from-port="0" to-layer="43" to-port="0" />
+		<edge from-layer="1" from-port="0" to-layer="52" to-port="1" />
+		<edge from-layer="2" from-port="0" to-layer="4" to-port="0" />
+		<edge from-layer="3" from-port="0" to-layer="4" to-port="1" />
+		<edge from-layer="4" from-port="2" to-layer="6" to-port="0" />
+		<edge from-layer="5" from-port="0" to-layer="6" to-port="1" />
+		<edge from-layer="6" from-port="2" to-layer="18" to-port="0" />
+		<edge from-layer="7" from-port="1" to-layer="10" to-port="0" />
+		<edge from-layer="7" from-port="1" to-layer="16" to-port="0" />
+		<edge from-layer="8" from-port="0" to-layer="10" to-port="1" />
+		<edge from-layer="9" from-port="0" to-layer="10" to-port="2" />
+		<edge from-layer="10" from-port="3" to-layer="12" to-port="0" />
+		<edge from-layer="10" from-port="3" to-layer="22" to-port="1" />
+		<edge from-layer="11" from-port="0" to-layer="12" to-port="1" />
+		<edge from-layer="12" from-port="2" to-layer="17" to-port="0" />
+		<edge from-layer="13" from-port="0" to-layer="17" to-port="1" />
+		<edge from-layer="14" from-port="0" to-layer="16" to-port="1" />
+		<edge from-layer="15" from-port="0" to-layer="16" to-port="2" />
+		<edge from-layer="16" from-port="3" to-layer="17" to-port="2" />
+		<edge from-layer="17" from-port="3" to-layer="18" to-port="1" />
+		<edge from-layer="18" from-port="2" to-layer="66" to-port="0" />
+		<edge from-layer="19" from-port="0" to-layer="23" to-port="0" />
+		<edge from-layer="20" from-port="0" to-layer="22" to-port="0" />
+		<edge from-layer="21" from-port="0" to-layer="22" to-port="2" />
+		<edge from-layer="22" from-port="3" to-layer="23" to-port="1" />
+		<edge from-layer="23" from-port="2" to-layer="29" to-port="0" />
+		<edge from-layer="24" from-port="0" to-layer="25" to-port="0" />
+		<edge from-layer="25" from-port="1" to-layer="27" to-port="0" />
+		<edge from-layer="26" from-port="0" to-layer="27" to-port="1" />
+		<edge from-layer="27" from-port="2" to-layer="28" to-port="0" />
+		<edge from-layer="28" from-port="1" to-layer="29" to-port="1" />
+		<edge from-layer="29" from-port="2" to-layer="31" to-port="0" />
+		<edge from-layer="30" from-port="0" to-layer="31" to-port="1" />
+		<edge from-layer="31" from-port="2" to-layer="33" to-port="0" />
+		<edge from-layer="32" from-port="0" to-layer="33" to-port="1" />
+		<edge from-layer="33" from-port="2" to-layer="35" to-port="0" />
+		<edge from-layer="34" from-port="0" to-layer="35" to-port="1" />
+		<edge from-layer="35" from-port="2" to-layer="37" to-port="0" />
+		<edge from-layer="36" from-port="0" to-layer="37" to-port="1" />
+		<edge from-layer="37" from-port="2" to-layer="65" to-port="0" />
+		<edge from-layer="38" from-port="0" to-layer="39" to-port="0" />
+		<edge from-layer="39" from-port="1" to-layer="41" to-port="0" />
+		<edge from-layer="40" from-port="0" to-layer="41" to-port="1" />
+		<edge from-layer="41" from-port="2" to-layer="42" to-port="0" />
+		<edge from-layer="42" from-port="1" to-layer="43" to-port="1" />
+		<edge from-layer="43" from-port="2" to-layer="45" to-port="0" />
+		<edge from-layer="44" from-port="0" to-layer="45" to-port="1" />
+		<edge from-layer="45" from-port="2" to-layer="47" to-port="0" />
+		<edge from-layer="46" from-port="0" to-layer="47" to-port="1" />
+		<edge from-layer="47" from-port="2" to-layer="49" to-port="0" />
+		<edge from-layer="48" from-port="0" to-layer="49" to-port="1" />
+		<edge from-layer="49" from-port="2" to-layer="51" to-port="0" />
+		<edge from-layer="50" from-port="0" to-layer="51" to-port="1" />
+		<edge from-layer="51" from-port="2" to-layer="52" to-port="0" />
+		<edge from-layer="51" from-port="2" to-layer="73" to-port="0" />
+		<edge from-layer="52" from-port="2" to-layer="58" to-port="0" />
+		<edge from-layer="53" from-port="0" to-layer="54" to-port="0" />
+		<edge from-layer="54" from-port="1" to-layer="56" to-port="0" />
+		<edge from-layer="55" from-port="0" to-layer="56" to-port="1" />
+		<edge from-layer="56" from-port="2" to-layer="57" to-port="0" />
+		<edge from-layer="57" from-port="1" to-layer="58" to-port="1" />
+		<edge from-layer="58" from-port="2" to-layer="60" to-port="0" />
+		<edge from-layer="59" from-port="0" to-layer="60" to-port="1" />
+		<edge from-layer="60" from-port="2" to-layer="62" to-port="0" />
+		<edge from-layer="61" from-port="0" to-layer="62" to-port="1" />
+		<edge from-layer="62" from-port="2" to-layer="64" to-port="0" />
+		<edge from-layer="63" from-port="0" to-layer="64" to-port="1" />
+		<edge from-layer="64" from-port="2" to-layer="65" to-port="1" />
+		<edge from-layer="65" from-port="2" to-layer="66" to-port="1" />
+		<edge from-layer="66" from-port="2" to-layer="67" to-port="0" />
+		<edge from-layer="67" from-port="1" to-layer="80" to-port="0" />
+		<edge from-layer="68" from-port="0" to-layer="69" to-port="0" />
+		<edge from-layer="69" from-port="1" to-layer="71" to-port="0" />
+		<edge from-layer="70" from-port="0" to-layer="71" to-port="1" />
+		<edge from-layer="71" from-port="2" to-layer="72" to-port="0" />
+		<edge from-layer="72" from-port="1" to-layer="73" to-port="1" />
+		<edge from-layer="73" from-port="2" to-layer="75" to-port="0" />
+		<edge from-layer="74" from-port="0" to-layer="75" to-port="1" />
+		<edge from-layer="75" from-port="2" to-layer="77" to-port="0" />
+		<edge from-layer="76" from-port="0" to-layer="77" to-port="1" />
+		<edge from-layer="77" from-port="2" to-layer="79" to-port="0" />
+		<edge from-layer="78" from-port="0" to-layer="79" to-port="1" />
+		<edge from-layer="79" from-port="2" to-layer="80" to-port="1" />
+		<edge from-layer="80" from-port="2" to-layer="82" to-port="0" />
+		<edge from-layer="81" from-port="0" to-layer="82" to-port="1" />
+		<edge from-layer="82" from-port="2" to-layer="84" to-port="0" />
+		<edge from-layer="83" from-port="0" to-layer="84" to-port="1" />
+		<edge from-layer="84" from-port="2" to-layer="90" to-port="0" />
+		<edge from-layer="85" from-port="0" to-layer="86" to-port="0" />
+		<edge from-layer="86" from-port="1" to-layer="88" to-port="0" />
+		<edge from-layer="87" from-port="0" to-layer="88" to-port="1" />
+		<edge from-layer="88" from-port="2" to-layer="89" to-port="0" />
+		<edge from-layer="89" from-port="1" to-layer="90" to-port="1" />
+		<edge from-layer="90" from-port="2" to-layer="92" to-port="0" />
+		<edge from-layer="91" from-port="0" to-layer="92" to-port="1" />
+		<edge from-layer="92" from-port="2" to-layer="94" to-port="0" />
+		<edge from-layer="93" from-port="0" to-layer="94" to-port="1" />
+		<edge from-layer="94" from-port="2" to-layer="96" to-port="0" />
+		<edge from-layer="95" from-port="0" to-layer="96" to-port="1" />
+		<edge from-layer="96" from-port="2" to-layer="98" to-port="0" />
+		<edge from-layer="97" from-port="0" to-layer="98" to-port="1" />
+		<edge from-layer="98" from-port="2" to-layer="100" to-port="0" />
+		<edge from-layer="99" from-port="0" to-layer="100" to-port="1" />
+		<edge from-layer="100" from-port="2" to-layer="102" to-port="0" />
+		<edge from-layer="101" from-port="0" to-layer="102" to-port="1" />
+		<edge from-layer="102" from-port="2" to-layer="108" to-port="0" />
+		<edge from-layer="103" from-port="0" to-layer="104" to-port="0" />
+		<edge from-layer="104" from-port="1" to-layer="106" to-port="0" />
+		<edge from-layer="105" from-port="0" to-layer="106" to-port="1" />
+		<edge from-layer="106" from-port="2" to-layer="107" to-port="0" />
+		<edge from-layer="107" from-port="1" to-layer="108" to-port="1" />
+		<edge from-layer="108" from-port="2" to-layer="109" to-port="0" />
+	</edges>
+	<rt_info>
+		<Runtime_version value="2026.0.0-20965-c6d6a13a886-releases/2026/0" />
+		<conversion_parameters>
+			<framework value="pytorch" />
+			<is_python_object value="True" />
+		</conversion_parameters>
+		<nncf>
+			<friendly_names_were_updated value="True" />
+			<version value="3.0.0" />
+			<weight_compression>
+				<advanced_parameters value="{'statistics_path': None, 'lora_adapter_rank': 256, 'group_size_fallback_mode': 'error', 'min_adjusted_group_size': 32, 'awq_params': {'subset_size': 32, 'percent_to_apply': 0.002, 'alpha_min': 0.0, 'alpha_max': 1.0, 'steps': 100, 'prefer_data_aware_scaling': True}, 'scale_estimation_params': {'subset_size': 64, 'initial_steps': 5, 'scale_steps': 5, 'weight_penalty': -1.0}, 'gptq_params': {'damp_percent': 0.1, 'block_size': 128, 'subset_size': 128}, 'lora_correction_params': {'adapter_rank': 8, 'num_iterations': 3, 'apply_regularization': True, 'subset_size': 128, 'use_int8_adapters': True}, 'backend_params': {}, 'codebook': None, 'adaptive_codebook_params': {'value_type': 'f8e4m3', 'across_blocks': False, 'num_elements': 16}}" />
+				<all_layers value="False" />
+				<awq value="False" />
+				<backup_mode value="int8_asym" />
+				<compression_format value="dequantize" />
+				<gptq value="False" />
+				<group_size value="-1" />
+				<ignored_scope value="[]" />
+				<lora_correction value="False" />
+				<mode value="int8_sym" />
+				<ratio value="1.0" />
+				<scale_estimation value="False" />
+				<sensitivity_metric value="weight_quantization_error" />
+			</weight_compression>
+		</nncf>
+		<optimum>
+			<nncf_version value="3.0.0" />
+			<optimum_intel_version value="1.27.0" />
+			<optimum_version value="2.1.0" />
+			<pytorch_version value="2.11.0" />
+			<transformers_version value="4.53.3" />
+		</optimum>
+	</rt_info>
+</net>

