Upload 26 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

README.md

@@ -0,0 +1,147 @@
+---
+license: mit
+language:
+- multilingual
+pipeline_tag: image-text-to-text
+tags:
+- nlp
+- vision
+- internvl
+base_model:
+- OpenGVLab/InternVL2-1B
+base_model_relation: quantized
+---
+
+# InternVL2-1B-int8-ov
+
+ * Model creator: [OpenGVLab](https://huggingface.co/OpenGVLab)
+ * Original model: [InternVL2-1B](https://huggingface.co/OpenGVLab/InternVL2-1B)
+
+## Description
+
+This is [OpenGVLab/InternVL2-1B](https://huggingface.co/OpenGVLab/InternVL2-1B) model converted to the [OpenVINO™ IR](https://docs.openvino.ai/2024/documentation/openvino-ir-format.html) (Intermediate Representation) format with weights compressed to INT8 by [NNCF](https://github.com/openvinotoolkit/nncf).
+
+
+## Quantization Parameters
+
+Weight compression was performed using `nncf.compress_weights` with the following parameters:
+* mode: **INT8_ASYM**
+
+## Compatibility
+
+The provided OpenVINO™ IR model is compatible with:
+
+* OpenVINO version 2025.0.0 and higher
+* Optimum Intel 1.21.0 and higher
+
+## Running Model Inference with [Optimum Intel](https://huggingface.co/docs/optimum/intel/index)
+
+1. Install packages required for using [Optimum Intel](https://huggingface.co/docs/optimum/intel/index) integration with the OpenVINO backend:
+
+```
+pip install --pre -U --extra-index-url https://storage.openvinotoolkit.org/simple/wheels/pre-release openvino_tokenizers openvino
+
+pip install git+https://github.com/huggingface/optimum-intel.git
+```
+
+2. Run model inference
+
+```
+from PIL import Image 
+import requests 
+from optimum.intel.openvino import OVModelForVisualCausalLM
+from transformers import AutoTokenizer, TextStreamer
+
+model_id = "OpenVINO/InternVL2-1B-int8-ov"
+
+tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
+
+ov_model = OVModelForVisualCausalLM.from_pretrained(model_id, trust_remote_code=True)
+prompt = "What is unusual on this picture?"
+
+url = "https://github.com/openvinotoolkit/openvino_notebooks/assets/29454499/d5fbbd1a-d484-415c-88cb-9986625b7b11"
+image = Image.open(requests.get(url, stream=True).raw)
+
+inputs = ov_model.preprocess_inputs(text=prompt, image=image, tokenizer=tokenizer, config=ov_model.config)
+
+generation_args = { 
+    "max_new_tokens": 100, 
+    "streamer": TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
+} 
+
+generate_ids = ov_model.generate(**inputs, **generation_args)
+
+generate_ids = generate_ids[:, inputs['input_ids'].shape[1]:]
+response = tokenizer.batch_decode(generate_ids, skip_special_tokens=True)[0]
+
+```
+
+## Running Model Inference with [OpenVINO GenAI](https://github.com/openvinotoolkit/openvino.genai)
+
+1. Install packages required for using OpenVINO GenAI.
+```
+pip install --pre -U --extra-index-url https://storage.openvinotoolkit.org/simple/wheels/pre-release openvino openvino-tokenizers openvino-genai
+
+pip install huggingface_hub
+```
+
+2. Download model from HuggingFace Hub
+   
+```
+import huggingface_hub as hf_hub
+
+model_id = "OpenVINO/InternVL2-1B-int8-ov"
+model_path = "InternVL2-1B-int8-ov"
+
+hf_hub.snapshot_download(model_id, local_dir=model_path)
+
+```
+
+1. Run model inference:
+
+```
+import openvino_genai as ov_genai
+import requests
+from PIL import Image
+from io import BytesIO
+import numpy as np
+import openvino as ov
+
+device = "CPU"
+pipe = ov_genai.VLMPipeline(model_path, device)
+
+def load_image(image_file):
+    if isinstance(image_file, str) and (image_file.startswith("http") or image_file.startswith("https")):
+        response = requests.get(image_file)
+        image = Image.open(BytesIO(response.content)).convert("RGB")
+    else:
+        image = Image.open(image_file).convert("RGB")
+    image_data = np.array(image.getdata()).reshape(1, image.size[1], image.size[0], 3).astype(np.byte)
+    return ov.Tensor(image_data)
+
+prompt = "What is unusual on this picture?"
+
+url = "https://github.com/openvinotoolkit/openvino_notebooks/assets/29454499/d5fbbd1a-d484-415c-88cb-9986625b7b11"
+image_tensor = load_image(url)
+
+def streamer(subword: str) -> bool:
+    print(subword, end="", flush=True)
+    return False
+
+pipe.start_chat()
+output = pipe.generate(prompt, image=image_tensor, max_new_tokens=100, streamer=streamer)
+pipe.finish_chat()
+```
+
+More GenAI usage examples can be found in OpenVINO GenAI library [docs](https://github.com/openvinotoolkit/openvino.genai/blob/master/src/README.md) and [samples](https://github.com/openvinotoolkit/openvino.genai?tab=readme-ov-file#openvino-genai-samples)
+
+
+## Limitations
+
+
+Check the original [model card](https://huggingface.co/OpenGVLab/InternVL2-1B) for limitations.
+
+## Legal information
+
+The original model is distributed under [MIT](https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/mit.md) license. More details can be found in [original model card](https://huggingface.co/OpenGVLab/InternVL2-1B).
+

added_tokens.json

@@ -0,0 +1,14 @@
+{
+  "</box>": 151654,
+  "</img>": 151647,
+  "</quad>": 151650,
+  "</ref>": 151652,
+  "<IMG_CONTEXT>": 151648,
+  "<box>": 151653,
+  "<img>": 151646,
+  "<quad>": 151649,
+  "<ref>": 151651,
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644
+}

config.json

@@ -0,0 +1,196 @@
+{
+  "_attn_implementation_autoset": true,
+  "_commit_hash": "3bc2e85347e17591b9387f283fef91b985b770c3",
+  "_name_or_path": "OpenGVLab/InternVL2-1B",
+  "architectures": [
+    "InternVLChatModel"
+  ],
+  "auto_map": {
+    "AutoConfig": "configuration_internvl_chat.InternVLChatConfig",
+    "AutoModel": "modeling_internvl_chat.InternVLChatModel",
+    "AutoModelForCausalLM": "modeling_internvl_chat.InternVLChatModel"
+  },
+  "downsample_ratio": 0.5,
+  "dynamic_image_size": true,
+  "force_image_size": 448,
+  "img_context_token_id": 151648,
+  "llm_config": {
+    "_attn_implementation_autoset": true,
+    "_name_or_path": "Qwen/Qwen2-0.5B-Instruct",
+    "add_cross_attention": false,
+    "architectures": [
+      "Qwen2ForCausalLM"
+    ],
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": 151643,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": 151645,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "silu",
+    "hidden_size": 896,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "intermediate_size": 4864,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "max_position_embeddings": 32768,
+    "max_window_layers": 24,
+    "min_length": 0,
+    "model_type": "qwen2",
+    "no_repeat_ngram_size": 0,
+    "num_attention_heads": 14,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_hidden_layers": 24,
+    "num_key_value_heads": 2,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "prefix": null,
+    "problem_type": null,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "rms_norm_eps": 1e-06,
+    "rope_scaling": null,
+    "rope_theta": 1000000.0,
+    "sep_token_id": null,
+    "sliding_window": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": false,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "bfloat16",
+    "torchscript": false,
+    "transformers_version": "4.47.0",
+    "typical_p": 1.0,
+    "use_bfloat16": true,
+    "use_cache": true,
+    "use_sliding_window": false,
+    "vocab_size": 151655
+  },
+  "max_dynamic_patch": 12,
+  "min_dynamic_patch": 1,
+  "model_type": "internvl_chat",
+  "ps_version": "v2",
+  "select_layer": -1,
+  "template": "Hermes-2",
+  "transformers_version": null,
+  "use_backbone_lora": 0,
+  "use_llm_lora": 0,
+  "use_thumbnail": true,
+  "vision_config": {
+    "_attn_implementation_autoset": true,
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": [
+      "InternVisionModel"
+    ],
+    "attention_dropout": 0.0,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "drop_path_rate": 0.0,
+    "dropout": 0.0,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_act": "gelu",
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "image_size": 448,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-06,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "intern_vit_6b",
+    "no_repeat_ngram_size": 0,
+    "norm_type": "layer_norm",
+    "num_attention_heads": 16,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_channels": 3,
+    "num_hidden_layers": 24,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "patch_size": 14,
+    "prefix": null,
+    "problem_type": null,
+    "pruned_heads": {},
+    "qk_normalization": false,
+    "qkv_bias": true,
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "bfloat16",
+    "torchscript": false,
+    "transformers_version": "4.47.0",
+    "typical_p": 1.0,
+    "use_bfloat16": true,
+    "use_flash_attn": false
+  }
+}

