Upload 16 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

InfiMed.py

@@ -0,0 +1,522 @@
+from pathlib import Path
+from typing import Any, Callable, List, Optional, Tuple, Union
+
+import torch
+import torch.nn as nn
+import os
+from accelerate import PartialState
+import PIL
+from transformers import PreTrainedModel, PretrainedConfig, GenerationConfig, AutoTokenizer, LlamaTokenizerFast
+from transformers.utils import WEIGHTS_NAME, SAFE_WEIGHTS_NAME
+from transformers import Qwen3ForCausalLM, SiglipImageProcessor
+from safetensors.torch import load_file
+from transformers.modeling_outputs import CausalLMOutputWithPast
+from modeling_siglip import SiglipVisionModel
+from configuration_siglip import SiglipVisionConfig
+from configuration_qwen3 import Qwen3Config
+from abc import ABC, abstractmethod
+from einops import rearrange
+
+
+IGNORE_INDEX = -100
+IMAGE_TOKEN_INDEX = -200
+
+
+class PromptBuilder(ABC):
+    def __init__(self, system_prompt: Optional[str] = None) -> None:
+        # Only some models define a system prompt => let subclasses handle this logic!
+        self.system_prompt = system_prompt
+
+    @abstractmethod
+    def add_turn(self, role: str, message: str) -> str: ...
+
+    @abstractmethod
+    def get_potential_prompt(self, user_msg: str) -> None: ...
+
+    @abstractmethod
+    def get_prompt(self) -> str: ...
+
+class Qwen3PromptBuilder(PromptBuilder):
+    def __init__(self, system_prompt: Optional[str] = None) -> None:
+        super().__init__(system_prompt)
+        self.system_prompt = "<s><|im_start|>system\nYou are a helpful assistant.<|im_end|>\n"
+        self.bos, self.eos = "<s>", "<|im_end|>"
+
+        # Get role-specific "wrap" functions
+        self.wrap_human = lambda msg: f"<|im_start|>user\n{msg}<|im_end|>assistant\n"
+        self.wrap_gpt = lambda msg: f"{msg if msg != '' else ' '}{self.eos}\n"
+
+        # === `self.prompt` gets built up over multiple turns ===
+        self.prompt, self.turn_count = "", 0
+
+    def add_turn(self, role: str, message: str) -> str:
+        # assert (role == "human") if (self.turn_count % 2 == 0) else (role == "gpt")
+        message = message.strip() #.replace("<image>", "").strip()
+
+        # Special Handling for "system" prompt (turn_count == 0)
+        if self.turn_count == 0:
+            sys_message = self.system_prompt + self.wrap_human(message)
+            wrapped_message = sys_message
+        elif (self.turn_count % 2) == 0:
+            human_message = self.wrap_human(message)
+            wrapped_message = human_message
+        else:
+            gpt_message = self.wrap_gpt(message)
+            wrapped_message = gpt_message
+
+        # Update Prompt
+        self.prompt += wrapped_message
+
+
+        # Bump Turn Counter
+        self.turn_count += 1
+
+        # Return "wrapped_message" (effective string added to context)
+        return wrapped_message
+
+    def get_potential_prompt(self, message: str) -> None:
+        # Assumes that it's always the user's (human's) turn!
+        prompt_copy = str(self.prompt)
+
+        # Special Handling for "system" prompt (turn_count == 0)
+        if self.turn_count == 0:
+            sys_message = self.system_prompt + self.wrap_human(message)
+            prompt_copy += sys_message
+
+        else:
+            human_message = self.wrap_human(message)
+            prompt_copy += human_message
+
+        # return prompt_copy.removeprefix(self.bos).rstrip()
+        return prompt_copy.rstrip()
+
+    def get_prompt(self) -> str:
+        # Remove prefix <bos> (if exists) because it gets auto-inserted by tokenizer!
+        # return self.prompt.removeprefix(self.bos).rstrip()
+        return self.prompt.rstrip()
+
+class InfiMedConfig(PretrainedConfig):
+    
+    def __init__(
+        self,
+        vision_config=None,
+        llm_config=None,
+        run_dir: str = None,
+        load_precision: str = "bf16",
+        max_length: int = 128,
+        temperature: float = 1.0,
+        **kwargs
+    ):
+        if vision_config is None:
+            vision_config = {}
+            print(
+                'vision_config is None. Initializing the SiglipVisionConfig with default values.')