openvino_text_embeddings_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:05203a06ed0e763ac81a62e3ec3d8b681df37b6212f73fa814fd3479d9685e58
+size 543874280

openvino_text_embeddings_model.xml

@@ -0,0 +1,179 @@
+<?xml version="1.0"?>
+<net name="Model9" version="11">
+	<layers>
+		<layer id="0" name="input" type="Parameter" version="opset1">
+			<data shape="?,?" element_type="i64" />
+			<output>
+				<port id="0" precision="I64" names="input">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="1" name="self.weight" type="Const" version="opset1">
+			<data element_type="i8" shape="151666, 3584" offset="0" size="543570944" />
+			<output>
+				<port id="0" precision="I8">
+					<dim>151666</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="2" name="Convert_1150972" type="Convert" version="opset1">
+			<data destination_type="f16" />
+			<input>
+				<port id="0" precision="I8">
+					<dim>151666</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP16">
+					<dim>151666</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="3" name="self.weight/scale" type="Const" version="opset1">
+			<data element_type="f16" shape="151666, 1" offset="543570944" size="303332" />
+			<output>
+				<port id="0" precision="FP16">
+					<dim>151666</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="4" name="self.weight/fq_weights_0" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>151666</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="FP16">
+					<dim>151666</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP16">
+					<dim>151666</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="5" name="ov_ext::embedding/Convert" type="Convert" version="opset1">
+			<data destination_type="f32" />
+			<rt_info>
+				<attribute name="decompression" version="0" />
+			</rt_info>
+			<input>
+				<port id="0" precision="FP16">
+					<dim>151666</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP32">
+					<dim>151666</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="6" name="ov_ext::embedding/Convert_1" type="Convert" version="opset1">
+			<data destination_type="i32" />
+			<input>
+				<port id="0" precision="I64">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="7" name="ov_ext::embedding/Constant" type="Const" version="opset1">
+			<data element_type="i32" shape="" offset="543874276" size="4" />
+			<output>
+				<port id="0" precision="I32" />
+			</output>
+		</layer>
+		<layer id="8" name="ov_ext::embedding/Gather" type="Gather" version="opset8">
+			<data batch_dims="0" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>151666</dim>
+					<dim>3584</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="I32" />
+			</input>
+			<output>
+				<port id="3" precision="FP32" names="inputs_embeds">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="9" name="Result_124398" type="Result" version="opset1" output_names="inputs_embeds">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>3584</dim>
+				</port>
+			</input>
+		</layer>
+	</layers>
+	<edges>
+		<edge from-layer="0" from-port="0" to-layer="6" to-port="0" />
+		<edge from-layer="1" from-port="0" to-layer="2" to-port="0" />
+		<edge from-layer="2" from-port="1" to-layer="4" to-port="0" />
+		<edge from-layer="3" from-port="0" to-layer="4" to-port="1" />
+		<edge from-layer="4" from-port="2" to-layer="5" to-port="0" />
+		<edge from-layer="5" from-port="1" to-layer="8" to-port="0" />
+		<edge from-layer="6" from-port="1" to-layer="8" to-port="1" />
+		<edge from-layer="7" from-port="0" to-layer="8" to-port="2" />
+		<edge from-layer="8" from-port="3" to-layer="9" to-port="0" />
+	</edges>
+	<rt_info>
+		<Runtime_version value="2026.0.0-20965-c6d6a13a886-releases/2026/0" />
+		<conversion_parameters>
+			<framework value="pytorch" />
+			<is_python_object value="True" />
+		</conversion_parameters>
+		<nncf>
+			<friendly_names_were_updated value="True" />
+			<version value="3.0.0" />
+			<weight_compression>
+				<advanced_parameters value="{'statistics_path': None, 'lora_adapter_rank': 256, 'group_size_fallback_mode': 'error', 'min_adjusted_group_size': 32, 'awq_params': {'subset_size': 32, 'percent_to_apply': 0.002, 'alpha_min': 0.0, 'alpha_max': 1.0, 'steps': 100, 'prefer_data_aware_scaling': True}, 'scale_estimation_params': {'subset_size': 64, 'initial_steps': 5, 'scale_steps': 5, 'weight_penalty': -1.0}, 'gptq_params': {'damp_percent': 0.1, 'block_size': 128, 'subset_size': 128}, 'lora_correction_params': {'adapter_rank': 8, 'num_iterations': 3, 'apply_regularization': True, 'subset_size': 128, 'use_int8_adapters': True}, 'backend_params': {}, 'codebook': None, 'adaptive_codebook_params': {'value_type': 'f8e4m3', 'across_blocks': False, 'num_elements': 16}}" />
+				<all_layers value="False" />
+				<awq value="False" />
+				<backup_mode value="int8_asym" />
+				<compression_format value="dequantize" />
+				<gptq value="False" />
+				<group_size value="-1" />
+				<ignored_scope value="[]" />
+				<lora_correction value="False" />
+				<mode value="int8_sym" />
+				<ratio value="1.0" />
+				<scale_estimation value="False" />
+				<sensitivity_metric value="weight_quantization_error" />
+			</weight_compression>
+		</nncf>
+		<optimum>
+			<nncf_version value="3.0.0" />
+			<optimum_intel_version value="1.27.0" />
+			<optimum_version value="2.1.0" />
+			<pytorch_version value="2.11.0" />
+			<transformers_version value="4.53.3" />
+		</optimum>
+	</rt_info>
+</net>

openvino_tokenizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5abec952803879f370a1216869d4752cba2c2f3b8e50edfb0924b09eef190679
+size 5588665