configuration_intern_vit.py

@@ -0,0 +1,120 @@
+# --------------------------------------------------------
+# InternVL
+# Copyright (c) 2024 OpenGVLab
+# Licensed under The MIT License [see LICENSE for details]
+# --------------------------------------------------------
+
+import os
+from typing import Union
+
+from transformers.configuration_utils import PretrainedConfig
+from transformers.utils import logging
+
+logger = logging.get_logger(__name__)
+
+
+class InternVisionConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`InternVisionModel`]. It is used to
+    instantiate a vision encoder according to the specified arguments, defining the model architecture.
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+    Args:
+        num_channels (`int`, *optional*, defaults to 3):
+            Number of color channels in the input images (e.g., 3 for RGB).
+        patch_size (`int`, *optional*, defaults to 14):
+            The size (resolution) of each patch.
+        image_size (`int`, *optional*, defaults to 224):
+            The size (resolution) of each image.
+        qkv_bias (`bool`, *optional*, defaults to `False`):
+            Whether to add a bias to the queries and values in the self-attention layers.
+        hidden_size (`int`, *optional*, defaults to 3200):
+            Dimensionality of the encoder layers and the pooler layer.
+        num_attention_heads (`int`, *optional*, defaults to 25):
+            Number of attention heads for each attention layer in the Transformer encoder.
+        intermediate_size (`int`, *optional*, defaults to 12800):
+            Dimensionality of the "intermediate" (i.e., feed-forward) layer in the Transformer encoder.
+        qk_normalization (`bool`, *optional*, defaults to `True`):
+            Whether to normalize the queries and keys in the self-attention layers.
+        num_hidden_layers (`int`, *optional*, defaults to 48):
+            Number of hidden layers in the Transformer encoder.
+        use_flash_attn (`bool`, *optional*, defaults to `True`):
+            Whether to use flash attention mechanism.
+        hidden_act (`str` or `function`, *optional*, defaults to `"gelu"`):
+            The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`,
+            `"relu"`, `"selu"` and `"gelu_new"` ``"gelu"` are supported.
+        layer_norm_eps (`float`, *optional*, defaults to 1e-6):
+            The epsilon used by the layer normalization layers.
+        dropout (`float`, *optional*, defaults to 0.0):
+            The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.
+        drop_path_rate (`float`, *optional*, defaults to 0.0):
+            Dropout rate for stochastic depth.
+        attention_dropout (`float`, *optional*, defaults to 0.0):
+            The dropout ratio for the attention probabilities.
+        initializer_range (`float`, *optional*, defaults to 0.02):
+            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
+        initializer_factor (`float`, *optional*, defaults to 0.1):
+            A factor for layer scale.
+    """
+
+    model_type = 'intern_vit_6b'
+
+    def __init__(
+            self,
+            num_channels=3,
+            patch_size=14,
+            image_size=224,
+            qkv_bias=False,
+            hidden_size=3200,
+            num_attention_heads=25,
+            intermediate_size=12800,
+            qk_normalization=True,
+            num_hidden_layers=48,
+            use_flash_attn=True,
+            hidden_act='gelu',
+            norm_type='rms_norm',
+            layer_norm_eps=1e-6,
+            dropout=0.0,
+            drop_path_rate=0.0,
+            attention_dropout=0.0,
+            initializer_range=0.02,
+            initializer_factor=0.1,
+            **kwargs,
+    ):
+        super().__init__(**kwargs)
+
+        self.hidden_size = hidden_size
+        self.intermediate_size = intermediate_size
+        self.dropout = dropout
+        self.drop_path_rate = drop_path_rate
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.num_channels = num_channels
+        self.patch_size = patch_size
+        self.image_size = image_size
+        self.initializer_range = initializer_range
+        self.initializer_factor = initializer_factor
+        self.attention_dropout = attention_dropout
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+        self.norm_type = norm_type
+        self.qkv_bias = qkv_bias
+        self.qk_normalization = qk_normalization
+        self.use_flash_attn = use_flash_attn
+
+    @classmethod
+    def from_pretrained(cls, pretrained_model_name_or_path: Union[str, os.PathLike], **kwargs) -> 'PretrainedConfig':
+        config_dict, kwargs = cls.get_config_dict(pretrained_model_name_or_path, **kwargs)
+
+        if 'vision_config' in config_dict:
+            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)

configuration_internvl_chat.py

@@ -0,0 +1,95 @@
+# --------------------------------------------------------
+# InternVL
+# Copyright (c) 2024 OpenGVLab
+# Licensed under The MIT License [see LICENSE for details]
+# --------------------------------------------------------
+
+import copy
+
+from transformers import AutoConfig, LlamaConfig, Qwen2Config
+from transformers.configuration_utils import PretrainedConfig
+from transformers.utils import logging
+
+from .configuration_intern_vit import InternVisionConfig
+
+logger = logging.get_logger(__name__)
+
+
+class InternVLChatConfig(PretrainedConfig):
+    model_type = 'internvl_chat'
+    is_composition = True
+
+    def __init__(
+            self,
+            vision_config=None,
+            llm_config=None,
+            use_backbone_lora=0,
+            use_llm_lora=0,
+            select_layer=-1,
+            force_image_size=None,
+            downsample_ratio=0.5,
+            template=None,
+            dynamic_image_size=False,
+            use_thumbnail=False,
+            ps_version='v1',
+            min_dynamic_patch=1,
+            max_dynamic_patch=6,
+            **kwargs):
+        super().__init__(**kwargs)
+
+        if vision_config is None:
+            vision_config = {'architectures': ['InternVisionModel']}
+            logger.info('vision_config is None. Initializing the InternVisionConfig with default values.')
+
+        if llm_config is None:
+            llm_config = {'architectures': ['Qwen2ForCausalLM']}
+            logger.info('llm_config is None. Initializing the LlamaConfig config with default values (`LlamaConfig`).')
+
+        self.vision_config = InternVisionConfig(**vision_config)
+        if llm_config.get('architectures')[0] == 'LlamaForCausalLM':
+            self.llm_config = LlamaConfig(**llm_config)
+        elif llm_config.get('architectures')[0] == 'Qwen2ForCausalLM':
+            self.llm_config = Qwen2Config(**llm_config)
+        else:
+            raise ValueError('Unsupported architecture: {}'.format(llm_config.get('architectures')[0]))
+        self.use_backbone_lora = use_backbone_lora
+        self.use_llm_lora = use_llm_lora
+        self.select_layer = select_layer
+        self.force_image_size = force_image_size
+        self.downsample_ratio = downsample_ratio
+        self.template = template
+        self.dynamic_image_size = dynamic_image_size
+        self.use_thumbnail = use_thumbnail
+        self.ps_version = ps_version  # pixel shuffle version
+        self.min_dynamic_patch = min_dynamic_patch
+        self.max_dynamic_patch = max_dynamic_patch
+
+        logger.info(f'vision_select_layer: {self.select_layer}')
+        logger.info(f'ps_version: {self.ps_version}')
+        logger.info(f'min_dynamic_patch: {self.min_dynamic_patch}')
+        logger.info(f'max_dynamic_patch: {self.max_dynamic_patch}')
+
+    def to_dict(self):
+        """
+        Serializes this instance to a Python dictionary. Override the default [`~PretrainedConfig.to_dict`].
+
+        Returns:
+            `Dict[str, any]`: Dictionary of all the attributes that make up this configuration instance,
+        """
+        output = copy.deepcopy(self.__dict__)
+        output['vision_config'] = self.vision_config.to_dict()
+        output['llm_config'] = self.llm_config.to_dict()
+        output['model_type'] = self.__class__.model_type
+        output['use_backbone_lora'] = self.use_backbone_lora
+        output['use_llm_lora'] = self.use_llm_lora
+        output['select_layer'] = self.select_layer
+        output['force_image_size'] = self.force_image_size
+        output['downsample_ratio'] = self.downsample_ratio
+        output['template'] = self.template
+        output['dynamic_image_size'] = self.dynamic_image_size
+        output['use_thumbnail'] = self.use_thumbnail
+        output['ps_version'] = self.ps_version
+        output['min_dynamic_patch'] = self.min_dynamic_patch
+        output['max_dynamic_patch'] = self.max_dynamic_patch
+
+        return output