+
+        if llm_config is None:
+            llm_config = {'architectures': ['Qwen3ForCausalLM']}
+            print(
+                'llm_config is None. Initializing the Qwen3Config config with default values')
+
+        self.vision_config = SiglipVisionConfig(**vision_config)
+        if llm_config['architectures'][0] == 'Qwen3ForCausalLM':
+            self.llm_config = Qwen3Config(**llm_config)
+        else:
+            raise ValueError('Unsupported architecture: {}'.format(
+                llm_config['architectures'][0]))
+        self.run_dir = run_dir
+        self.load_precision = load_precision
+        self.max_length = max_length
+        self.temperature = temperature
+        super().__init__(**kwargs)
+
+class AvgPoolProjector(nn.Module):
+    def __init__(
+        self,
+        layer_num: int = 2,
+        query_num: int = 144,
+        mm_hidden_size: int = 1024,
+        llm_hidden_size: int = 4096,
+    ):
+        super().__init__()
+        self.layer_num = layer_num
+        self.query_num = query_num
+        self.mm_hidden_size = mm_hidden_size
+        self.llm_hidden_size = llm_hidden_size
+        self.build_net()
+        
+    def build_net(self):
+        hw = int(self.query_num ** 0.5)
+        sampler = nn.AdaptiveAvgPool2d((hw, hw))
+        self.sampler = sampler
+        modules = [nn.Linear(self.mm_hidden_size, self.llm_hidden_size)]
+        for _ in range(1, self.layer_num):
+            modules.append(nn.GELU())
+            modules.append(nn.Linear(self.llm_hidden_size, self.llm_hidden_size))
+        self.mlp_projector = nn.Sequential(*modules)
+        print(f"patch size {hw} average pooling layer initialized")
+        
+    def forward(self, visual_feat: torch.Tensor) -> torch.Tensor:
+        batch_size, seq_len, h_dim = visual_feat.shape 
+        hw = int(seq_len ** 0.5) 
+        shaped_visual_feat = rearrange(visual_feat, "b (h w) d -> b d h w", h=hw, w=hw) 
+        pooled_visual_feat = self.sampler(shaped_visual_feat) 
+        reshaped_visual_feat = rearrange(pooled_visual_feat, "b d h w -> b (h w) d") 
+        output_feat = self.mlp_projector(reshaped_visual_feat) 
+        return output_feat
+
+class InfiMed(PreTrainedModel):
+    config_class = InfiMedConfig
+    
+    def __init__(self, config: InfiMedConfig, vision_model=None, language_model=None):
+        super().__init__(config)
+        self.run_dir = Path(config.run_dir) if config.run_dir else None
+        self.model_dtype = {"fp32": torch.float32, "fp16": torch.float16, "bf16": torch.bfloat16}[config.load_precision]
+        self.distributed_state = PartialState()
+        self.max_new_tokens = config.max_length
+        self.temperature = config.temperature
+        self.top_p = config.top_p
+        self.repetition_penalty = config.repetition_penalty
+
+
+        if vision_model is not None:
+            self.vision_model = vision_model
+        else:
+            self.vision_model = SiglipVisionModel.from_pretrained(config.vision_config._name_or_path, hidden_act = "gelu")
+
+        if language_model is not None:
+            self.language_model = language_model
+            self.config.llm_config = language_model.config
+        else:
+            if config.llm_config.architectures[0] == 'Qwen3ForCausalLM':
+                self.language_model = Qwen3ForCausalLM.from_pretrained(config.llm_config._name_or_path, pad_token_id = 151670, bos_token_id = 128245, eos_token_id = 151645, tie_word_embeddings = False)
+            else:
+                raise NotImplementedError(
+                    f'{config.llm_config.architectures[0]} is not implemented.')
+    
+        self.tokenizer = AutoTokenizer.from_pretrained(config.llm_config._name_or_path, use_fast=True)
+        self.tokenizer.add_special_tokens({"additional_special_tokens": ["<|endofchunk|>", "<s>", "<|pad|>"]})
+        self.tokenizer.pad_token = "<|pad|>"
+        self.tokenizer.bos_token = "<s>"
+
+        self.offset = 1 if self.tokenizer.encode("\n")[0] == self.tokenizer.bos_token_id else 0
+
+        if "finetune" in config.run_dir:
+            self.arch_specifier = "full-align+729-avgpool"
+        else:
+            self.arch_specifier = "no-align+avgpool"
+
+        if self.arch_specifier.split("+")[-1].split("-")[0] != "avgpool":
+            query_dim = int(self.arch_specifier.split("+")[-1].split("-")[0])
+        else:
+            query_dim = 144
+        self.projector = AvgPoolProjector(query_num=query_dim, mm_hidden_size=config.vision_config.hidden_size, llm_hidden_size=config.llm_config.hidden_size)
+
+        self.vision_backbone_requires_grad = False
+
+        self.img_context_token_id = 151655
+
+        self.image_processor = SiglipImageProcessor.from_pretrained(
+            config.vision_config._name_or_path,
+            size={"height": 384, "width": 384},  
+            resample=PIL.Image.Resampling.BICUBIC,  
+            crop_size={"height": 384, "width": 384},  
+            do_center_crop=True,  
+            do_normalize=True,  
+            image_mean=[0.5, 0.5, 0.5],  
+            image_std=[0.5, 0.5, 0.5],  
+            do_convert_rgb=True  
+        )
+
+
+    @classmethod
+    # load model from .pt file
+    def from_pretrained_ckpt(cls, pretrained_model_name_or_path, *args, **kwargs):
+        config = InfiMedConfig.from_pretrained(pretrained_model_name_or_path, *args, **kwargs)
+        model = cls(config)
+        ckpt_base_path = os.path.join(os.path.dirname(pretrained_model_name_or_path), "checkpoints")
+        if (Path(ckpt_base_path) / SAFE_WEIGHTS_NAME).exists():
+            state_dict = load_file(Path(ckpt_base_path) / SAFE_WEIGHTS_NAME)
+        elif (Path(ckpt_base_path) / WEIGHTS_NAME).exists():
+            state_dict = torch.load(Path(ckpt_base_path) / WEIGHTS_NAME, map_location="cpu")["model"]
+        elif (Path(ckpt_base_path) / "latest-checkpoint.pt").exists():
+            state_dict = torch.load(Path(ckpt_base_path) / "latest-checkpoint.pt", map_location="cpu")["model"]
+        else:
+            raise FileNotFoundError("No model weights found in the directory.")