openvino_tokenizer.xml

@@ -0,0 +1,771 @@
+<?xml version="1.0"?>
+<net name="tokenizer" version="11">
+	<layers>
+		<layer id="0" name="Parameter_126838" type="Parameter" version="opset1">
+			<data shape="?" element_type="string" />
+			<output>
+				<port id="0" precision="STRING" names="Parameter_126838">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="1" name="Constant_126844" type="Const" version="opset1">
+			<data element_type="i64" shape="" offset="0" size="8" />
+			<output>
+				<port id="0" precision="I64" />
+			</output>
+		</layer>
+		<layer id="2" name="StringTensorUnpack_126839" type="StringTensorUnpack" version="opset15">
+			<input>
+				<port id="0" precision="STRING">
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="3" precision="U8">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="3" name="ShapeOf_126840" type="ShapeOf" version="opset3">
+			<data output_type="i64" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="I64">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="4" name="Constant_126841" type="Const" version="opset1">
+			<data element_type="i64" shape="" offset="0" size="8" />
+			<output>
+				<port id="0" precision="I64" />
+			</output>
+		</layer>
+		<layer id="5" name="Constant_126842" type="Const" version="opset1">
+			<data element_type="i64" shape="" offset="0" size="8" />
+			<output>
+				<port id="0" precision="I64" />
+			</output>
+		</layer>
+		<layer id="6" name="Gather_126843" type="Gather" version="opset8">
+			<data batch_dims="0" />
+			<input>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+				<port id="1" precision="I64" />
+				<port id="2" precision="I64" />
+			</input>
+			<output>
+				<port id="3" precision="I64" />
+			</output>
+		</layer>
+		<layer id="7" name="Constant_126845" type="Const" version="opset1">
+			<data element_type="i64" shape="" offset="8" size="8" />
+			<output>
+				<port id="0" precision="I64" />
+			</output>
+		</layer>
+		<layer id="8" name="Range_126846" type="Range" version="opset4">
+			<data output_type="i32" />
+			<input>
+				<port id="0" precision="I64" />
+				<port id="1" precision="I64" />
+				<port id="2" precision="I64" />
+			</input>
+			<output>
+				<port id="3" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="9" name="Constant_126847" type="Const" version="opset1">
+			<data element_type="i64" shape="" offset="8" size="8" />
+			<output>
+				<port id="0" precision="I64" />
+			</output>
+		</layer>
+		<layer id="10" name="Constant_126848" type="Const" version="opset1">
+			<data element_type="i64" shape="" offset="8" size="8" />
+			<output>
+				<port id="0" precision="I64" />
+			</output>
+		</layer>
+		<layer id="11" name="Add_126849" type="Add" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="I64" />
+				<port id="1" precision="I64" />
+			</input>
+			<output>
+				<port id="2" precision="I64" />
+			</output>
+		</layer>
+		<layer id="12" name="Constant_126850" type="Const" version="opset1">
+			<data element_type="i64" shape="" offset="8" size="8" />
+			<output>
+				<port id="0" precision="I64" />
+			</output>
+		</layer>
+		<layer id="13" name="Range_126851" type="Range" version="opset4">
+			<data output_type="i32" />
+			<input>
+				<port id="0" precision="I64" />
+				<port id="1" precision="I64" />
+				<port id="2" precision="I64" />
+			</input>
+			<output>
+				<port id="3" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="14" name="Constant_126915" type="Const" version="opset1">
+			<data element_type="u8" shape="407" offset="16" size="407" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>407</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="15" name="SpecialTokensSplit_126916" type="SpecialTokensSplit" version="extension">
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="3" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="4" precision="U8">
+					<dim>-1</dim>
+				</port>
+				<port id="5" precision="U8">
+					<dim>407</dim>
+				</port>
+			</input>
+			<output>
+				<port id="6" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="7" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="8" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="9" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="10" precision="U8">
+					<dim>-1</dim>
+				</port>
+				<port id="11" precision="BOOL">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="16" name="CharsMapNormalization_126917" type="CharsMapNormalization" version="extension">
+			<data add_dummy_prefix="false" remove_extra_whitespaces="false" escape_whitespaces="false" normalization_form="nfc" case_fold="false" nmt="false" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="U8">
+					<dim>-1</dim>
+				</port>
+				<port id="3" precision="BOOL">
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="4" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="5" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="6" precision="U8">
+					<dim>-1</dim>
+				</port>
+				<port id="7" precision="BOOL">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="17" name="Constant_126919" type="Const" version="opset1">
+			<data element_type="u8" shape="110" offset="423" size="110" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>110</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="18" name="RegexSplit_126920" type="RegexSplit" version="extension">
+			<data behaviour="isolate" invert="false" max_splits="-1" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="3" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="4" precision="U8">
+					<dim>-1</dim>
+				</port>
+				<port id="5" precision="BOOL">
+					<dim>-1</dim>
+				</port>
+				<port id="6" precision="U8">
+					<dim>110</dim>
+				</port>
+			</input>
+			<output>
+				<port id="7" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="8" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="9" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="10" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="11" precision="U8">
+					<dim>-1</dim>
+				</port>
+				<port id="12" precision="BOOL">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="19" name="Constant_126922" type="Const" version="opset1">
+			<data element_type="i32" shape="151666" offset="533" size="606664" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151666</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="20" name="Constant_126924" type="Const" version="opset1">
+			<data element_type="i32" shape="151666" offset="607197" size="606664" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151666</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="21" name="Constant_126926" type="Const" version="opset1">
+			<data element_type="u8" shape="976273" offset="1213861" size="976273" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>976273</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="22" name="Constant_126934" type="Const" version="opset1">
+			<data element_type="i32" shape="151387" offset="2190134" size="605548" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151387</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="23" name="Constant_126936" type="Const" version="opset1">