conversation.py

@@ -0,0 +1,391 @@
+"""
+Conversation prompt templates.
+
+We kindly request that you import fastchat instead of copying this file if you wish to use it.
+If you have changes in mind, please contribute back so the community can benefit collectively and continue to maintain these valuable templates.
+
+Modified from https://github.com/lm-sys/FastChat/blob/main/fastchat/conversation.py
+"""
+
+import dataclasses
+from enum import IntEnum, auto
+from typing import Dict, List, Tuple, Union
+
+
+class SeparatorStyle(IntEnum):
+    """Separator styles."""
+
+    ADD_COLON_SINGLE = auto()
+    ADD_COLON_TWO = auto()
+    ADD_COLON_SPACE_SINGLE = auto()
+    NO_COLON_SINGLE = auto()
+    NO_COLON_TWO = auto()
+    ADD_NEW_LINE_SINGLE = auto()
+    LLAMA2 = auto()
+    CHATGLM = auto()
+    CHATML = auto()
+    CHATINTERN = auto()
+    DOLLY = auto()
+    RWKV = auto()
+    PHOENIX = auto()
+    ROBIN = auto()
+    FALCON_CHAT = auto()
+    CHATGLM3 = auto()
+    INTERNVL_ZH = auto()
+    MPT = auto()
+
+
+@dataclasses.dataclass
+class Conversation:
+    """A class that manages prompt templates and keeps all conversation history."""
+
+    # The name of this template
+    name: str
+    # The template of the system prompt
+    system_template: str = '{system_message}'
+    # The system message
+    system_message: str = ''
+    # The names of two roles
+    roles: Tuple[str] = ('USER', 'ASSISTANT')
+    # All messages. Each item is (role, message).
+    messages: List[List[str]] = ()
+    # The number of few shot examples
+    offset: int = 0
+    # The separator style and configurations
+    sep_style: SeparatorStyle = SeparatorStyle.ADD_COLON_SINGLE
+    sep: str = '\n'
+    sep2: str = None
+    # Stop criteria (the default one is EOS token)
+    stop_str: Union[str, List[str]] = None
+    # Stops generation if meeting any token in this list
+    stop_token_ids: List[int] = None
+
+    def get_prompt(self) -> str:
+        """Get the prompt for generation."""
+        system_prompt = self.system_template.format(system_message=self.system_message)
+        if self.sep_style == SeparatorStyle.ADD_COLON_SINGLE:
+            ret = system_prompt + self.sep
+            for role, message in self.messages:
+                if message:
+                    ret += role + ': ' + message + self.sep
+                else:
+                    ret += role + ':'
+            return ret
+        elif self.sep_style == SeparatorStyle.ADD_COLON_TWO:
+            seps = [self.sep, self.sep2]
+            ret = system_prompt + seps[0]
+            for i, (role, message) in enumerate(self.messages):
+                if message:
+                    ret += role + ': ' + message + seps[i % 2]
+                else:
+                    ret += role + ':'
+            return ret
+        elif self.sep_style == SeparatorStyle.ADD_COLON_SPACE_SINGLE:
+            ret = system_prompt + self.sep
+            for role, message in self.messages:
+                if message:
+                    ret += role + ': ' + message + self.sep
+                else:
+                    ret += role + ': '  # must be end with a space
+            return ret
+        elif self.sep_style == SeparatorStyle.ADD_NEW_LINE_SINGLE:
+            ret = '' if system_prompt == '' else system_prompt + self.sep
+            for role, message in self.messages:
+                if message:
+                    ret += role + '\n' + message + self.sep
+                else:
+                    ret += role + '\n'
+            return ret
+        elif self.sep_style == SeparatorStyle.NO_COLON_SINGLE:
+            ret = system_prompt
+            for role, message in self.messages:
+                if message:
+                    ret += role + message + self.sep
+                else:
+                    ret += role
+            return ret
+        elif self.sep_style == SeparatorStyle.NO_COLON_TWO:
+            seps = [self.sep, self.sep2]
+            ret = system_prompt
+            for i, (role, message) in enumerate(self.messages):
+                if message:
+                    ret += role + message + seps[i % 2]
+                else:
+                    ret += role
+            return ret
+        elif self.sep_style == SeparatorStyle.RWKV:
+            ret = system_prompt
+            for i, (role, message) in enumerate(self.messages):
+                if message:
+                    ret += (
+                        role
+                        + ': '
+                        + message.replace('\r\n', '\n').replace('\n\n', '\n')
+                    )
+                    ret += '\n\n'
+                else:
+                    ret += role + ':'
+            return ret
+        elif self.sep_style == SeparatorStyle.LLAMA2:
+            seps = [self.sep, self.sep2]
+            if self.system_message:
+                ret = system_prompt
+            else:
+                ret = '[INST] '
+            for i, (role, message) in enumerate(self.messages):
+                tag = self.roles[i % 2]
+                if message:
+                    if i == 0:
+                        ret += message + ' '
+                    else:
+                        ret += tag + ' ' + message + seps[i % 2]
+                else:
+                    ret += tag
+            return ret
+        elif self.sep_style == SeparatorStyle.CHATGLM:
+            # source: https://huggingface.co/THUDM/chatglm-6b/blob/1d240ba371910e9282298d4592532d7f0f3e9f3e/modeling_chatglm.py#L1302-L1308
+            # source2: https://huggingface.co/THUDM/chatglm2-6b/blob/e186c891cf64310ac66ef10a87e6635fa6c2a579/modeling_chatglm.py#L926
+            round_add_n = 1 if self.name == 'chatglm2' else 0
+            if system_prompt:
+                ret = system_prompt + self.sep
+            else:
+                ret = ''
+
+            for i, (role, message) in enumerate(self.messages):
+                if i % 2 == 0:
+                    ret += f'[Round {i//2 + round_add_n}]{self.sep}'
+
+                if message:
+                    ret += f'{role}：{message}{self.sep}'
+                else:
+                    ret += f'{role}：'
+            return ret
+        elif self.sep_style == SeparatorStyle.CHATML:
+            ret = '' if system_prompt == '' else system_prompt + self.sep + '\n'
+            for role, message in self.messages:
+                if message:
+                    ret += role + '\n' + message + self.sep + '\n'
+                else:
+                    ret += role + '\n'
+            return ret
+        elif self.sep_style == SeparatorStyle.CHATGLM3:
+            ret = ''
+            if self.system_message:
+                ret += system_prompt
+            for role, message in self.messages:
+                if message:
+                    ret += role + '\n' + ' ' + message
+                else:
+                    ret += role
+            return ret
+        elif self.sep_style == SeparatorStyle.CHATINTERN:
+            # source: https://huggingface.co/internlm/internlm-chat-7b-8k/blob/bd546fa984b4b0b86958f56bf37f94aa75ab8831/modeling_internlm.py#L771
+            seps = [self.sep, self.sep2]
+            ret = system_prompt
+            for i, (role, message) in enumerate(self.messages):
+                # if i % 2 == 0:
+                #     ret += "<s>"
+                if message:
+                    ret += role + ':' + message + seps[i % 2] + '\n'
+                else:
+                    ret += role + ':'
+            return ret
+        elif self.sep_style == SeparatorStyle.DOLLY:
+            seps = [self.sep, self.sep2]
+            ret = system_prompt
+            for i, (role, message) in enumerate(self.messages):
+                if message:
+                    ret += role + ':\n' + message + seps[i % 2]
+                    if i % 2 == 1:
+                        ret += '\n\n'
+                else:
+                    ret += role + ':\n'
+            return ret
+        elif self.sep_style == SeparatorStyle.PHOENIX:
+            ret = system_prompt
+            for role, message in self.messages:
+                if message:
+                    ret += role + ': ' + '<s>' + message + '</s>'
+                else:
+                    ret += role + ': ' + '<s>'
+            return ret
+        elif self.sep_style == SeparatorStyle.ROBIN:
+            ret = system_prompt + self.sep
+            for role, message in self.messages:
+                if message:
+                    ret += role + ':\n' + message + self.sep
+                else:
+                    ret += role + ':\n'
+            return ret
+        elif self.sep_style == SeparatorStyle.FALCON_CHAT:
+            ret = ''
+            if self.system_message:
+                ret += system_prompt + self.sep
+            for role, message in self.messages:
+                if message:
+                    ret += role + ': ' + message + self.sep
+                else:
+                    ret += role + ':'
+
+            return ret
+        elif self.sep_style == SeparatorStyle.INTERNVL_ZH:
+            seps = [self.sep, self.sep2]
+            ret = self.system_message + seps[0]
+            for i, (role, message) in enumerate(self.messages):
+                if message:
+                    ret += role + ': ' + message + seps[i % 2]
+                else:
+                    ret += role + ':'
+            return ret
+        elif self.sep_style == SeparatorStyle.MPT:
+            ret = system_prompt + self.sep
+            for role, message in self.messages:
+                if message:
+                    if type(message) is tuple:
+                        message, _, _ = message
+                    ret += role + message + self.sep
+                else:
+                    ret += role
+            return ret
+        else:
+            raise ValueError(f'Invalid style: {self.sep_style}')
+
+    def set_system_message(self, system_message: str):
+        """Set the system message."""
+        self.system_message = system_message
+
+    def append_message(self, role: str, message: str):
+        """Append a new message."""
+        self.messages.append([role, message])
+
+    def update_last_message(self, message: str):
+        """Update the last output.
+
+        The last message is typically set to be None when constructing the prompt,
+        so we need to update it in-place after getting the response from a model.
+        """
+        self.messages[-1][1] = message
+
+    def to_gradio_chatbot(self):
+        """Convert the conversation to gradio chatbot format."""
+        ret = []
+        for i, (role, msg) in enumerate(self.messages[self.offset :]):
+            if i % 2 == 0:
+                ret.append([msg, None])
+            else:
+                ret[-1][-1] = msg
+        return ret
+
+    def to_openai_api_messages(self):
+        """Convert the conversation to OpenAI chat completion format."""
+        ret = [{'role': 'system', 'content': self.system_message}]
+
+        for i, (_, msg) in enumerate(self.messages[self.offset :]):
+            if i % 2 == 0:
+                ret.append({'role': 'user', 'content': msg})
+            else:
+                if msg is not None:
+                    ret.append({'role': 'assistant', 'content': msg})
+        return ret
+
+    def copy(self):
+        return Conversation(
+            name=self.name,
+            system_template=self.system_template,
+            system_message=self.system_message,
+            roles=self.roles,
+            messages=[[x, y] for x, y in self.messages],
+            offset=self.offset,
+            sep_style=self.sep_style,
+            sep=self.sep,
+            sep2=self.sep2,
+            stop_str=self.stop_str,
+            stop_token_ids=self.stop_token_ids,
+        )
+
+    def dict(self):
+        return {
+            'template_name': self.name,
+            'system_message': self.system_message,
+            'roles': self.roles,
+            'messages': self.messages,
+            'offset': self.offset,
+        }
+
+
+# A global registry for all conversation templates
+conv_templates: Dict[str, Conversation] = {}
+
+
+def register_conv_template(template: Conversation, override: bool = False):
+    """Register a new conversation template."""
+    if not override:
+        assert (
+            template.name not in conv_templates
+        ), f'{template.name} has been registered.'
+
+    conv_templates[template.name] = template
+
+
+def get_conv_template(name: str) -> Conversation:
+    """Get a conversation template."""
+    return conv_templates[name].copy()
+
+
+# Both Hermes-2 and internlm2-chat are chatml-format conversation templates. The difference
+# is that during training, the preprocessing function for the Hermes-2 template doesn't add
+# <s> at the beginning of the tokenized sequence, while the internlm2-chat template does.
+# Therefore, they are completely equivalent during inference.
+register_conv_template(
+    Conversation(
+        name='Hermes-2',
+        system_template='<|im_start|>system\n{system_message}',
+        # note: The new system prompt was not used here to avoid changes in benchmark performance.
+        # system_message='我是书生·万象，英文名是InternVL，是由上海人工智能实验室、清华大学及多家合作单位联合开发的多模态大语言模型。',
+        system_message='你是由上海人工智能实验室联合商汤科技开发的书生多模态大模型，英文名叫InternVL, 是一个有用无害的人工智能助手。',
+        roles=('<|im_start|>user\n', '<|im_start|>assistant\n'),
+        sep_style=SeparatorStyle.MPT,
+        sep='<|im_end|>',
+        stop_str='<|endoftext|>',
+    )
+)
+
+
+register_conv_template(
+    Conversation(
+        name='internlm2-chat',
+        system_template='<|im_start|>system\n{system_message}',
+        # note: The new system prompt was not used here to avoid changes in benchmark performance.
+        # system_message='我是书生·万象，英文名是InternVL，是由上海人工智能实验室、清华大学及多家合作单位联合开发的多模态大语言模型。',
+        system_message='你是由上海人工智能实验室联合商汤科技开发的书生多模态大模型，英文名叫InternVL, 是一个有用无害的人工智能助手。',
+        roles=('<|im_start|>user\n', '<|im_start|>assistant\n'),
+        sep_style=SeparatorStyle.MPT,
+        sep='<|im_end|>',
+    )
+)
+
+
+register_conv_template(
+    Conversation(
+        name='phi3-chat',
+        system_template='<|system|>\n{system_message}',
+        # note: The new system prompt was not used here to avoid changes in benchmark performance.
+        # system_message='我是书生·万象，英文名是InternVL，是由上海人工智能实验室、清华大学及多家合作单位联合开发的多模态大语言模型。',
+        system_message='你是由上海人工智能实验室联合商汤科技开发的书生多模态大模型，英文名叫InternVL, 是一个有用无害的人工智能助手。',
+        roles=('<|user|>\n', '<|assistant|>\n'),
+        sep_style=SeparatorStyle.MPT,
+        sep='<|end|>',
+    )
+)
+
+
+register_conv_template(
+    Conversation(
+        name='internvl2_5',
+        system_template='<|im_start|>system\n{system_message}',
+        system_message='你是书生·万象，英文名是InternVL，是由上海人工智能实验室、清华大学及多家合作单位联合开发的多模态大语言模型。',
+        roles=('<|im_start|>user\n', '<|im_start|>assistant\n'),
+        sep_style=SeparatorStyle.MPT,
+        sep='<|im_end|>\n',
+    )
+)