+        if "vision_backbone" in state_dict:
+            model.vision_model.load_state_dict(state_dict["vision_backbone"])
+
+        new_state_dict = {}
+        for key, value in state_dict["llm_backbone"].items():
+            new_key = key.replace("llm.", "")
+            new_state_dict[new_key] = value
+        model.language_model.load_state_dict(new_state_dict)
+        model.projector.load_state_dict(state_dict["projector"])
+
+        model.to("cuda", dtype=torch.bfloat16)
+
+        model.requires_grad_(False)
+        model.eval()
+        return model
+
+    def save_checkpoint(self, save_path):
+        os.makedirs(save_path, exist_ok=True)
+        self.save_pretrained(save_path)
+        self.tokenizer.save_pretrained(save_path)
+        self.image_processor.save_pretrained(save_path)
+
+
+    def process_messages(self,messages):
+        prompt_builder = Qwen3PromptBuilder()
+        if "image" in messages:
+            processed_prompt = "<image>" + "\n" + messages['prompt'].replace("<image>", '')
+        elif "images" in messages:
+            processed_prompt = ""
+            for i, image in enumerate(messages['images']):
+                processed_prompt += f"<image_{i+1}>: "
+            processed_prompt += "\n" + messages['prompt'].replace("<image>", '')
+
+        msg = prompt_builder.add_turn("user", processed_prompt)
+        msg = msg.strip()
+
+        if isinstance(self.tokenizer, LlamaTokenizerFast):
+            msg = msg.rstrip()
+        else:
+            pass
+
+        turn_input_ids, _ = tokenizer_image_token(msg, self.tokenizer)
+
+        result = []
+        for x in turn_input_ids:
+            if x == -200:
+                result.extend([self.img_context_token_id] * 729)
+            else:
+                result.append(x)
+
+        turn_input_ids = result
+
+        input_ids = torch.tensor(turn_input_ids)
+
+        input_ids = input_ids[: self.tokenizer.model_max_length]
+
+        input_ids = input_ids.unsqueeze(0)
+
+        if "image" in messages:
+            pixel_values = self.image_processor(images=messages["image"], return_tensors="pt")["pixel_values"]
+        else:
+            pixel_values = None
+
+        input_ids = input_ids.to("cuda")
+        pixel_values = pixel_values.to("cuda") if pixel_values is not None else None
+
+        return input_ids, pixel_values
+
+    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,
+    ) -> CausalLMOutputWithPast:
+
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        vit_embeds = self.extract_feature(pixel_values)
+
+        input_embeds = self.language_model.get_input_embeddings()(input_ids)
+
+        vit_batch_size = pixel_values.shape[0]
+
+        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)
+        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)
+
+        if attention_mask is None:
+            batch_size = input_embeds.shape[0]
+            max_len = input_embeds.shape[1]
+            attention_mask = torch.zeros((batch_size, max_len), device=input_embeds.device).bool()
+            for index in range(batch_size):
+                if getattr(self.tokenizer, 'tokenizer_padding_side', 'right') == 'left':
+                    attention_mask[index, -max_len:] = True
+                else:
+                    attention_mask[index, :max_len] = True
+
+
+        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 extract_feature(self, pixel_values):
+        vit_embeds = self.vision_model(
+            pixel_values=pixel_values,
+            output_hidden_states=True,
+            return_dict=True).hidden_states[-2]
+
+        h = w = int(vit_embeds.shape[1] ** 0.5)
+        vit_embeds = vit_embeds.reshape(vit_embeds.shape[0], h, w, -1)
+        vit_embeds = vit_embeds.reshape(
+            vit_embeds.shape[0], -1, vit_embeds.shape[-1])
+        vit_embeds = self.projector(vit_embeds)
+        return vit_embeds
+
+
+    @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,
+            return_dict: 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)
+
+
+        if attention_mask is None:
+            batch_size = input_embeds.shape[0]
+            max_len = input_embeds.shape[1]
+            attention_mask = torch.zeros((batch_size, max_len), device=input_embeds.device).bool()
+            for index in range(batch_size):
+                if getattr(self.tokenizer, 'tokenizer_padding_side', 'right') == 'left':
+                    attention_mask[index, -max_len:] = True
+                else:
+                    attention_mask[index, :max_len] = True
+
+        outputs = self.language_model.generate(
+            # input_ids=origin_input_ids,
+            inputs_embeds=input_embeds,
+            attention_mask=attention_mask,
+            generation_config=generation_config,
+            output_hidden_states=output_hidden_states,
+            # return_dict=return_dict,
+            use_cache=True,
+            **generate_kwargs,
+        )
+
+        return outputs
+
+
+
+    @torch.no_grad()
+    def generate_output(self,messages):
+        input_ids, pixel_values = self.process_messages(messages)
+        do_sample = False if self.temperature == 0 else True
+        generated_ids = self.generate(pixel_values=pixel_values, input_ids=input_ids, temperature=self.temperature,top_p=self.top_p,repetition_penalty=self.repetition_penalty,max_new_tokens=self.max_new_tokens,do_sample = do_sample)
+        generated_ids_trimmed = generated_ids
+        output_text = self.tokenizer.batch_decode(
+            generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
+        )
+        return output_text[0]
+    
+    def generate_outputs(self,messages_list):
+        res = []
+        for messages in messages_list:
+            result = self.generate_output(messages)
+            res.append(result)
+        return res
+
+
+def tokenizer_image_token(prompt, tokenizer, image_token_index=IMAGE_TOKEN_INDEX, return_tensors=None):
+    prompt_chunks = [tokenizer(chunk).input_ids for chunk in prompt.split('<image>')]
+
+    def insert_separator(X, sep):
+        return [ele for sublist in zip(X, [sep]*len(X)) for ele in sublist][:-1]
+
+    input_ids = []
+    labels = []
+    offset = 0
+    if len(prompt_chunks) > 0 and len(prompt_chunks[0]) > 0 and prompt_chunks[0][0] == tokenizer.bos_token_id:
+        offset = 1
+        input_ids.append(prompt_chunks[0][0])
+        labels.append(prompt_chunks[0][0])
+
+    for x in insert_separator(prompt_chunks, [image_token_index] * (offset + 1)):
+        input_ids.extend(x[offset:])
+    
+    for x in insert_separator(prompt_chunks, [IGNORE_INDEX] * (offset + 1)):
+        labels.extend(x[offset:])
+
+    if return_tensors is not None:
+        if return_tensors == 'pt':
+            return torch.tensor(input_ids, dtype=torch.long), torch.tensor(labels, dtype=torch.long)
+        raise ValueError(f'Unsupported tensor type: {return_tensors}')
+    return input_ids, labels

added_tokens.json

@@ -0,0 +1,30 @@
+{
+  "</think>": 151668,
+  "</tool_call>": 151658,
+  "</tool_response>": 151666,
+  "<think>": 151667,
+  "<tool_call>": 151657,