+			<data element_type="i32" shape="151387" offset="2795682" size="605548" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151387</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="24" name="Constant_126938" type="Const" version="opset1">
+			<data element_type="u8" shape="491359" offset="3401230" size="491359" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>491359</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="25" name="Constant_126940" type="Const" version="opset1">
+			<data element_type="i32" shape="151387" offset="3892589" size="605548" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151387</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="26" name="Constant_126942" type="Const" version="opset1">
+			<data element_type="i32" shape="151387" offset="4498137" size="605548" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151387</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="27" name="Constant_126944" type="Const" version="opset1">
+			<data element_type="u8" shape="484354" offset="5103685" size="484354" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>484354</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="28" name="Constant_126928" type="Const" version="opset1">
+			<data element_type="i32" shape="24" offset="5588039" size="96" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>24</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="29" name="Constant_126930" type="Const" version="opset1">
+			<data element_type="i32" shape="24" offset="5588135" size="96" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>24</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="30" name="Constant_126932" type="Const" version="opset1">
+			<data element_type="u8" shape="309" offset="5588231" size="309" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>309</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="31" name="Constant_126945" type="Const" version="opset1">
+			<data element_type="i32" shape="24" offset="5588540" size="96" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>24</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="32" name="BPETokenizer_126946" type="BPETokenizer" version="extension">
+			<data unk_token="" fuse_unk="false" suffix_indicator="" end_suffix="" byte_fallback="false" cache_capacity="30328" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="3" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="4" precision="U8">
+					<dim>-1</dim>
+				</port>
+				<port id="5" precision="I32">
+					<dim>151666</dim>
+				</port>
+				<port id="6" precision="I32">
+					<dim>151666</dim>
+				</port>
+				<port id="7" precision="U8">
+					<dim>976273</dim>
+				</port>
+				<port id="8" precision="I32">
+					<dim>151387</dim>
+				</port>
+				<port id="9" precision="I32">
+					<dim>151387</dim>
+				</port>
+				<port id="10" precision="U8">
+					<dim>491359</dim>
+				</port>
+				<port id="11" precision="I32">
+					<dim>151387</dim>
+				</port>
+				<port id="12" precision="I32">
+					<dim>151387</dim>
+				</port>
+				<port id="13" precision="U8">
+					<dim>484354</dim>
+				</port>
+				<port id="14" precision="I32">
+					<dim>24</dim>
+				</port>
+				<port id="15" precision="I32">
+					<dim>24</dim>
+				</port>
+				<port id="16" precision="U8">
+					<dim>309</dim>
+				</port>
+				<port id="17" precision="I32">
+					<dim>24</dim>
+				</port>
+			</input>
+			<output>
+				<port id="18" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="19" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="20" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="33" name="Constant_126947" type="Const" version="opset1">
+			<data element_type="i32" shape="" offset="5588636" size="4" />
+			<output>
+				<port id="0" precision="I32" />
+			</output>
+		</layer>
+		<layer id="34" name="Constant_126949" type="Const" version="opset1">
+			<data element_type="u8" shape="4" offset="5588640" size="4" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="35" name="Constant_126951" type="Const" version="opset1">
+			<data element_type="u8" shape="13" offset="5588644" size="13" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>13</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="36" name="Truncate_126952" type="Truncate" version="extension">
+			<data m_num_inputs="1" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="3" precision="I32" />
+				<port id="4" precision="U8">
+					<dim>4</dim>
+				</port>
+				<port id="5" precision="U8">
+					<dim>13</dim>
+				</port>
+			</input>
+			<output>
+				<port id="6" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="7" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="8" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="37" name="Constant_126953" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="5588657" size="4" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="38" name="CombineSegments_126954" type="CombineSegments" version="extension">
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="3" precision="I32">
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="4" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="5" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="6" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="7" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="8" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="9" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="39" name="Subtract_126955" type="Subtract" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="40" name="Constant_126956" type="Const" version="opset1">
+			<data element_type="i32" shape="" offset="5588657" size="4" />
+			<output>
+				<port id="0" precision="I32" />
+			</output>
+		</layer>
+		<layer id="41" name="ReduceMax_126957" type="ReduceMax" version="opset1">
+			<data keep_dims="false" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32" />
+			</input>
+			<output>
+				<port id="2" precision="I32" />
+			</output>
+		</layer>
+		<layer id="42" name="Constant_126958" type="Const" version="opset1">
+			<data element_type="i32" shape="" offset="5588661" size="4" />
+			<output>
+				<port id="0" precision="I32" />
+			</output>
+		</layer>
+		<layer id="43" name="RaggedToDense_126959" type="RaggedToDense" version="extension">
+			<data pad_right="false" m_pad_max_length="false" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="2" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="3" precision="I32" />
+				<port id="4" precision="I32" />
+			</input>
+			<output>
+				<port id="5" precision="I32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+				<port id="6" precision="BOOL">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="44" name="Convert_126960" type="Convert" version="opset1">
+			<data destination_type="i32" />
+			<input>
+				<port id="0" precision="BOOL">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="I32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="45" name="Convert_126960.0" type="Convert" version="opset1">
+			<data destination_type="i64" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="I64" names="attention_mask">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="46" name="RaggedToDense_126959.0" type="Convert" version="opset1">