generation_config.json

@@ -0,0 +1,8 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": [
+    151644,
+    151645
+  ],
+  "transformers_version": "4.47.0"
+}

modeling_intern_vit.py

@@ -0,0 +1,430 @@
+# --------------------------------------------------------
+# InternVL
+# Copyright (c) 2024 OpenGVLab
+# Licensed under The MIT License [see LICENSE for details]
+# --------------------------------------------------------
+
+from typing import Optional, Tuple, Union
+
+import torch
+import torch.nn.functional as F
+import torch.utils.checkpoint
+from einops import rearrange
+from timm.models.layers import DropPath
+from torch import nn
+from transformers.activations import ACT2FN
+from transformers.modeling_outputs import (BaseModelOutput,
+                                           BaseModelOutputWithPooling)
+from transformers.modeling_utils import PreTrainedModel
+from transformers.utils import logging
+
+from .configuration_intern_vit import InternVisionConfig
+
+try:
+    from flash_attn.bert_padding import pad_input, unpad_input
+    from flash_attn.flash_attn_interface import \
+        flash_attn_varlen_qkvpacked_func
+    has_flash_attn = True
+except:
+    print('FlashAttention2 is not installed.')
+    has_flash_attn = False
+
+logger = logging.get_logger(__name__)
+
+
+class FlashAttention(nn.Module):
+    """Implement the scaled dot product attention with softmax.
+    Arguments
+    ---------
+        softmax_scale: The temperature to use for the softmax attention.
+                      (default: 1/sqrt(d_keys) where d_keys is computed at
+                      runtime)
+        attention_dropout: The dropout rate to apply to the attention
+                           (default: 0.0)
+    """
+
+    def __init__(self, softmax_scale=None, attention_dropout=0.0, device=None, dtype=None):
+        super().__init__()
+        self.softmax_scale = softmax_scale
+        self.dropout_p = attention_dropout
+
+    def forward(self, qkv, key_padding_mask=None, causal=False, cu_seqlens=None,
+                max_s=None, need_weights=False):
+        """Implements the multihead softmax attention.
+        Arguments
+        ---------
+            qkv: The tensor containing the query, key, and value. (B, S, 3, H, D) if key_padding_mask is None
+                if unpadded: (nnz, 3, h, d)
+            key_padding_mask: a bool tensor of shape (B, S)
+        """
+        assert not need_weights
+        assert qkv.dtype in [torch.float16, torch.bfloat16]
+        assert qkv.is_cuda
+
+        if cu_seqlens is None:
+            batch_size = qkv.shape[0]
+            seqlen = qkv.shape[1]
+            if key_padding_mask is None:
+                qkv = rearrange(qkv, 'b s ... -> (b s) ...')
+                max_s = seqlen
+                cu_seqlens = torch.arange(0, (batch_size + 1) * seqlen, step=seqlen, dtype=torch.int32,
+                                          device=qkv.device)
+                output = flash_attn_varlen_qkvpacked_func(
+                    qkv, cu_seqlens, max_s, self.dropout_p if self.training else 0.0,
+                    softmax_scale=self.softmax_scale, causal=causal
+                )
+                output = rearrange(output, '(b s) ... -> b s ...', b=batch_size)
+            else:
+                nheads = qkv.shape[-2]
+                x = rearrange(qkv, 'b s three h d -> b s (three h d)')
+                x_unpad, indices, cu_seqlens, max_s = unpad_input(x, key_padding_mask)
+                x_unpad = rearrange(x_unpad, 'nnz (three h d) -> nnz three h d', three=3, h=nheads)
+                output_unpad = flash_attn_varlen_qkvpacked_func(
+                    x_unpad, cu_seqlens, max_s, self.dropout_p if self.training else 0.0,
+                    softmax_scale=self.softmax_scale, causal=causal
+                )
+                output = rearrange(pad_input(rearrange(output_unpad, 'nnz h d -> nnz (h d)'),
+                                             indices, batch_size, seqlen),
+                                   'b s (h d) -> b s h d', h=nheads)
+        else:
+            assert max_s is not None
+            output = flash_attn_varlen_qkvpacked_func(
+                qkv, cu_seqlens, max_s, self.dropout_p if self.training else 0.0,
+                softmax_scale=self.softmax_scale, causal=causal
+            )
+
+        return output, None
+
+
+class InternRMSNorm(nn.Module):
+    def __init__(self, hidden_size, eps=1e-6):
+        super().__init__()
+        self.weight = nn.Parameter(torch.ones(hidden_size))
+        self.variance_epsilon = eps
+
+    def forward(self, hidden_states):
+        input_dtype = hidden_states.dtype
+        hidden_states = hidden_states.to(torch.float32)
+        variance = hidden_states.pow(2).mean(-1, keepdim=True)
+        hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
+        return self.weight * hidden_states.to(input_dtype)
+
+
+try:
+    from apex.normalization import FusedRMSNorm
+
+    InternRMSNorm = FusedRMSNorm  # noqa
+
+    logger.info('Discovered apex.normalization.FusedRMSNorm - will use it instead of InternRMSNorm')
+except ImportError:
+    # using the normal InternRMSNorm
+    pass
+except Exception:
+    logger.warning('discovered apex but it failed to load, falling back to InternRMSNorm')
+    pass
+
+
+NORM2FN = {
+    'rms_norm': InternRMSNorm,
+    'layer_norm': nn.LayerNorm,
+}
+
+
+class InternVisionEmbeddings(nn.Module):
+    def __init__(self, config: InternVisionConfig):
+        super().__init__()
+        self.config = config
+        self.embed_dim = config.hidden_size
+        self.image_size = config.image_size
+        self.patch_size = config.patch_size
+
+        self.class_embedding = nn.Parameter(
+            torch.randn(1, 1, self.embed_dim),
+        )
+
+        self.patch_embedding = nn.Conv2d(
+            in_channels=3, out_channels=self.embed_dim, kernel_size=self.patch_size, stride=self.patch_size
+        )
+
+        self.num_patches = (self.image_size // self.patch_size) ** 2
+        self.num_positions = self.num_patches + 1
+
+        self.position_embedding = nn.Parameter(torch.randn(1, self.num_positions, self.embed_dim))
+
+    def _get_pos_embed(self, pos_embed, H, W):
+        target_dtype = pos_embed.dtype
+        pos_embed = pos_embed.float().reshape(
+            1, self.image_size // self.patch_size, self.image_size // self.patch_size, -1).permute(0, 3, 1, 2)
+        pos_embed = F.interpolate(pos_embed, size=(H, W), mode='bicubic', align_corners=False). \
+            reshape(1, -1, H * W).permute(0, 2, 1).to(target_dtype)
+        return pos_embed
+
+    def forward(self, pixel_values: torch.FloatTensor) -> torch.Tensor:
+        target_dtype = self.patch_embedding.weight.dtype
+        patch_embeds = self.patch_embedding(pixel_values)  # shape = [*, channel, width, height]
+        batch_size, _, height, width = patch_embeds.shape
+        patch_embeds = patch_embeds.flatten(2).transpose(1, 2)
+        class_embeds = self.class_embedding.expand(batch_size, 1, -1).to(target_dtype)
+        embeddings = torch.cat([class_embeds, patch_embeds], dim=1)
+        position_embedding = torch.cat([
+            self.position_embedding[:, :1, :],
+            self._get_pos_embed(self.position_embedding[:, 1:, :], height, width)
+        ], dim=1)
+        embeddings = embeddings + position_embedding.to(target_dtype)
+        return embeddings
+
+
+class InternAttention(nn.Module):
+    """Multi-headed attention from 'Attention Is All You Need' paper"""
+
+    def __init__(self, config: InternVisionConfig):
+        super().__init__()
+        self.config = config
+        self.embed_dim = config.hidden_size
+        self.num_heads = config.num_attention_heads
+        self.use_flash_attn = config.use_flash_attn and has_flash_attn
+        if config.use_flash_attn and not has_flash_attn:
+            print('Warning: Flash Attention is not available, use_flash_attn is set to False.')
+        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.qkv = nn.Linear(self.embed_dim, 3 * self.embed_dim, bias=config.qkv_bias)
+        self.attn_drop = nn.Dropout(config.attention_dropout)
+        self.proj_drop = nn.Dropout(config.dropout)
+
+        self.qk_normalization = config.qk_normalization
+
+        if self.qk_normalization:
+            self.q_norm = InternRMSNorm(self.embed_dim, eps=config.layer_norm_eps)
+            self.k_norm = InternRMSNorm(self.embed_dim, eps=config.layer_norm_eps)
+
+        if self.use_flash_attn:
+            self.inner_attn = FlashAttention(attention_dropout=config.attention_dropout)
+        self.proj = nn.Linear(self.embed_dim, self.embed_dim)
+
+    def _naive_attn(self, x):
+        B, N, C = x.shape
+        qkv = self.qkv(x).reshape(B, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
+        q, k, v = qkv.unbind(0)  # make torchscript happy (cannot use tensor as tuple)
+
+        if self.qk_normalization:
+            B_, H_, N_, D_ = q.shape
+            q = self.q_norm(q.transpose(1, 2).flatten(-2, -1)).view(B_, N_, H_, D_).transpose(1, 2)
+            k = self.k_norm(k.transpose(1, 2).flatten(-2, -1)).view(B_, N_, H_, D_).transpose(1, 2)
+
+        attn = ((q * self.scale) @ k.transpose(-2, -1))
+        attn = attn.softmax(dim=-1)
+        attn = self.attn_drop(attn)
+
+        x = (attn @ v).transpose(1, 2).reshape(B, N, C)
+        x = self.proj(x)
+        x = self.proj_drop(x)
+        return x
+
+    def _flash_attn(self, x, key_padding_mask=None, need_weights=False):
+        qkv = self.qkv(x)
+        qkv = rearrange(qkv, 'b s (three h d) -> b s three h d', three=3, h=self.num_heads)
+
+        if self.qk_normalization:
+            q, k, v = qkv.unbind(2)
+            q = self.q_norm(q.flatten(-2, -1)).view(q.shape)
+            k = self.k_norm(k.flatten(-2, -1)).view(k.shape)
+            qkv = torch.stack([q, k, v], dim=2)
+
+        context, _ = self.inner_attn(
+            qkv, key_padding_mask=key_padding_mask, need_weights=need_weights, causal=False
+        )
+        outs = self.proj(rearrange(context, 'b s h d -> b s (h d)'))
+        outs = self.proj_drop(outs)
+        return outs
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        x = self._naive_attn(hidden_states) if not self.use_flash_attn else self._flash_attn(hidden_states)
+        return x
+
+
+class InternMLP(nn.Module):
+    def __init__(self, config: InternVisionConfig):
+        super().__init__()
+        self.config = config
+        self.act = 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.act(hidden_states)
+        hidden_states = self.fc2(hidden_states)
+        return hidden_states
+
+
+class InternVisionEncoderLayer(nn.Module):
+    def __init__(self, config: InternVisionConfig, drop_path_rate: float):
+        super().__init__()
+        self.embed_dim = config.hidden_size
+        self.intermediate_size = config.intermediate_size
+        self.norm_type = config.norm_type
+
+        self.attn = InternAttention(config)
+        self.mlp = InternMLP(config)
+        self.norm1 = NORM2FN[self.norm_type](self.embed_dim, eps=config.layer_norm_eps)
+        self.norm2 = NORM2FN[self.norm_type](self.embed_dim, eps=config.layer_norm_eps)
+
+        self.ls1 = nn.Parameter(config.initializer_factor * torch.ones(self.embed_dim))
+        self.ls2 = nn.Parameter(config.initializer_factor * torch.ones(self.embed_dim))
+        self.drop_path1 = DropPath(drop_path_rate) if drop_path_rate > 0. else nn.Identity()
+        self.drop_path2 = DropPath(drop_path_rate) if drop_path_rate > 0. else nn.Identity()
+
+    def forward(
+            self,
+            hidden_states: torch.Tensor,
+    ) -> Tuple[torch.FloatTensor, Optional[torch.FloatTensor], Optional[Tuple[torch.FloatTensor]]]:
+        """
+        Args:
+            hidden_states (`Tuple[torch.FloatTensor, Optional[torch.FloatTensor]]`): input to the layer of shape `(batch, seq_len, embed_dim)`
+        """
+        hidden_states = hidden_states + self.drop_path1(self.attn(self.norm1(hidden_states).to(hidden_states.dtype)) * self.ls1)
+
+        hidden_states = hidden_states + self.drop_path2(self.mlp(self.norm2(hidden_states).to(hidden_states.dtype)) * self.ls2)
+
+        return hidden_states
+
+
+class InternVisionEncoder(nn.Module):
+    """
+    Transformer encoder consisting of `config.num_hidden_layers` self attention layers. Each layer is a
+    [`InternEncoderLayer`].
+
+    Args:
+        config (`InternConfig`):
+            The corresponding vision configuration for the `InternEncoder`.
+    """
+
+    def __init__(self, config: InternVisionConfig):
+        super().__init__()
+        self.config = config
+        # stochastic depth decay rule
+        dpr = [x.item() for x in torch.linspace(0, config.drop_path_rate, config.num_hidden_layers)]
+        self.layers = nn.ModuleList([
+            InternVisionEncoderLayer(config, dpr[idx]) for idx in range(config.num_hidden_layers)])
+        self.gradient_checkpointing = True
+
+    def forward(
+            self,
+            inputs_embeds,
+            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)`):
+                Embedded representation of the inputs. Should be float, not int tokens.
+            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_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
+        hidden_states = inputs_embeds
+
+        for idx, encoder_layer in enumerate(self.layers):
+            if output_hidden_states:
+                encoder_states = encoder_states + (hidden_states,)
+            if self.gradient_checkpointing and self.training:
+                layer_outputs = torch.utils.checkpoint.checkpoint(
+                    encoder_layer,
+                    hidden_states)
+            else:
+                layer_outputs = encoder_layer(
+                    hidden_states,
+                )
+            hidden_states = layer_outputs
+
+        if output_hidden_states:
+            encoder_states = encoder_states + (hidden_states,)
+
+        if not return_dict:
+            return tuple(v for v in [hidden_states, encoder_states] if v is not None)
+        return BaseModelOutput(
+            last_hidden_state=hidden_states, hidden_states=encoder_states
+        )
+
+
+class InternVisionModel(PreTrainedModel):
+    main_input_name = 'pixel_values'
+    _supports_flash_attn_2 = True
+    config_class = InternVisionConfig
+    _no_split_modules = ['InternVisionEncoderLayer']
+
+    def __init__(self, config: InternVisionConfig):
+        super().__init__(config)
+        self.config = config
+
+        self.embeddings = InternVisionEmbeddings(config)
+        self.encoder = InternVisionEncoder(config)
+
+    def resize_pos_embeddings(self, old_size, new_size, patch_size):
+        pos_emb = self.embeddings.position_embedding
+        _, num_positions, embed_dim = pos_emb.shape
+        cls_emb = pos_emb[:, :1, :]
+        pos_emb = pos_emb[:, 1:, :].reshape(1, old_size // patch_size, old_size // patch_size, -1).permute(0, 3, 1, 2)
+        pos_emb = F.interpolate(pos_emb.float(), size=new_size // patch_size, mode='bicubic', align_corners=False)
+        pos_emb = pos_emb.to(cls_emb.dtype).reshape(1, embed_dim, -1).permute(0, 2, 1)
+        pos_emb = torch.cat([cls_emb, pos_emb], dim=1)
+        self.embeddings.position_embedding = nn.Parameter(pos_emb)
+        self.embeddings.image_size = new_size
+        logger.info('Resized position embeddings from {} to {}'.format(old_size, new_size))
+
+    def get_input_embeddings(self):
+        return self.embeddings
+
+    def forward(
+            self,
+            pixel_values: Optional[torch.FloatTensor] = None,
+            output_hidden_states: Optional[bool] = None,
+            return_dict: Optional[bool] = None,
+            pixel_embeds: Optional[torch.FloatTensor] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        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
+
+        if pixel_values is None and pixel_embeds is None:
+            raise ValueError('You have to specify pixel_values or pixel_embeds')
+
+        if pixel_embeds is not None:
+            hidden_states = pixel_embeds
+        else:
+            if len(pixel_values.shape) == 4:
+                hidden_states = self.embeddings(pixel_values)
+            else:
+                raise ValueError(f'wrong pixel_values size: {pixel_values.shape}')
+        encoder_outputs = self.encoder(
+            inputs_embeds=hidden_states,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        last_hidden_state = encoder_outputs.last_hidden_state
+        pooled_output = last_hidden_state[:, 0, :]
+
+        if not return_dict:
+            return (last_hidden_state, pooled_output) + encoder_outputs[1:]
+
+        return BaseModelOutputWithPooling(
+            last_hidden_state=last_hidden_state,
+            pooler_output=pooled_output,
+            hidden_states=encoder_outputs.hidden_states,
+            attentions=encoder_outputs.attentions,
+        )