+  "<tool_response>": 151665,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endofchunk|>": 151669,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|pad|>": 151670,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- endif %}
+{%- endif %}

config.json

@@ -0,0 +1,70 @@
+{
+  "architectures": [
+    "InfiMed"
+  ],
+  "llm_config": {
+    "_name_or_path": "/lustre/projects/polyullm/models/Qwen3/Qwen3-4B",
+    "architectures": [
+      "Qwen3ForCausalLM"
+    ],
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "eos_token_id": 151645,
+    "head_dim": 128,
+    "hidden_act": "silu",
+    "hidden_size": 2560,
+    "initializer_range": 0.02,
+    "intermediate_size": 9728,
+    "max_position_embeddings": 40960,
+    "max_window_layers": 36,
+    "model_type": "qwen3",
+    "num_attention_heads": 32,
+    "num_hidden_layers": 36,
+    "num_key_value_heads": 8,
+    "rms_norm_eps": 1e-06,
+    "rope_scaling": null,
+    "rope_theta": 1000000,
+    "sliding_window": null,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_cache": true,
+    "use_sliding_window": false,
+    "vocab_size": 151936
+  },
+  "load_precision": "bf16",
+  "max_length": 32,
+  "repetition_penalty": 1.2,
+  "run_dir": "qwen3-4b-instruct-continue-training+stage-finetune+x7",
+  "seed": 7,
+  "stage": "finetune",
+  "temperature": 0.0,
+  "top_p": 0.0001,
+  "torch_dtype": "bfloat16",
+  "trackers": [
+    "jsonl",
+    "wandb"
+  ],
+  "transformers_version": "4.52.4",
+  "vision_config": {
+    "_name_or_path": "/home/projects/polyullm/guanghao/train_code/siglip-so400m-patch14-384",
+    "architectures": [
+      "SiglipModel"
+    ],
+    "attention_dropout": 0.0,
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "image_size": 384,
+    "initializer_factor": 1.0,
+    "intermediate_size": 4304,
+    "layer_norm_eps": 1e-06,
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 16,
+    "num_channels": 3,
+    "num_hidden_layers": 27,
+    "patch_size": 14,
+    "torch_dtype": "bfloat16"
+  },
+  "wandb_entity": null,
+  "wandb_project": "mmpretrain"
+}

configuration_qwen3.py

@@ -0,0 +1,212 @@
+# coding=utf-8
+# Copyright 2024 The Qwen team, Alibaba Group and the HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Qwen3 model configuration"""
+
+from transformers.configuration_utils import PretrainedConfig
+from transformers.modeling_rope_utils import rope_config_validation
+from transformers.utils import logging
+
+
+logger = logging.get_logger(__name__)
+
+
+class Qwen3Config(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`Qwen3Model`]. It is used to instantiate a
+    Qwen3 model according to the specified arguments, defining the model architecture. Instantiating a configuration
+    with the defaults will yield a similar configuration to that of
+    Qwen3-8B [Qwen/Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B).
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+
+    Args:
+        vocab_size (`int`, *optional*, defaults to 151936):
+            Vocabulary size of the Qwen3 model. Defines the number of different tokens that can be represented by the
+            `inputs_ids` passed when calling [`Qwen3Model`]
+        hidden_size (`int`, *optional*, defaults to 4096):
+            Dimension of the hidden representations.
+        intermediate_size (`int`, *optional*, defaults to 22016):
+            Dimension of the MLP representations.
+        num_hidden_layers (`int`, *optional*, defaults to 32):
+            Number of hidden layers in the Transformer encoder.
+        num_attention_heads (`int`, *optional*, defaults to 32):
+            Number of attention heads for each attention layer in the Transformer encoder.
+        num_key_value_heads (`int`, *optional*, defaults to 32):
+            This is the number of key_value heads that should be used to implement Grouped Query Attention. If
+            `num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
+            `num_key_value_heads=1` the model will use Multi Query Attention (MQA) otherwise GQA is used. When
+            converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
+            by meanpooling all the original heads within that group. For more details checkout [this
+            paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to `32`.
+        head_dim (`int`, *optional*, defaults to 128):
+            The attention head dimension.
+        hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
+            The non-linear activation function (function or string) in the decoder.
+        max_position_embeddings (`int`, *optional*, defaults to 32768):
+            The maximum sequence length that this model might ever be used with.
+        initializer_range (`float`, *optional*, defaults to 0.02):
+            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
+        rms_norm_eps (`float`, *optional*, defaults to 1e-06):
+            The epsilon used by the rms normalization layers.
+        use_cache (`bool`, *optional*, defaults to `True`):
+            Whether or not the model should return the last key/values attentions (not used by all models). Only
+            relevant if `config.is_decoder=True`.
+        tie_word_embeddings (`bool`, *optional*, defaults to `False`):
+            Whether the model's input and output word embeddings should be tied.
+        rope_theta (`float`, *optional*, defaults to 10000.0):
+            The base period of the RoPE embeddings.
+        rope_scaling (`Dict`, *optional*):
+            Dictionary containing the scaling configuration for the RoPE embeddings. NOTE: if you apply new rope type
+            and you expect the model to work on longer `max_position_embeddings`, we recommend you to update this value
+            accordingly.
+            Expected contents:
+                `rope_type` (`str`):
+                    The sub-variant of RoPE to use. Can be one of ['default', 'linear', 'dynamic', 'yarn', 'longrope',
+                    'llama3'], with 'default' being the original RoPE implementation.
+                `factor` (`float`, *optional*):
+                    Used with all rope types except 'default'. The scaling factor to apply to the RoPE embeddings. In
+                    most scaling types, a `factor` of x will enable the model to handle sequences of length x *
+                    original maximum pre-trained length.
+                `original_max_position_embeddings` (`int`, *optional*):
+                    Used with 'dynamic', 'longrope' and 'llama3'. The original max position embeddings used during
+                    pretraining.
+                `attention_factor` (`float`, *optional*):
+                    Used with 'yarn' and 'longrope'. The scaling factor to be applied on the attention
+                    computation. If unspecified, it defaults to value recommended by the implementation, using the
+                    `factor` field to infer the suggested value.
+                `beta_fast` (`float`, *optional*):
+                    Only used with 'yarn'. Parameter to set the boundary for extrapolation (only) in the linear
+                    ramp function. If unspecified, it defaults to 32.