+			<data destination_type="i64" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="I64" names="input_ids">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="47" name="Result_126963" type="Result" version="opset1" output_names="input_ids">
+			<input>
+				<port id="0" precision="I64">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+		</layer>
+		<layer id="48" name="Result_126965" type="Result" version="opset1" output_names="attention_mask">
+			<input>
+				<port id="0" precision="I64">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+		</layer>
+	</layers>
+	<edges>
+		<edge from-layer="0" from-port="0" to-layer="2" to-port="0" />
+		<edge from-layer="1" from-port="0" to-layer="8" to-port="0" />
+		<edge from-layer="2" from-port="1" to-layer="3" to-port="0" />
+		<edge from-layer="2" from-port="1" to-layer="15" to-port="2" />
+		<edge from-layer="2" from-port="2" to-layer="15" to-port="3" />
+		<edge from-layer="2" from-port="3" to-layer="15" to-port="4" />
+		<edge from-layer="3" from-port="1" to-layer="6" to-port="0" />
+		<edge from-layer="4" from-port="0" to-layer="6" to-port="1" />
+		<edge from-layer="5" from-port="0" to-layer="6" to-port="2" />
+		<edge from-layer="6" from-port="3" to-layer="8" to-port="1" />
+		<edge from-layer="6" from-port="3" to-layer="11" to-port="0" />
+		<edge from-layer="7" from-port="0" to-layer="8" to-port="2" />
+		<edge from-layer="8" from-port="3" to-layer="15" to-port="0" />
+		<edge from-layer="9" from-port="0" to-layer="13" to-port="0" />
+		<edge from-layer="10" from-port="0" to-layer="11" to-port="1" />
+		<edge from-layer="11" from-port="2" to-layer="13" to-port="1" />
+		<edge from-layer="12" from-port="0" to-layer="13" to-port="2" />
+		<edge from-layer="13" from-port="3" to-layer="15" to-port="1" />
+		<edge from-layer="14" from-port="0" to-layer="15" to-port="5" />
+		<edge from-layer="15" from-port="8" to-layer="16" to-port="0" />
+		<edge from-layer="15" from-port="9" to-layer="16" to-port="1" />
+		<edge from-layer="15" from-port="10" to-layer="16" to-port="2" />
+		<edge from-layer="15" from-port="11" to-layer="16" to-port="3" />
+		<edge from-layer="15" from-port="6" to-layer="18" to-port="0" />
+		<edge from-layer="15" from-port="7" to-layer="18" to-port="1" />
+		<edge from-layer="16" from-port="4" to-layer="18" to-port="2" />
+		<edge from-layer="16" from-port="5" to-layer="18" to-port="3" />
+		<edge from-layer="16" from-port="6" to-layer="18" to-port="4" />
+		<edge from-layer="16" from-port="7" to-layer="18" to-port="5" />
+		<edge from-layer="17" from-port="0" to-layer="18" to-port="6" />
+		<edge from-layer="18" from-port="7" to-layer="32" to-port="0" />
+		<edge from-layer="18" from-port="8" to-layer="32" to-port="1" />
+		<edge from-layer="18" from-port="9" to-layer="32" to-port="2" />
+		<edge from-layer="18" from-port="10" to-layer="32" to-port="3" />
+		<edge from-layer="18" from-port="11" to-layer="32" to-port="4" />
+		<edge from-layer="19" from-port="0" to-layer="32" to-port="5" />
+		<edge from-layer="20" from-port="0" to-layer="32" to-port="6" />
+		<edge from-layer="21" from-port="0" to-layer="32" to-port="7" />
+		<edge from-layer="22" from-port="0" to-layer="32" to-port="8" />
+		<edge from-layer="23" from-port="0" to-layer="32" to-port="9" />
+		<edge from-layer="24" from-port="0" to-layer="32" to-port="10" />
+		<edge from-layer="25" from-port="0" to-layer="32" to-port="11" />
+		<edge from-layer="26" from-port="0" to-layer="32" to-port="12" />
+		<edge from-layer="27" from-port="0" to-layer="32" to-port="13" />
+		<edge from-layer="28" from-port="0" to-layer="32" to-port="14" />
+		<edge from-layer="29" from-port="0" to-layer="32" to-port="15" />
+		<edge from-layer="30" from-port="0" to-layer="32" to-port="16" />
+		<edge from-layer="31" from-port="0" to-layer="32" to-port="17" />
+		<edge from-layer="32" from-port="18" to-layer="36" to-port="0" />
+		<edge from-layer="32" from-port="19" to-layer="36" to-port="1" />
+		<edge from-layer="32" from-port="20" to-layer="36" to-port="2" />
+		<edge from-layer="33" from-port="0" to-layer="36" to-port="3" />
+		<edge from-layer="34" from-port="0" to-layer="36" to-port="4" />
+		<edge from-layer="35" from-port="0" to-layer="36" to-port="5" />
+		<edge from-layer="36" from-port="6" to-layer="38" to-port="0" />
+		<edge from-layer="36" from-port="7" to-layer="38" to-port="1" />
+		<edge from-layer="36" from-port="8" to-layer="38" to-port="2" />
+		<edge from-layer="37" from-port="0" to-layer="38" to-port="3" />
+		<edge from-layer="38" from-port="5" to-layer="39" to-port="0" />
+		<edge from-layer="38" from-port="4" to-layer="39" to-port="1" />
+		<edge from-layer="38" from-port="4" to-layer="43" to-port="0" />
+		<edge from-layer="38" from-port="5" to-layer="43" to-port="1" />
+		<edge from-layer="38" from-port="6" to-layer="43" to-port="2" />
+		<edge from-layer="39" from-port="2" to-layer="41" to-port="0" />
+		<edge from-layer="40" from-port="0" to-layer="41" to-port="1" />
+		<edge from-layer="41" from-port="2" to-layer="43" to-port="3" />
+		<edge from-layer="42" from-port="0" to-layer="43" to-port="4" />
+		<edge from-layer="43" from-port="6" to-layer="44" to-port="0" />
+		<edge from-layer="43" from-port="5" to-layer="46" to-port="0" />
+		<edge from-layer="44" from-port="1" to-layer="45" to-port="0" />
+		<edge from-layer="45" from-port="1" to-layer="48" to-port="0" />
+		<edge from-layer="46" from-port="1" to-layer="47" to-port="0" />
+	</edges>
+	<rt_info>
+		<add_attention_mask value="True" />
+		<add_prefix_space />
+		<add_special_tokens value="True" />
+		<bos_token_id value="151644" />
+		<chat_template value="{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '&lt;|im_start|>system&#10;You are a helpful assistant.&lt;|im_end|>&#10;' }}{% endif %}{{'&lt;|im_start|>' + message['role'] + '&#10;' + message['content'] + '&lt;|im_end|>' + '&#10;'}}{% endfor %}{% if add_generation_prompt %}{{ '&lt;|im_start|>assistant&#10;' }}{% endif %}" />
+		<clean_up_tokenization_spaces />
+		<detokenizer_input_type value="i64" />
+		<eos_token_id value="151645" />
+		<handle_special_tokens_with_re />
+		<max_length />
+		<number_of_inputs value="1" />
+		<openvino_tokenizers_version value="2026.0.0.0-632-47cea02a2d4" />
+		<openvino_version value="2026.0.0-20965-c6d6a13a886-releases/2026/0" />
+		<original_post_processor_template value="{&quot;type&quot;: &quot;ByteLevel&quot;, &quot;add_prefix_space&quot;: false, &quot;trim_offsets&quot;: false, &quot;use_regex&quot;: false}" />
+		<original_tokenizer_class value="&lt;class 'transformers_modules.openbmb.MiniCPM-V-2_6.6c04d9e3022bcff6e6738dfb1fc19a5cfd2a855f.tokenization_minicpmv_fast.MiniCPMVTokenizerFast'>" />
+		<pad_token_id value="151643" />
+		<processed_post_processor_template value="{&quot;single&quot;: {&quot;ids&quot;: [-1], &quot;type_ids&quot;: [0]}, &quot;pair&quot;: {&quot;ids&quot;: [-1, -2], &quot;type_ids&quot;: [0, 0]}}" />
+		<skip_special_tokens value="True" />
+		<streaming_detokenizer value="False" />
+		<tokenizer_output_type value="i64" />
+		<tokenizers_version value="0.21.4" />
+		<transformers_version value="4.53.3" />
+		<use_max_padding value="False" />
+		<use_sentencepiece_backend value="False" />
+		<utf8_replace_mode value="replace" />
+		<with_detokenizer value="True" />
+	</rt_info>
+</net>