modeling_internvl_chat.py

@@ -0,0 +1,348 @@
+# --------------------------------------------------------
+# InternVL
+# Copyright (c) 2024 OpenGVLab
+# Licensed under The MIT License [see LICENSE for details]
+# --------------------------------------------------------
+
+import warnings
+from typing import List, Optional, Tuple, Union
+
+import torch.utils.checkpoint
+import transformers
+from torch import nn
+from torch.nn import CrossEntropyLoss
+from transformers import (AutoModel, GenerationConfig, LlamaForCausalLM,
+                          Qwen2ForCausalLM)
+from transformers.modeling_outputs import CausalLMOutputWithPast
+from transformers.modeling_utils import PreTrainedModel
+from transformers.utils import ModelOutput, logging
+
+from .configuration_internvl_chat import InternVLChatConfig
+from .conversation import get_conv_template
+from .modeling_intern_vit import InternVisionModel, has_flash_attn
+
+logger = logging.get_logger(__name__)
+
+
+def version_cmp(v1, v2, op='eq'):
+    import operator
+
+    from packaging import version
+    op_func = getattr(operator, op)
+    return op_func(version.parse(v1), version.parse(v2))
+
+
+class InternVLChatModel(PreTrainedModel):
+    config_class = InternVLChatConfig
+    main_input_name = 'pixel_values'
+    base_model_prefix = 'language_model'
+    _supports_flash_attn_2 = True
+    _no_split_modules = ['InternVisionModel', 'LlamaDecoderLayer', 'Qwen2DecoderLayer']
+
+    def __init__(self, config: InternVLChatConfig, vision_model=None, language_model=None, use_flash_attn=True):
+        super().__init__(config)
+
+        assert version_cmp(transformers.__version__, '4.37.0', 'ge')
+        image_size = config.force_image_size or config.vision_config.image_size
+        patch_size = config.vision_config.patch_size
+        self.patch_size = patch_size
+        self.select_layer = config.select_layer
+        self.template = config.template
+        self.num_image_token = int((image_size // patch_size) ** 2 * (config.downsample_ratio ** 2))
+        self.downsample_ratio = config.downsample_ratio
+        self.ps_version = config.ps_version
+        use_flash_attn = use_flash_attn if has_flash_attn else False
+        config.vision_config.use_flash_attn = True if use_flash_attn else False
+        config.llm_config._attn_implementation = 'flash_attention_2' if use_flash_attn else 'eager'
+
+        logger.info(f'num_image_token: {self.num_image_token}')
+        logger.info(f'ps_version: {self.ps_version}')
+        if vision_model is not None:
+            self.vision_model = vision_model
+        else:
+            self.vision_model = InternVisionModel(config.vision_config)
+        if language_model is not None:
+            self.language_model = language_model
+        else:
+            if config.llm_config.architectures[0] == 'LlamaForCausalLM':
+                self.language_model = LlamaForCausalLM(config.llm_config)
+            elif config.llm_config.architectures[0] == 'Qwen2ForCausalLM':
+                self.language_model = Qwen2ForCausalLM(config.llm_config)
+            else:
+                raise NotImplementedError(f'{config.llm_config.architectures[0]} is not implemented.')
+
+        vit_hidden_size = config.vision_config.hidden_size
+        llm_hidden_size = config.llm_config.hidden_size
+
+        self.mlp1 = nn.Sequential(
+            nn.LayerNorm(vit_hidden_size * int(1 / self.downsample_ratio) ** 2),
+            nn.Linear(vit_hidden_size * int(1 / self.downsample_ratio) ** 2, llm_hidden_size),
+            nn.GELU(),
+            nn.Linear(llm_hidden_size, llm_hidden_size)
+        )
+
+        self.img_context_token_id = None
+        self.conv_template = get_conv_template(self.template)
+        self.system_message = self.conv_template.system_message
+
+    def forward(
+            self,
+            pixel_values: torch.FloatTensor,
+            input_ids: torch.LongTensor = None,
+            attention_mask: Optional[torch.Tensor] = None,
+            position_ids: Optional[torch.LongTensor] = None,
+            image_flags: Optional[torch.LongTensor] = None,
+            past_key_values: Optional[List[torch.FloatTensor]] = None,
+            labels: Optional[torch.LongTensor] = None,
+            use_cache: Optional[bool] = None,
+            output_attentions: Optional[bool] = None,
+            output_hidden_states: Optional[bool] = None,
+            return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, CausalLMOutputWithPast]:
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        image_flags = image_flags.squeeze(-1)
+        input_embeds = self.language_model.get_input_embeddings()(input_ids).clone()
+
+        vit_embeds = self.extract_feature(pixel_values)
+        vit_embeds = vit_embeds[image_flags == 1]
+        vit_batch_size = pixel_values.shape[0]
+
+        B, N, C = input_embeds.shape
+        input_embeds = input_embeds.reshape(B * N, C)
+
+        if torch.distributed.is_initialized() and torch.distributed.get_rank() == 0:
+            print(f'dynamic ViT batch size: {vit_batch_size}, images per sample: {vit_batch_size / B}, dynamic token length: {N}')
+
+        input_ids = input_ids.reshape(B * N)
+        selected = (input_ids == self.img_context_token_id)
+        try:
+            input_embeds[selected] = input_embeds[selected] * 0.0 + vit_embeds.reshape(-1, C)
+        except Exception as e:
+            vit_embeds = vit_embeds.reshape(-1, C)
+            print(f'warning: {e}, input_embeds[selected].shape={input_embeds[selected].shape}, '
+                  f'vit_embeds.shape={vit_embeds.shape}')
+            n_token = selected.sum()
+            input_embeds[selected] = input_embeds[selected] * 0.0 + vit_embeds[:n_token]
+
+        input_embeds = input_embeds.reshape(B, N, C)
+
+        outputs = self.language_model(
+            inputs_embeds=input_embeds,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_values=past_key_values,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        logits = outputs.logits
+
+        loss = None
+        if labels is not None:
+            # Shift so that tokens < n predict n
+            shift_logits = logits[..., :-1, :].contiguous()
+            shift_labels = labels[..., 1:].contiguous()
+            # Flatten the tokens
+            loss_fct = CrossEntropyLoss()
+            shift_logits = shift_logits.view(-1, self.language_model.config.vocab_size)
+            shift_labels = shift_labels.view(-1)
+            # Enable model parallelism
+            shift_labels = shift_labels.to(shift_logits.device)
+            loss = loss_fct(shift_logits, shift_labels)
+
+        if not return_dict:
+            output = (logits,) + outputs[1:]
+            return (loss,) + output if loss is not None else output
+
+        return CausalLMOutputWithPast(
+            loss=loss,
+            logits=logits,
+            past_key_values=outputs.past_key_values,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+    def pixel_shuffle(self, x, scale_factor=0.5):
+        n, w, h, c = x.size()
+        # N, W, H, C --> N, W, H * scale, C // scale
+        x = x.view(n, w, int(h * scale_factor), int(c / scale_factor))
+        # N, W, H * scale, C // scale --> N, H * scale, W, C // scale
+        x = x.permute(0, 2, 1, 3).contiguous()
+        # N, H * scale, W, C // scale --> N, H * scale, W * scale, C // (scale ** 2)
+        x = x.view(n, int(h * scale_factor), int(w * scale_factor),
+                   int(c / (scale_factor * scale_factor)))
+        if self.ps_version == 'v1':
+            warnings.warn("In ps_version 'v1', the height and width have not been swapped back, "
+                          'which results in a transposed image.')
+        else:
+            x = x.permute(0, 2, 1, 3).contiguous()
+        return x
+
+    def extract_feature(self, pixel_values):
+        if self.select_layer == -1:
+            vit_embeds = self.vision_model(
+                pixel_values=pixel_values,
+                output_hidden_states=False,
+                return_dict=True).last_hidden_state
+        else:
+            vit_embeds = self.vision_model(
+                pixel_values=pixel_values,
+                output_hidden_states=True,
+                return_dict=True).hidden_states[self.select_layer]
+        vit_embeds = vit_embeds[:, 1:, :]
+
+        h = w = int(vit_embeds.shape[1] ** 0.5)
+        vit_embeds = vit_embeds.reshape(vit_embeds.shape[0], h, w, -1)
+        vit_embeds = self.pixel_shuffle(vit_embeds, scale_factor=self.downsample_ratio)
+        vit_embeds = vit_embeds.reshape(vit_embeds.shape[0], -1, vit_embeds.shape[-1])
+        vit_embeds = self.mlp1(vit_embeds)
+        return vit_embeds
+
+    def batch_chat(self, tokenizer, pixel_values, questions, generation_config, num_patches_list=None,
+                   history=None, return_history=False, IMG_START_TOKEN='<img>', IMG_END_TOKEN='</img>',
+                   IMG_CONTEXT_TOKEN='<IMG_CONTEXT>', verbose=False, image_counts=None):
+        if history is not None or return_history:
+            print('Now multi-turn chat is not supported in batch_chat.')
+            raise NotImplementedError
+
+        if image_counts is not None:
+            num_patches_list = image_counts
+            print('Warning: `image_counts` is deprecated. Please use `num_patches_list` instead.')
+
+        img_context_token_id = tokenizer.convert_tokens_to_ids(IMG_CONTEXT_TOKEN)
+        self.img_context_token_id = img_context_token_id
+
+        if verbose and pixel_values is not None:
+            image_bs = pixel_values.shape[0]
+            print(f'dynamic ViT batch size: {image_bs}')
+
+        queries = []
+        for idx, num_patches in enumerate(num_patches_list):
+            question = questions[idx]
+            if pixel_values is not None and '<image>' not in question:
+                question = '<image>\n' + question
+            template = get_conv_template(self.template)
+            template.system_message = self.system_message
+            template.append_message(template.roles[0], question)
+            template.append_message(template.roles[1], None)
+            query = template.get_prompt()
+
+            image_tokens = IMG_START_TOKEN + IMG_CONTEXT_TOKEN * self.num_image_token * num_patches + IMG_END_TOKEN
+            query = query.replace('<image>', image_tokens, 1)
+            queries.append(query)
+
+        tokenizer.padding_side = 'left'
+        model_inputs = tokenizer(queries, return_tensors='pt', padding=True)
+        input_ids = model_inputs['input_ids'].to(self.device)
+        attention_mask = model_inputs['attention_mask'].to(self.device)
+        eos_token_id = tokenizer.convert_tokens_to_ids(template.sep.strip())
+        generation_config['eos_token_id'] = eos_token_id
+        generation_output = self.generate(
+            pixel_values=pixel_values,
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            **generation_config
+        )
+        responses = tokenizer.batch_decode(generation_output, skip_special_tokens=True)
+        responses = [response.split(template.sep.strip())[0].strip() for response in responses]
+        return responses
+
+    def chat(self, tokenizer, pixel_values, question, generation_config, history=None, return_history=False,
+             num_patches_list=None, IMG_START_TOKEN='<img>', IMG_END_TOKEN='</img>', IMG_CONTEXT_TOKEN='<IMG_CONTEXT>',
+             verbose=False):
+
+        if history is None and pixel_values is not None and '<image>' not in question:
+            question = '<image>\n' + question
+
+        if num_patches_list is None:
+            num_patches_list = [pixel_values.shape[0]] if pixel_values is not None else []
+        assert pixel_values is None or len(pixel_values) == sum(num_patches_list)
+
+        img_context_token_id = tokenizer.convert_tokens_to_ids(IMG_CONTEXT_TOKEN)
+        self.img_context_token_id = img_context_token_id
+
+        template = get_conv_template(self.template)
+        template.system_message = self.system_message
+        eos_token_id = tokenizer.convert_tokens_to_ids(template.sep.strip())
+
+        history = [] if history is None else history
+        for (old_question, old_answer) in history:
+            template.append_message(template.roles[0], old_question)
+            template.append_message(template.roles[1], old_answer)
+        template.append_message(template.roles[0], question)
+        template.append_message(template.roles[1], None)
+        query = template.get_prompt()
+
+        if verbose and pixel_values is not None:
+            image_bs = pixel_values.shape[0]
+            print(f'dynamic ViT batch size: {image_bs}')
+
+        for num_patches in num_patches_list:
+            image_tokens = IMG_START_TOKEN + IMG_CONTEXT_TOKEN * self.num_image_token * num_patches + IMG_END_TOKEN
+            query = query.replace('<image>', image_tokens, 1)
+
+        model_inputs = tokenizer(query, return_tensors='pt')
+        input_ids = model_inputs['input_ids'].to(self.device)
+        attention_mask = model_inputs['attention_mask'].to(self.device)
+        generation_config['eos_token_id'] = eos_token_id
+        generation_output = self.generate(
+            pixel_values=pixel_values,
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            **generation_config
+        )
+        response = tokenizer.batch_decode(generation_output, skip_special_tokens=True)[0]
+        response = response.split(template.sep.strip())[0].strip()
+        history.append((question, response))
+        if return_history:
+            return response, history
+        else:
+            query_to_print = query.replace(IMG_CONTEXT_TOKEN, '')
+            query_to_print = query_to_print.replace(f'{IMG_START_TOKEN}{IMG_END_TOKEN}', '<image>')
+            if verbose:
+                print(query_to_print, response)
+            return response
+
+    @torch.no_grad()
+    def generate(
+            self,
+            pixel_values: Optional[torch.FloatTensor] = None,
+            input_ids: Optional[torch.FloatTensor] = None,
+            attention_mask: Optional[torch.LongTensor] = None,
+            visual_features: Optional[torch.FloatTensor] = None,
+            generation_config: Optional[GenerationConfig] = None,
+            output_hidden_states: Optional[bool] = None,
+            **generate_kwargs,
+    ) -> torch.LongTensor:
+
+        assert self.img_context_token_id is not None
+        if pixel_values is not None:
+            if visual_features is not None:
+                vit_embeds = visual_features
+            else:
+                vit_embeds = self.extract_feature(pixel_values)
+            input_embeds = self.language_model.get_input_embeddings()(input_ids)
+            B, N, C = input_embeds.shape
+            input_embeds = input_embeds.reshape(B * N, C)
+
+            input_ids = input_ids.reshape(B * N)
+            selected = (input_ids == self.img_context_token_id)
+            assert selected.sum() != 0
+            input_embeds[selected] = vit_embeds.reshape(-1, C).to(input_embeds.device)
+
+            input_embeds = input_embeds.reshape(B, N, C)
+        else:
+            input_embeds = self.language_model.get_input_embeddings()(input_ids)
+
+        outputs = self.language_model.generate(
+            inputs_embeds=input_embeds,
+            attention_mask=attention_mask,
+            generation_config=generation_config,
+            output_hidden_states=output_hidden_states,
+            use_cache=True,
+            **generate_kwargs,
+        )
+
+        return outputs