+                `beta_slow` (`float`, *optional*):
+                    Only used with 'yarn'. Parameter to set the boundary for interpolation (only) in the linear
+                    ramp function. If unspecified, it defaults to 1.
+                `short_factor` (`List[float]`, *optional*):
+                    Only used with 'longrope'. The scaling factor to be applied to short contexts (<
+                    `original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden
+                    size divided by the number of attention heads divided by 2
+                `long_factor` (`List[float]`, *optional*):
+                    Only used with 'longrope'. The scaling factor to be applied to long contexts (<
+                    `original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden
+                    size divided by the number of attention heads divided by 2
+                `low_freq_factor` (`float`, *optional*):
+                    Only used with 'llama3'. Scaling factor applied to low frequency components of the RoPE
+                `high_freq_factor` (`float`, *optional*):
+                    Only used with 'llama3'. Scaling factor applied to high frequency components of the RoPE
+        attention_bias (`bool`, defaults to `False`, *optional*, defaults to `False`):
+            Whether to use a bias in the query, key, value and output projection layers during self-attention.
+        use_sliding_window (`bool`, *optional*, defaults to `False`):
+            Whether to use sliding window attention.
+        sliding_window (`int`, *optional*, defaults to 4096):
+            Sliding window attention (SWA) window size. If not specified, will default to `4096`.
+        max_window_layers (`int`, *optional*, defaults to 28):
+            The number of layers that use SWA (Sliding Window Attention). The bottom layers use SWA while the top use full attention.
+        attention_dropout (`float`, *optional*, defaults to 0.0):
+            The dropout ratio for the attention probabilities.
+
+    ```python
+    >>> from transformers import Qwen3Model, Qwen3Config
+
+    >>> # Initializing a Qwen3 style configuration
+    >>> configuration = Qwen3Config()
+
+    >>> # Initializing a model from the Qwen3-8B style configuration
+    >>> model = Qwen3Model(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "qwen3"
+    keys_to_ignore_at_inference = ["past_key_values"]
+
+    # Default tensor parallel plan for base model `Qwen3`
+    base_model_tp_plan = {
+        "layers.*.self_attn.q_proj": "colwise",
+        "layers.*.self_attn.k_proj": "colwise",
+        "layers.*.self_attn.v_proj": "colwise",
+        "layers.*.self_attn.o_proj": "rowwise",
+        "layers.*.mlp.gate_proj": "colwise",
+        "layers.*.mlp.up_proj": "colwise",
+        "layers.*.mlp.down_proj": "rowwise",
+    }
+    base_model_pp_plan = {
+        "embed_tokens": (["input_ids"], ["inputs_embeds"]),
+        "layers": (["hidden_states", "attention_mask"], ["hidden_states"]),
+        "norm": (["hidden_states"], ["hidden_states"]),
+    }
+
+    def __init__(
+        self,
+        vocab_size=151936,
+        hidden_size=4096,
+        intermediate_size=22016,
+        num_hidden_layers=32,
+        num_attention_heads=32,
+        num_key_value_heads=32,
+        head_dim=128,
+        hidden_act="silu",
+        max_position_embeddings=32768,
+        initializer_range=0.02,
+        rms_norm_eps=1e-6,
+        use_cache=True,
+        tie_word_embeddings=False,
+        rope_theta=10000.0,
+        rope_scaling=None,
+        attention_bias=False,
+        use_sliding_window=False,
+        sliding_window=4096,
+        max_window_layers=28,
+        attention_dropout=0.0,
+        **kwargs,
+    ):
+        self.vocab_size = vocab_size
+        self.max_position_embeddings = max_position_embeddings
+        self.hidden_size = hidden_size
+        self.intermediate_size = intermediate_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.use_sliding_window = use_sliding_window
+        self.sliding_window = sliding_window  # we check `use_sliding_window` in the modeling code
+        self.max_window_layers = max_window_layers
+
+        # for backward compatibility
+        if num_key_value_heads is None:
+            num_key_value_heads = num_attention_heads
+
+        self.num_key_value_heads = num_key_value_heads
+        self.head_dim = head_dim
+        self.hidden_act = hidden_act
+        self.initializer_range = initializer_range
+        self.rms_norm_eps = rms_norm_eps
+        self.use_cache = use_cache
+        self.rope_theta = rope_theta
+        self.rope_scaling = rope_scaling
+        self.attention_bias = attention_bias
+        self.attention_dropout = attention_dropout
+        # Validate the correctness of rotary position embeddings parameters
+        # BC: if there is a 'type' field, move it to 'rope_type'.
+        if self.rope_scaling is not None and "type" in self.rope_scaling:
+            self.rope_scaling["rope_type"] = self.rope_scaling["type"]
+        rope_config_validation(self)
+
+        super().__init__(
+            tie_word_embeddings=tie_word_embeddings,
+            **kwargs,
+        )
+
+
+__all__ = ["Qwen3Config"]

configuration_siglip.py

@@ -0,0 +1,265 @@
+# coding=utf-8
+# Copyright 2024 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Siglip model configuration"""
+
+from transformers.configuration_utils import PretrainedConfig
+from transformers.utils import logging
+
+
+logger = logging.get_logger(__name__)
+
+
+class SiglipTextConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`SiglipTextModel`]. It is used to instantiate a
+    Siglip text encoder according to the specified arguments, defining the model architecture. Instantiating a
+    configuration with the defaults will yield a similar configuration to that of the text encoder of the Siglip
+    [google/siglip-base-patch16-224](https://huggingface.co/google/siglip-base-patch16-224) architecture.
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+    Args:
+        vocab_size (`int`, *optional*, defaults to 32000):
+            Vocabulary size of the Siglip text model. Defines the number of different tokens that can be represented by
+            the `inputs_ids` passed when calling [`SiglipModel`].
+        hidden_size (`int`, *optional*, defaults to 768):
+            Dimensionality of the encoder layers and the pooler layer.
+        intermediate_size (`int`, *optional*, defaults to 3072):
+            Dimensionality of the "intermediate" (i.e., feed-forward) layer in the Transformer encoder.