openvino_vision_embeddings_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f41cb38674e7630d4a0914a72b9f3ad605d8c40c34ce736b78bb4d70c4477f7
+size 419003300

preprocessor_config.json

@@ -0,0 +1,47 @@
+{
+  "auto_map": {
+    "AutoImageProcessor": "image_processing_minicpmv.MiniCPMVImageProcessor",
+    "AutoProcessor": "processing_minicpmv.MiniCPMVProcessor"
+  },
+  "im_end": "</image>",
+  "im_end_token": "</image>",
+  "im_id_end": "</image_id>",
+  "im_id_start": "<image_id>",
+  "im_start": "<image>",
+  "im_start_token": "<image>",
+  "image_feature_size": 64,
+  "image_processor_type": "MiniCPMVImageProcessor",
+  "max_slice_nums": 9,
+  "mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "norm_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "norm_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "patch_size": 14,
+  "processor_class": "MiniCPMVProcessor",
+  "scale_resolution": 448,
+  "slice_end": "</slice>",
+  "slice_end_token": "</slice>",
+  "slice_mode": true,
+  "slice_start": "<slice>",
+  "slice_start_token": "<slice>",
+  "std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "unk": "<unk>",
+  "unk_token": "<unk>",
+  "use_image_id": true,
+  "version": 2.6
+}