openvino_config.json

@@ -0,0 +1,28 @@
+{
+  "compression": null,
+  "dtype": "int8",
+  "input_info": null,
+  "optimum_version": "1.24.0.dev0",
+  "quantization_config": {
+    "all_layers": null,
+    "backup_precision": null,
+    "bits": 8,
+    "dataset": null,
+    "gptq": null,
+    "group_size": -1,
+    "ignored_scope": null,
+    "lora_correction": null,
+    "num_samples": null,
+    "processor": null,
+    "quant_method": "default",
+    "ratio": 1,
+    "scale_estimation": null,
+    "sensitivity_metric": null,
+    "sym": false,
+    "tokenizer": null,
+    "trust_remote_code": false,
+    "weight_format": "int8"
+  },
+  "save_onnx_model": false,
+  "transformers_version": "4.47.0"
+}

openvino_detokenizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a75fa1f91293966b14a7e48d848bf123b5478c286d745dcb79b042a2a3ab63cc
+size 2189371

openvino_detokenizer.xml

@@ -0,0 +1,279 @@
+<?xml version="1.0"?>
+<net name="detokenizer" version="11">
+	<layers>
+		<layer id="0" name="Parameter_134755" type="Parameter" version="opset1">
+			<data shape="?,?" element_type="i64" />
+			<output>
+				<port id="0" precision="I64" names="Parameter_134755">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="1" name="Convert_134771" 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_134718" type="Const" version="opset1">
+			<data element_type="i32" shape="151655" offset="0" size="606620" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151655</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="3" name="Constant_134720" type="Const" version="opset1">
+			<data element_type="i32" shape="151655" offset="606620" size="606620" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151655</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="4" name="Constant_134722" type="Const" version="opset1">
+			<data element_type="u8" shape="976063" offset="1213240" size="976063" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>976063</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="5" name="Constant_134759" type="Const" version="opset1">
+			<data element_type="i32" shape="12" offset="2189303" size="48" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>12</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="6" name="Constant_134757" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="2189351" size="4" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="7" name="Constant_134756" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="2189355" size="4" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="8" name="Constant_134758" type="Const" version="opset1">
+			<data element_type="i32" shape="1" offset="2189359" size="4" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="9" name="Constant_134761" type="Const" version="opset1">
+			<data element_type="i64" shape="1" offset="2189363" size="8" />
+			<output>
+				<port id="0" precision="I64">
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="10" name="Slice_134760" type="Slice" version="opset8">
+			<input>
+				<port id="0" precision="I32">
+					<dim>12</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="I64">
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="5" precision="I32">
+					<dim>12</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="11" name="VocabDecoder_134762" 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>151655</dim>
+				</port>
+				<port id="2" precision="I32">
+					<dim>151655</dim>
+				</port>
+				<port id="3" precision="U8">
+					<dim>976063</dim>
+				</port>
+				<port id="4" precision="I32">
+					<dim>12</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="12" name="FuzeRagged_134763" 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="13" name="UTF8Validate_134764" 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="14" name="StringTensorPack_134765" 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="string_output">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="15" name="Result_134766" type="Result" version="opset1">
+			<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="11" to-port="0" />
+		<edge from-layer="2" from-port="0" to-layer="11" to-port="1" />
+		<edge from-layer="3" from-port="0" to-layer="11" to-port="2" />
+		<edge from-layer="4" from-port="0" to-layer="11" to-port="3" />
+		<edge from-layer="5" from-port="0" to-layer="10" to-port="0" />
+		<edge from-layer="6" from-port="0" to-layer="10" to-port="1" />
+		<edge from-layer="7" from-port="0" to-layer="10" to-port="2" />
+		<edge from-layer="8" from-port="0" to-layer="10" to-port="3" />
+		<edge from-layer="9" from-port="0" to-layer="10" to-port="4" />
+		<edge from-layer="10" from-port="5" to-layer="11" to-port="4" />
+		<edge from-layer="11" from-port="5" to-layer="12" to-port="0" />
+		<edge from-layer="11" from-port="9" to-layer="13" to-port="2" />
+		<edge from-layer="11" from-port="8" to-layer="12" to-port="3" />
+		<edge from-layer="11" from-port="7" to-layer="12" to-port="2" />
+		<edge from-layer="11" from-port="6" to-layer="12" to-port="1" />
+		<edge from-layer="12" from-port="4" to-layer="13" to-port="0" />
+		<edge from-layer="12" from-port="5" to-layer="13" to-port="1" />
+		<edge from-layer="13" from-port="3" to-layer="14" to-port="0" />
+		<edge from-layer="13" from-port="4" to-layer="14" to-port="1" />
+		<edge from-layer="13" from-port="5" to-layer="14" to-port="2" />
+		<edge from-layer="14" from-port="3" to-layer="15" to-port="0" />
+	</edges>
+	<rt_info>
+		<add_attention_mask value="True" />
+		<add_prefix_space />
+		<add_special_tokens value="True" />
+		<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 />
+		<number_of_inputs value="1" />
+		<openvino_tokenizers_version value="2025.0.0.0-471-7f2ac4f9648" />
+		<openvino_version value="2025.0.0-17908-513dcc5c7b7-releases/2025/0" />
+		<original_tokenizer_class value="&lt;class 'transformers.models.qwen2.tokenization_qwen2_fast.Qwen2TokenizerFast'>" />
+		<pad_token_id value="151643" />
+		<sentencepiece_version value="0.2.0" />
+		<skip_special_tokens value="True" />
+		<streaming_detokenizer value="False" />
+		<tiktoken_version value="0.7.0" />
+		<tokenizer_output_type value="i64" />
+		<tokenizers_version value="0.21.0" />
+		<transformers_version value="4.47.0" />
+		<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 sha256:1082d96fb77e1f96e7df4797ff48f1cfd9c9d2577a4004aba4bd0415c4b95b73
+size 495363421

openvino_text_embeddings_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3b56cdebfe4b6d884386a1fece4ebd7a9350f0ea575aeba1c02b465f87cd180
+size 136186194

openvino_text_embeddings_model.xml

@@ -0,0 +1,173 @@
+<?xml version="1.0"?>
+<net name="Model6" 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="151655, 896" offset="0" size="135882880" />
+			<output>
+				<port id="0" precision="I8">
+					<dim>151655</dim>
+					<dim>896</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="2" name="Convert_128423" type="Convert" version="opset1">
+			<data destination_type="f16" />
+			<input>
+				<port id="0" precision="I8">
+					<dim>151655</dim>
+					<dim>896</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP16">
+					<dim>151655</dim>
+					<dim>896</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="3" name="self.weight/scale" type="Const" version="opset1">
+			<data element_type="f16" shape="151655, 1" offset="135882880" size="303310" />
+			<output>
+				<port id="0" precision="FP16">
+					<dim>151655</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>151655</dim>
+					<dim>896</dim>
+				</port>
+				<port id="1" precision="FP16">
+					<dim>151655</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP16">
+					<dim>151655</dim>
+					<dim>896</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="5" name="self.weight/fq_weights_0/convert" type="Convert" version="opset1">
+			<data destination_type="f32" />
+			<input>
+				<port id="0" precision="FP16">
+					<dim>151655</dim>
+					<dim>896</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" precision="FP32">
+					<dim>151655</dim>
+					<dim>896</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="6" name="aten::embedding/Convert" 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="aten::embedding/Constant" type="Const" version="opset1">
+			<data element_type="i32" shape="" offset="136186190" size="4" />
+			<output>
+				<port id="0" precision="I32" />
+			</output>
+		</layer>
+		<layer id="8" name="aten::embedding/Gather" type="Gather" version="opset8">
+			<data batch_dims="0" />
+			<input>
+				<port id="0" precision="FP32">
+					<dim>151655</dim>
+					<dim>896</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>896</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="9" name="Result_99850" type="Result" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>-1</dim>
+					<dim>-1</dim>
+					<dim>896</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="2025.0.0-17908-513dcc5c7b7-releases/2025/0" />
+		<conversion_parameters>
+			<framework value="pytorch" />
+			<is_python_object value="True" />
+		</conversion_parameters>
+		<nncf>
+			<friendly_names_were_updated value="True" />
+			<weight_compression>
+				<advanced_parameters value="{'statistics_path': None, 'awq_params': {'subset_size': 32, 'percent_to_apply': 0.002, 'alpha_min': 0.0, 'alpha_max': 1.0, 'steps': 100}, '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}}" />
+				<all_layers value="False" />
+				<awq value="False" />
+				<backup_mode value="int8_asym" />
+				<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>
+			<optimum_intel_version value="1.22.0.dev0+2b0d642" />
+			<optimum_version value="1.24.0.dev0" />
+			<pytorch_version value="2.5.1+cpu" />
+			<transformers_version value="4.47.0" />
+		</optimum>
+	</rt_info>
+</net>

openvino_tokenizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3a335c7a9e16bcc547408ea0e61cda12049f13d512689757a358310492ace17
+size 5584911