+        num_hidden_layers (`int`, *optional*, defaults to 12):
+            Number of hidden layers in the Transformer encoder.
+        num_attention_heads (`int`, *optional*, defaults to 12):
+            Number of attention heads for each attention layer in the Transformer encoder.
+        max_position_embeddings (`int`, *optional*, defaults to 64):
+            The maximum sequence length that this model might ever be used with. Typically set this to something large
+            just in case (e.g., 512 or 1024 or 2048).
+        hidden_act (`str` or `function`, *optional*, defaults to `"gelu_pytorch_tanh"`):
+            The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`,
+            `"relu"`, `"selu"` and `"gelu_new"` `"quick_gelu"` are supported.
+        layer_norm_eps (`float`, *optional*, defaults to 1e-06):
+            The epsilon used by the layer normalization layers.
+        attention_dropout (`float`, *optional*, defaults to 0.0):
+            The dropout ratio for the attention probabilities.
+        pad_token_id (`int`, *optional*, defaults to 1):
+            The id of the padding token in the vocabulary.
+        bos_token_id (`int`, *optional*, defaults to 49406):
+            The id of the beginning-of-sequence token in the vocabulary.
+        eos_token_id (`int`, *optional*, defaults to 49407):
+            The id of the end-of-sequence token in the vocabulary.
+
+    Example:
+
+    ```python
+    >>> from transformers import SiglipTextConfig, SiglipTextModel
+
+    >>> # Initializing a SiglipTextConfig with google/siglip-base-patch16-224 style configuration
+    >>> configuration = SiglipTextConfig()
+
+    >>> # Initializing a SiglipTextModel (with random weights) from the google/siglip-base-patch16-224 style configuration
+    >>> model = SiglipTextModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip_text_model"
+    base_config_key = "text_config"
+
+    def __init__(
+        self,
+        vocab_size=32000,
+        hidden_size=768,
+        intermediate_size=3072,
+        num_hidden_layers=12,
+        num_attention_heads=12,
+        max_position_embeddings=64,
+        hidden_act="gelu_pytorch_tanh",
+        layer_norm_eps=1e-6,
+        attention_dropout=0.0,
+        # This differs from `CLIPTokenizer`'s default and from openai/siglip
+        # See https://github.com/huggingface/transformers/pull/24773#issuecomment-1632287538
+        pad_token_id=1,
+        bos_token_id=49406,
+        eos_token_id=49407,
+        **kwargs,
+    ):
+        super().__init__(pad_token_id=pad_token_id, bos_token_id=bos_token_id, eos_token_id=eos_token_id, **kwargs)
+
+        self.vocab_size = vocab_size
+        self.hidden_size = hidden_size
+        self.intermediate_size = intermediate_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.max_position_embeddings = max_position_embeddings
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+        self.attention_dropout = attention_dropout
+
+
+class SiglipVisionConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`SiglipVisionModel`]. It is used to instantiate a
+    Siglip vision encoder according to the specified arguments, defining the model architecture. Instantiating a
+    configuration with the defaults will yield a similar configuration to that of the vision encoder of the Siglip
+    [google/siglip-base-patch16-224](https://huggingface.co/google/siglip-base-patch16-224) architecture.
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+    Args:
+        hidden_size (`int`, *optional*, defaults to 768):
+            Dimensionality of the encoder layers and the pooler layer.
+        intermediate_size (`int`, *optional*, defaults to 3072):
+            Dimensionality of the "intermediate" (i.e., feed-forward) layer in the Transformer encoder.
+        num_hidden_layers (`int`, *optional*, defaults to 12):
+            Number of hidden layers in the Transformer encoder.
+        num_attention_heads (`int`, *optional*, defaults to 12):
+            Number of attention heads for each attention layer in the Transformer encoder.
+        num_channels (`int`, *optional*, defaults to 3):
+            Number of channels in the input images.
+        image_size (`int`, *optional*, defaults to 224):
+            The size (resolution) of each image.
+        patch_size (`int`, *optional*, defaults to 16):
+            The size (resolution) of each patch.
+        hidden_act (`str` or `function`, *optional*, defaults to `"gelu_pytorch_tanh"`):
+            The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`,
+            `"relu"`, `"selu"` and `"gelu_new"` `"quick_gelu"` are supported.
+        layer_norm_eps (`float`, *optional*, defaults to 1e-06):
+            The epsilon used by the layer normalization layers.
+        attention_dropout (`float`, *optional*, defaults to 0.0):
+            The dropout ratio for the attention probabilities.
+
+    Example:
+
+    ```python
+    >>> from transformers import SiglipVisionConfig, SiglipVisionModel
+
+    >>> # Initializing a SiglipVisionConfig with google/siglip-base-patch16-224 style configuration
+    >>> configuration = SiglipVisionConfig()
+
+    >>> # Initializing a SiglipVisionModel (with random weights) from the google/siglip-base-patch16-224 style configuration
+    >>> model = SiglipVisionModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip_vision_model"
+    base_config_key = "vision_config"
+
+    def __init__(
+        self,
+        hidden_size=768,
+        intermediate_size=3072,
+        num_hidden_layers=12,
+        num_attention_heads=12,
+        num_channels=3,
+        image_size=224,
+        patch_size=16,
+        hidden_act="gelu_pytorch_tanh",
+        layer_norm_eps=1e-6,
+        attention_dropout=0.0,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+
+        self.hidden_size = hidden_size
+        self.intermediate_size = intermediate_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.num_channels = num_channels
+        self.patch_size = patch_size
+        self.image_size = image_size
+        self.attention_dropout = attention_dropout
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+
+
+class SiglipConfig(PretrainedConfig):
+    r"""
+    [`SiglipConfig`] is the configuration class to store the configuration of a [`SiglipModel`]. It is used to
+    instantiate a Siglip model according to the specified arguments, defining the text model and vision model configs.