processing_minicpmv.py

@@ -0,0 +1,240 @@
+# coding=utf-8
+# Copyright 2024 The HuggingFace Inc. team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""
+Processor class for MiniCPMV.
+"""
+
+from typing import List, Optional, Union, Dict, Any
+import torch
+import re
+
+from transformers.image_processing_utils import BatchFeature
+from transformers.image_utils import ImageInput
+from transformers.processing_utils import ProcessorMixin
+from transformers.tokenization_utils_base import PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy
+from transformers.utils import TensorType, requires_backends, is_torch_dtype, is_torch_device
+
+from .image_processing_minicpmv import MiniCPMVBatchFeature
+
+
+class MiniCPMVProcessor(ProcessorMixin):
+    r"""
+    Constructs a MiniCPMV processor which wraps a MiniCPMV image processor and a MiniCPMV tokenizer into a single processor.
+
+    [`MiniCPMVProcessor`] offers all the functionalities of [`MiniCPMVImageProcessor`] and [`LlamaTokenizerWrapper`]. See the
+    [`~MiniCPMVProcessor.__call__`] and [`~MiniCPMVProcessor.decode`] for more information.
+
+    Args:
+        image_processor ([`MiniCPMVImageProcessor`], *optional*):
+            The image processor is a required input.
+        tokenizer ([`LlamaTokenizerWrapper`], *optional*):
+            The tokenizer is a required input.
+    """
+    attributes = ["image_processor", "tokenizer"]
+    image_processor_class = "AutoImageProcessor"
+    tokenizer_class = "AutoTokenizer"
+
+    def __init__(self, image_processor=None, tokenizer=None):
+        super().__init__(image_processor, tokenizer)
+        self.version = image_processor.version
+    
+    def __call__(
+        self,
+        text: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]],
+        images: ImageInput = None,
+        max_length: Optional[int] = None,
+        do_pad: Optional[bool] = True,
+        max_slice_nums: int = None,
+        use_image_id: bool = None,
+        return_tensors: Optional[Union[str, TensorType]] = TensorType.PYTORCH,
+        **kwargs
+    ) -> MiniCPMVBatchFeature:
+
+        if images is not None:
+            image_inputs = self.image_processor(images, do_pad=do_pad, max_slice_nums=max_slice_nums, return_tensors=return_tensors)
+        return self._convert_images_texts_to_inputs(image_inputs, text, max_slice_nums=max_slice_nums, use_image_id=use_image_id, max_length=max_length, **kwargs)
+    
+    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.batch_decode with CLIP->Llama
+    def batch_decode(self, *args, **kwargs):
+        """
+        This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please
+        refer to the docstring of this method for more information.
+        """
+        output_ids = args[0]
+        result_text = []
+        for result in output_ids:
+            result = result[result != 0]
+            if result[0] == self.tokenizer.bos_id:
+                result = result[1:]
+            if result[-1] == self.tokenizer.eos_id:
+                result = result[:-1]
+            result_text.append(self.tokenizer.decode(result, *args[1:], **kwargs).strip())
+        return result_text
+        # return self.tokenizer.batch_decode(*args, **kwargs)
+    
+    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.decode with CLIP->Llama
+    def decode(self, *args, **kwargs):
+        """
+        This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to
+        the docstring of this method for more information.
+        """
+        result = args[0]
+        result = result[result != 0]
+        if result[0] == self.tokenizer.bos_id:
+            result = result[1:]
+        if result[-1] == self.tokenizer.eos_id or (hasattr(self.tokenizer, "eot_id") and result[-1] == self.tokenizer.eot_id):
+            result = result[:-1]
+        return self.tokenizer.decode(result, *args[1:], **kwargs).strip()
+
+    def _convert(
+        self, input_str, max_inp_length: Optional[int] = None
+    ):
+        if self.version > 2.5 or not getattr(self.tokenizer, "add_bos_token", False):
+            input_ids = self.tokenizer.encode(input_str)
+        else:
+            input_ids = [self.tokenizer.bos_id] + self.tokenizer.encode(input_str)
+        if max_inp_length is not None:
+            input_ids = input_ids[:max_inp_length]
+        input_ids = torch.tensor(input_ids, dtype=torch.int32)
+
+        start_cond = (input_ids == self.tokenizer.im_start_id) | (input_ids == self.tokenizer.slice_start_id)
+        end_cond = (input_ids == self.tokenizer.im_end_id) | (input_ids == self.tokenizer.slice_end_id)
+
+        image_start_tokens = torch.where(start_cond)[0]
+        image_start_tokens += 1
+        image_end_tokens = torch.where(end_cond)[0]
+
+        valid_image_nums = max(len(image_start_tokens), len(image_end_tokens))
+
+        image_bounds = torch.hstack(
+            [
+                image_start_tokens[:valid_image_nums].unsqueeze(-1),
+                image_end_tokens[:valid_image_nums].unsqueeze(-1),
+            ]
+        )
+        return input_ids, image_bounds
+
+    def _convert_images_texts_to_inputs(
+            self, 
+            images, 
+            texts: Union[str, List[str]], 
+            truncation=None, 
+            max_length=None,
+            max_slice_nums=None,
+            use_image_id=None, 
+            return_tensors=None,
+            **kwargs
+        ):
+        if images is None or not len(images):
+            model_inputs = self.tokenizer(texts, return_tensors=return_tensors, truncation=truncation, max_length=max_length, **kwargs)
+            return MiniCPMVBatchFeature(data={**model_inputs})
+        
+        pattern = "(<image>./</image>)"
+        images, image_sizes, tgt_sizes = images["pixel_values"], images["image_sizes"], images["tgt_sizes"]
+        
+        if isinstance(texts, str):
+            texts = [texts]
+        input_ids_list = []
+        image_bounds_list = []
+        for index, text in enumerate(texts):
+            image_tags = re.findall(pattern, text)
+            assert len(image_tags) == len(image_sizes[index])
+            text_chunks = text.split(pattern)
+            final_text = ""
+            for i in range(len(image_tags)):
+                final_text = final_text + text_chunks[i] + \
+                    self.image_processor.get_slice_image_placeholder(
+                        image_sizes[index][i], 
+                        i,
+                        max_slice_nums,
+                        use_image_id
+                    )
+            final_text += text_chunks[-1]
+            input_ids, image_bounds = self._convert(final_text, max_length)
+            input_ids_list.append(input_ids)
+            image_bounds_list.append(image_bounds)
+        padded_input_ids, padding_lengths = self.pad(
+            input_ids_list,
+            padding_side="left"
+        )
+        for i, length in enumerate(padding_lengths):
+            image_bounds_list[i] = image_bounds_list[i] + length
+        attention_mask = padded_input_ids.ne(0)
+
+        return MiniCPMVBatchFeature(data={
+            "input_ids": padded_input_ids,
+            "attention_mask": attention_mask,
+            "pixel_values": images,
+            "image_sizes": image_sizes,
+            "image_bound": image_bounds_list,
+            "tgt_sizes": tgt_sizes
+        })
+
+    @property
+    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.model_input_names
+    def model_input_names(self):
+        tokenizer_input_names = self.tokenizer.model_input_names
+        image_processor_input_names = self.image_processor.model_input_names
+        return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names))
+
+
+    def pad(self, inputs, max_length=None, padding_value=0, padding_side="left"):
+        items = []
+        if isinstance(inputs[0], list):
+            assert isinstance(inputs[0][0], torch.Tensor)
+            for it in inputs:
+                for tr in it:
+                    items.append(tr)
+        else:
+            assert isinstance(inputs[0], torch.Tensor)
+            items = inputs
+
+        batch_size = len(items)
+        shape = items[0].shape
+        dim = len(shape)
+        assert dim <= 2
+        if max_length is None:
+            max_length = 0
+        max_length = max(max_length, max(item.shape[-1] for item in items))
+        min_length = min(item.shape[-1] for item in items)
+        dtype = items[0].dtype
+
+        if dim == 0:
+            return torch.stack([item for item in items], dim=0), [0]
+        elif dim == 1:
+            if max_length == min_length:
+                return torch.stack([item for item in items], dim=0), [0] * batch_size
+            tensor = torch.zeros((batch_size, max_length), dtype=dtype) + padding_value
+        else:
+            tensor = (
+                torch.zeros((batch_size, max_length, shape[-1]), dtype=dtype)
+                + padding_value
+            )
+
+        padding_length = []
+        for i, item in enumerate(items):
+            if dim == 1:
+                if padding_side == "left":
+                    tensor[i, -len(item) :] = item.clone()
+                else:
+                    tensor[i, : len(item)] = item.clone()
+            elif dim == 2:
+                if padding_side == "left":
+                    tensor[i, -len(item) :, :] = item.clone()
+                else:
+                    tensor[i, : len(item), :] = item.clone()
+            padding_length.append(tensor.shape[-1] - len(item))
+
+        return tensor, padding_length