openvino_tokenizer.xml

@@ -0,0 +1,719 @@
+<?xml version="1.0"?>
+<net name="tokenizer" version="11">
+	<layers>
+		<layer id="0" name="Parameter_134636" type="Parameter" version="opset1">
+			<data shape="?" element_type="string" />
+			<output>
+				<port id="0" precision="STRING" names="Parameter_134636">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="1" name="Constant_134642" 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_134637" 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_134638" 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_134639" 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_134640" 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_134641" 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_134643" 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_134644" 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_134645" 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_134646" 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_134647" 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_134648" 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_134649" 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_134711" type="Const" version="opset1">
+			<data element_type="u8" shape="163" offset="16" size="163" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>163</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="15" name="SpecialTokensSplit_134712" 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>163</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="NormalizeUnicode_134713" type="NormalizeUnicode" version="extension">
+			<data normalization_form="NFC" />
+			<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_134715" type="Const" version="opset1">
+			<data element_type="u8" shape="110" offset="179" size="110" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>110</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="18" name="RegexSplit_134716" 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_134718" type="Const" version="opset1">
+			<data element_type="i32" shape="151655" offset="289" size="606620" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151655</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="20" name="Constant_134720" type="Const" version="opset1">
+			<data element_type="i32" shape="151655" offset="606909" size="606620" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151655</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="21" name="Constant_134722" type="Const" version="opset1">
+			<data element_type="u8" shape="976063" offset="1213529" size="976063" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>976063</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="22" name="Constant_134730" type="Const" version="opset1">
+			<data element_type="i32" shape="151291" offset="2189592" size="605164" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151291</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="23" name="Constant_134732" type="Const" version="opset1">
+			<data element_type="i32" shape="151291" offset="2794756" size="605164" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151291</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="24" name="Constant_134734" type="Const" version="opset1">
+			<data element_type="u8" shape="490634" offset="3399920" size="490634" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>490634</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="25" name="Constant_134736" type="Const" version="opset1">
+			<data element_type="i32" shape="151291" offset="3890554" size="605164" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151291</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="26" name="Constant_134738" type="Const" version="opset1">
+			<data element_type="i32" shape="151291" offset="4495718" size="605164" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>151291</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="27" name="Constant_134740" type="Const" version="opset1">
+			<data element_type="u8" shape="483779" offset="5100882" size="483779" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>483779</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="28" name="Constant_134724" type="Const" version="opset1">
+			<data element_type="i32" shape="12" offset="5584661" size="48" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>12</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="29" name="Constant_134726" type="Const" version="opset1">
+			<data element_type="i32" shape="12" offset="5584709" size="48" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>12</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="30" name="Constant_134728" type="Const" version="opset1">
+			<data element_type="u8" shape="94" offset="5584757" size="94" />
+			<output>
+				<port id="0" precision="U8">
+					<dim>94</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="31" name="Constant_134741" type="Const" version="opset1">
+			<data element_type="i32" shape="12" offset="5584851" size="48" />
+			<output>
+				<port id="0" precision="I32">
+					<dim>12</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="32" name="BPETokenizer_134742" 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>151655</dim>
+				</port>
+				<port id="6" precision="I32">
+					<dim>151655</dim>
+				</port>
+				<port id="7" precision="U8">
+					<dim>976063</dim>
+				</port>
+				<port id="8" precision="I32">
+					<dim>151291</dim>
+				</port>
+				<port id="9" precision="I32">
+					<dim>151291</dim>
+				</port>
+				<port id="10" precision="U8">
+					<dim>490634</dim>
+				</port>
+				<port id="11" precision="I32">
+					<dim>151291</dim>
+				</port>
+				<port id="12" precision="I32">
+					<dim>151291</dim>
+				</port>
+				<port id="13" precision="U8">
+					<dim>483779</dim>
+				</port>
+				<port id="14" precision="I32">
+					<dim>12</dim>
+				</port>
+				<port id="15" precision="I32">
+					<dim>12</dim>
+				</port>
+				<port id="16" precision="U8">
+					<dim>94</dim>
+				</port>
+				<port id="17" precision="I32">
+					<dim>12</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="Subtract_134743" 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="34" name="Constant_134744" type="Const" version="opset1">
+			<data element_type="i32" shape="" offset="5584899" size="4" />
+			<output>
+				<port id="0" precision="I32" />
+			</output>
+		</layer>
+		<layer id="35" name="Minimum_134745" type="Minimum" version="opset1">
+			<data auto_broadcast="numpy" />
+			<input>
+				<port id="0" precision="I32">
+					<dim>-1</dim>
+				</port>
+				<port id="1" precision="I32" />
+			</input>
+			<output>
+				<port id="2" precision="I32">
+					<dim>-1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="36" name="Add_134746" type="Add" 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="37" name="Subtract_134747" 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="38" name="Constant_134748" type="Const" version="opset1">
+			<data element_type="i32" shape="" offset="5584903" size="4" />
+			<output>
+				<port id="0" precision="I32" />
+			</output>
+		</layer>
+		<layer id="39" name="ReduceMax_134749" 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="40" name="Constant_134750" type="Const" version="opset1">
+			<data element_type="i32" shape="" offset="5584907" size="4" />
+			<output>
+				<port id="0" precision="I32" />
+			</output>
+		</layer>
+		<layer id="41" name="RaggedToDense_134751" type="RaggedToDense" version="extension">
+			<data pad_right="true" />
+			<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="42" name="Convert_134752" 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="43" name="Convert_134752.0" 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="45" name="RaggedToDense_134751.0" 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="46" name="Result_134753" type="Result" version="opset1">
+			<input>
+				<port id="0" precision="I64">
+					<dim>-1</dim>
+					<dim>-1</dim>
+				</port>
+			</input>
+		</layer>
+		<layer id="44" name="Result_134754" type="Result" version="opset1">
+			<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="3" to-layer="15" to-port="4" />
+		<edge from-layer="2" from-port="2" to-layer="15" to-port="3" />
+		<edge from-layer="2" from-port="1" to-layer="15" to-port="2" />
+		<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="11" to-port="0" />
+		<edge from-layer="6" from-port="3" 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="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="7" to-layer="18" to-port="1" />
+		<edge from-layer="15" from-port="6" to-layer="18" to-port="0" />
+		<edge from-layer="15" from-port="11" to-layer="16" to-port="3" />
+		<edge from-layer="15" from-port="10" to-layer="16" to-port="2" />
+		<edge from-layer="15" from-port="9" to-layer="16" 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="11" to-layer="32" to-port="4" />
+		<edge from-layer="18" from-port="10" to-layer="32" to-port="3" />
+		<edge from-layer="18" from-port="8" to-layer="32" to-port="1" />
+		<edge from-layer="18" from-port="7" to-layer="32" to-port="0" />
+		<edge from-layer="18" from-port="9" to-layer="32" to-port="2" />
+		<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="19" to-layer="33" to-port="0" />
+		<edge from-layer="32" from-port="18" to-layer="33" to-port="1" />
+		<edge from-layer="32" from-port="20" to-layer="41" to-port="2" />
+		<edge from-layer="32" from-port="18" to-layer="36" to-port="0" />
+		<edge from-layer="32" from-port="18" to-layer="41" to-port="0" />
+		<edge from-layer="32" from-port="18" to-layer="37" 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="36" to-port="1" />
+		<edge from-layer="36" from-port="2" to-layer="37" to-port="0" />
+		<edge from-layer="36" from-port="2" to-layer="41" to-port="1" />
+		<edge from-layer="37" from-port="2" to-layer="39" to-port="0" />
+		<edge from-layer="38" from-port="0" to-layer="39" to-port="1" />
+		<edge from-layer="39" from-port="2" to-layer="41" to-port="3" />
+		<edge from-layer="40" from-port="0" to-layer="41" to-port="4" />
+		<edge from-layer="41" from-port="6" to-layer="42" to-port="0" />
+		<edge from-layer="41" from-port="5" to-layer="45" to-port="0" />
+		<edge from-layer="42" from-port="1" to-layer="43" to-port="0" />
+		<edge from-layer="43" from-port="1" to-layer="44" to-port="0" />
+		<edge from-layer="45" from-port="1" to-layer="46" to-port="0" />
+	</edges>
+	<rt_info>
+		<add_attention_mask value="True" />
+		<add_prefix_space />
+		<add_special_tokens value="True" />
+		<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 />
+		<number_of_inputs value="1" />
+		<openvino_tokenizers_version value="2025.0.0.0-471-7f2ac4f9648" />
+		<openvino_version value="2025.0.0-17908-513dcc5c7b7-releases/2025/0" />
+		<original_tokenizer_class value="&lt;class 'transformers.models.qwen2.tokenization_qwen2_fast.Qwen2TokenizerFast'>" />
+		<pad_token_id value="151643" />
+		<sentencepiece_version value="0.2.0" />
+		<skip_special_tokens value="True" />
+		<streaming_detokenizer value="False" />
+		<tiktoken_version value="0.7.0" />
+		<tokenizer_output_type value="i64" />
+		<tokenizers_version value="0.21.0" />
+		<transformers_version value="4.47.0" />
+		<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:d8995f4e5a7e271498532645b80d33339058533a754961c5bd557a93e3c583f7
+size 313037732

preprocessor_config.json

@@ -0,0 +1,27 @@
+{
+  "crop_size": {
+    "height": 448,
+    "width": 448
+  },
+  "do_center_crop": true,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.485,
+    0.456,
+    0.406
+  ],
+  "image_processor_type": "CLIPFeatureExtractor",
+  "image_std": [
+    0.229,
+    0.224,
+    0.225
+  ],
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 448
+  }
+}

special_tokens_map.json

@@ -0,0 +1,29 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<img>",
+    "</img>",
+    "<IMG_CONTEXT>",
+    "<quad>",
+    "</quad>",
+    "<ref>",
+    "</ref>",
+    "<box>",
+    "</box>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:63a3df087afa0c8890b416fe2c4f3f2593f06181adda4aacf93db8eb34cf0208
+size 11414752

tokenizer_config.json

@@ -0,0 +1,126 @@
+{
+  "add_eos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<img>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "</img>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<IMG_CONTEXT>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<quad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "</quad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<ref>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "</ref>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<box>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "</box>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<img>",
+    "</img>",
+    "<IMG_CONTEXT>",
+    "<quad>",
+    "</quad>",
+    "<ref>",
+    "</ref>",
+    "<box>",
+    "</box>"
+  ],
+  "bos_token": null,
+  "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}