+    Instantiating a configuration with the defaults will yield a similar configuration to that of the Siglip
+    [google/siglip-base-patch16-224](https://huggingface.co/google/siglip-base-patch16-224) architecture.
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+    Args:
+        text_config (`dict`, *optional*):
+            Dictionary of configuration options used to initialize [`SiglipTextConfig`].
+        vision_config (`dict`, *optional*):
+            Dictionary of configuration options used to initialize [`SiglipVisionConfig`].
+        kwargs (*optional*):
+            Dictionary of keyword arguments.
+
+    Example:
+
+    ```python
+    >>> from transformers import SiglipConfig, SiglipModel
+
+    >>> # Initializing a SiglipConfig with google/siglip-base-patch16-224 style configuration
+    >>> configuration = SiglipConfig()
+
+    >>> # Initializing a SiglipModel (with random weights) from the google/siglip-base-patch16-224 style configuration
+    >>> model = SiglipModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+
+    >>> # We can also initialize a SiglipConfig from a SiglipTextConfig and a SiglipVisionConfig
+    >>> from transformers import SiglipTextConfig, SiglipVisionConfig
+
+    >>> # Initializing a SiglipText and SiglipVision configuration
+    >>> config_text = SiglipTextConfig()
+    >>> config_vision = SiglipVisionConfig()
+
+    >>> config = SiglipConfig.from_text_vision_configs(config_text, config_vision)
+    ```"""
+
+    model_type = "siglip"
+    sub_configs = {"text_config": SiglipTextConfig, "vision_config": SiglipVisionConfig}
+
+    def __init__(self, text_config=None, vision_config=None, **kwargs):
+        super().__init__(**kwargs)
+
+        if text_config is None:
+            text_config = {}
+            logger.info("`text_config` is `None`. Initializing the `SiglipTextConfig` with default values.")
+
+        if vision_config is None:
+            vision_config = {}
+            logger.info("`vision_config` is `None`. initializing the `SiglipVisionConfig` with default values.")
+
+        self.text_config = SiglipTextConfig(**text_config)
+        self.vision_config = SiglipVisionConfig(**vision_config)
+
+        self.initializer_factor = 1.0
+
+    @classmethod
+    def from_text_vision_configs(cls, text_config: SiglipTextConfig, vision_config: SiglipVisionConfig, **kwargs):
+        r"""
+        Instantiate a [`SiglipConfig`] (or a derived class) from siglip text model configuration and siglip vision
+        model configuration.
+
+        Returns:
+            [`SiglipConfig`]: An instance of a configuration object
+        """
+
+        return cls(text_config=text_config.to_dict(), vision_config=vision_config.to_dict(), **kwargs)
+
+
+__all__ = ["SiglipConfig", "SiglipTextConfig", "SiglipVisionConfig"]

model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1031cef38e5bc62149fd5d1b68bd67d00d7d723a82180c97fd17090c70e63449
+size 4966471968

model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3b8d072383149ce00cc8183badf0f5abf42ee7fa2e4a795dcfed5208e18b438
+size 4731957576

model.safetensors.index.json

@@ -0,0 +1,858 @@
+{
+  "metadata": {
+    "total_size": 9698315392
+  },
+  "weight_map": {
+    "language_model.lm_head.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.16.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.16.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.16.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.16.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.16.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.16.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.17.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.17.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.17.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.17.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.17.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.17.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.17.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.17.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.17.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.17.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.17.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.18.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.19.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.2.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.2.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.20.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.20.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.20.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.20.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.20.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.20.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.20.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.20.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.28.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.29.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.3.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.3.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.30.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.30.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.30.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.30.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.30.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.30.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.30.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.30.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.30.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.30.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.30.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.31.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.32.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.33.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.34.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.35.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "language_model.model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.4.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.4.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "language_model.model.norm.weight": "model-00002-of-00002.safetensors",
+    "projector.mlp_projector.0.bias": "model-00002-of-00002.safetensors",
+    "projector.mlp_projector.0.weight": "model-00002-of-00002.safetensors",
+    "projector.mlp_projector.2.bias": "model-00002-of-00002.safetensors",
+    "projector.mlp_projector.2.weight": "model-00002-of-00002.safetensors",
+    "vision_model.vision_model.embeddings.patch_embedding.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.embeddings.patch_embedding.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.embeddings.position_embedding.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.21.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.22.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.23.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.24.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.25.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.26.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.layer_norm1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.layer_norm1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.layer_norm2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.layer_norm2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.self_attn.k_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.self_attn.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.self_attn.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.self_attn.q_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.self_attn.v_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.encoder.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.attention.in_proj_bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.attention.in_proj_weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.attention.out_proj.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.attention.out_proj.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.layernorm.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.layernorm.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.head.probe": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.post_layernorm.bias": "model-00001-of-00002.safetensors",
+    "vision_model.vision_model.post_layernorm.weight": "model-00001-of-00002.safetensors"
+  }
+}

modeling_siglip.py

@@ -0,0 +1,1585 @@
+# coding=utf-8
+# Copyright 2024 Google AI and The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""PyTorch Siglip model."""