processor_config.json

@@ -0,0 +1,6 @@
+{
+  "auto_map": {
+    "AutoProcessor": "processing_minicpmv.MiniCPMVProcessor"
+  },
+  "processor_class": "MiniCPMVProcessor"
+}

special_tokens_map.json

@@ -0,0 +1,52 @@
+{
+  "additional_special_tokens": [
+    "<image>",
+    "</image>",
+    "<ref>",
+    "</ref>",
+    "<box>",
+    "</box>",
+    "<quad>",
+    "</quad>",
+    "<point>",
+    "</point>",
+    "<slice>",
+    "</slice>",
+    "<image_id>",
+    "</image_id>",
+    "<|reserved_special_token_0|>",
+    "<|reserved_special_token_1|>",
+    "<|reserved_special_token_2|>",
+    "<|reserved_special_token_3|>",
+    "<|reserved_special_token_4|>",
+    "<|reserved_special_token_5|>"
+  ],
+  "bos_token": {
+    "content": "<|im_start|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "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
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenization_minicpmv_fast.py

@@ -0,0 +1,66 @@
+from transformers.models.qwen2 import Qwen2TokenizerFast
+
+
+class MiniCPMVTokenizerFast(Qwen2TokenizerFast):
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        self.im_start = "<image>"
+        self.im_end = "</image>"
+        self.ref_start = "<ref>"
+        self.ref_end = "</ref>"
+        self.box_start = "<box>"
+        self.box_end = "</box>"
+        self.quad_start = "<quad>"
+        self.quad_end = "</quad>"
+        self.slice_start = "<slice>"
+        self.slice_end = "</slice>"
+        self.im_id_start = "<image_id>"
+        self.im_id_end = "</image_id>"
+
+    @property
+    def eos_id(self):
+        return self.eos_token_id
+
+    @property
+    def bos_id(self):
+        return self.bos_token_id
+
+    @property
+    def unk_id(self):
+        return self.unk_token_id
+
+    @property
+    def im_start_id(self):
+        return self.convert_tokens_to_ids(self.im_start)
+
+    @property
+    def im_end_id(self):
+        return self.convert_tokens_to_ids(self.im_end)
+
+    @property
+    def slice_start_id(self):
+        return self.convert_tokens_to_ids(self.slice_start)
+    
+    @property
+    def slice_end_id(self):
+        return self.convert_tokens_to_ids(self.slice_end)
+
+    @property
+    def im_id_start_id(self):
+        return self.convert_tokens_to_ids(self.im_id_start)
+    
+    @property
+    def im_id_end_id(self):
+        return self.convert_tokens_to_ids(self.im_id_end)
+    
+    @property
+    def newline_id(self):
+        return self.convert_tokens_to_ids('\n')
+
+    @staticmethod
+    def escape(text: str) -> str:
+        return text
+
+    @staticmethod
+    def unescape(text: str) -> str:
+        return text

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9de76ce95f90e336b4d2b0ec11d37f3d5404f2dad0f7ac95298405474b2a3a90
+size 11422257

tokenizer_config.json

@@ -0,0 +1,237 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "128244": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "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": "<image>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "</image>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<ref>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "</ref>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<box>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "</box>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<quad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "</quad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<point>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "</point>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<slice>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "</slice>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151658": {
+      "content": "<image_id>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151659": {
+      "content": "</image_id>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151660": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151661": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151662": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151663": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151664": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151665": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<image>",
+    "</image>",
+    "<ref>",
+    "</ref>",
+    "<box>",
+    "</box>",
+    "<quad>",
+    "</quad>",
+    "<point>",
+    "</point>",
+    "<slice>",
+    "</slice>",
+    "<image_id>",
+    "</image_id>",
+    "<|reserved_special_token_0|>",
+    "<|reserved_special_token_1|>",
+    "<|reserved_special_token_2|>",
+    "<|reserved_special_token_3|>",
+    "<|reserved_special_token_4|>",
+    "<|reserved_special_token_5|>"
+  ],
+  "auto_map": {
+    "AutoProcessor": "processing_minicpmv.MiniCPMVProcessor",
+    "AutoTokenizer": [
+      "tokenization_qwen2.Qwen2Tokenizer",
+      "tokenization_minicpmv_fast.MiniCPMVTokenizerFast"
+    ]
+  },
+  "bos_token": "<|im_start|>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<|endoftext|>",
+  "processor_class": "MiniCPMVProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "MiniCPMVTokenizer",
+  "unk_token": "<unk>"
+}