+
+import math
+import warnings
+from dataclasses import dataclass
+from typing import Any, Optional, Tuple, Union
+
+import numpy as np
+import torch
+import torch.utils.checkpoint
+from torch import nn
+from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
+from torch.nn.init import _calculate_fan_in_and_fan_out
+
+from transformers.activations import ACT2FN
+from transformers.modeling_attn_mask_utils import _prepare_4d_attention_mask
+from transformers.modeling_outputs import BaseModelOutput, BaseModelOutputWithPooling, ImageClassifierOutput
+from transformers.modeling_utils import PreTrainedModel
+from transformers.utils import (
+    ModelOutput,
+    add_start_docstrings,
+    add_start_docstrings_to_model_forward,
+    is_flash_attn_2_available,
+    is_flash_attn_greater_or_equal_2_10,
+    logging,
+    replace_return_docstrings,
+    torch_int,
+)
+from configuration_siglip import SiglipConfig, SiglipTextConfig, SiglipVisionConfig
+
+
+if is_flash_attn_2_available():
+    from transformers.modeling_flash_attention_utils import _flash_attention_forward
+
+
+logger = logging.get_logger(__name__)
+
+# General docstring
+_CONFIG_FOR_DOC = "SiglipConfig"
+_CHECKPOINT_FOR_DOC = "google/siglip-base-patch16-224"
+
+
+def _trunc_normal_(tensor, mean, std, a, b):
+    # Cut & paste from PyTorch official master until it's in a few official releases - RW
+    # Method based on https://people.sc.fsu.edu/~jburkardt/presentations/truncated_normal.pdf
+    def norm_cdf(x):
+        # Computes standard normal cumulative distribution function
+        return (1.0 + math.erf(x / math.sqrt(2.0))) / 2.0
+
+    if (mean < a - 2 * std) or (mean > b + 2 * std):
+        warnings.warn(
+            "mean is more than 2 std from [a, b] in nn.init.trunc_normal_. "
+            "The distribution of values may be incorrect.",
+            stacklevel=2,
+        )
+
+    # Values are generated by using a truncated uniform distribution and
+    # then using the inverse CDF for the normal distribution.
+    # Get upper and lower cdf values
+    l = norm_cdf((a - mean) / std)
+    u = norm_cdf((b - mean) / std)
+
+    # Uniformly fill tensor with values from [l, u], then translate to
+    # [2l-1, 2u-1].
+    tensor.uniform_(2 * l - 1, 2 * u - 1)
+
+    # Use inverse cdf transform for normal distribution to get truncated
+    # standard normal
+    tensor.erfinv_()
+
+    # Transform to proper mean, std
+    tensor.mul_(std * math.sqrt(2.0))
+    tensor.add_(mean)
+
+    # Clamp to ensure it's in the proper range
+    tensor.clamp_(min=a, max=b)
+
+
+def trunc_normal_tf_(
+    tensor: torch.Tensor, mean: float = 0.0, std: float = 1.0, a: float = -2.0, b: float = 2.0
+) -> torch.Tensor:
+    """Fills the input Tensor with values drawn from a truncated
+    normal distribution. The values are effectively drawn from the
+    normal distribution :math:`\\mathcal{N}(\text{mean}, \text{std}^2)`
+    with values outside :math:`[a, b]` redrawn until they are within
+    the bounds. The method used for generating the random values works
+    best when :math:`a \\leq \text{mean} \\leq b`.
+
+    NOTE: this 'tf' variant behaves closer to Tensorflow / JAX impl where the
+    bounds [a, b] are applied when sampling the normal distribution with mean=0, std=1.0
+    and the result is subsequently scaled and shifted by the mean and std args.
+
+    Args:
+        tensor: an n-dimensional `torch.Tensor`
+        mean: the mean of the normal distribution
+        std: the standard deviation of the normal distribution
+        a: the minimum cutoff value
+        b: the maximum cutoff value
+    """
+    with torch.no_grad():
+        _trunc_normal_(tensor, 0, 1.0, a, b)
+        tensor.mul_(std).add_(mean)
+
+
+def variance_scaling_(tensor, scale=1.0, mode="fan_in", distribution="normal"):
+    fan_in, fan_out = _calculate_fan_in_and_fan_out(tensor)
+    if mode == "fan_in":
+        denom = fan_in
+    elif mode == "fan_out":
+        denom = fan_out
+    elif mode == "fan_avg":
+        denom = (fan_in + fan_out) / 2
+
+    variance = scale / denom
+
+    if distribution == "truncated_normal":
+        # constant is stddev of standard normal truncated to (-2, 2)
+        trunc_normal_tf_(tensor, std=math.sqrt(variance) / 0.87962566103423978)
+    elif distribution == "normal":
+        with torch.no_grad():
+            tensor.normal_(std=math.sqrt(variance))
+    elif distribution == "uniform":
+        bound = math.sqrt(3 * variance)
+        with torch.no_grad():
+            tensor.uniform_(-bound, bound)
+    else:
+        raise ValueError(f"invalid distribution {distribution}")
+
+
+def lecun_normal_(tensor):
+    variance_scaling_(tensor, mode="fan_in", distribution="truncated_normal")
+
+
+def default_flax_embed_init(tensor):
+    variance_scaling_(tensor, mode="fan_in", distribution="normal")
+
+
+@dataclass
+# Copied from transformers.models.clip.modeling_clip.CLIPVisionModelOutput with CLIP->Siglip
+class SiglipVisionModelOutput(ModelOutput):
+    """
+    Base class for vision model's outputs that also contains image embeddings of the pooling of the last hidden states.
+
+    Args:
+        image_embeds (`torch.FloatTensor` of shape `(batch_size, output_dim)` *optional* returned when model is initialized with `with_projection=True`):
+            The image embeddings obtained by applying the projection layer to the pooler_output.
+        last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
+            Sequence of hidden-states at the output of the last layer of the model.
+        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
+            Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
+            one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.
+
+            Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
+        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
+            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
+            sequence_length)`.
+
+            Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
+            heads.
+    """
+
+    image_embeds: Optional[torch.FloatTensor] = None
+    last_hidden_state: torch.FloatTensor = None
+    hidden_states: Optional[Tuple[torch.FloatTensor, ...]] = None
+    attentions: Optional[Tuple[torch.FloatTensor, ...]] = None
+
+
+@dataclass
+# Copied from transformers.models.clip.modeling_clip.CLIPTextModelOutput with CLIP->Siglip
+class SiglipTextModelOutput(ModelOutput):
+    """
+    Base class for text model's outputs that also contains a pooling of the last hidden states.
+
+    Args:
+        text_embeds (`torch.FloatTensor` of shape `(batch_size, output_dim)` *optional* returned when model is initialized with `with_projection=True`):
+            The text embeddings obtained by applying the projection layer to the pooler_output.
+        last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
+            Sequence of hidden-states at the output of the last layer of the model.
+        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
+            Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
+            one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.
+
+            Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
+        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
+            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
+            sequence_length)`.
+
+            Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
+            heads.
+    """
+
+    text_embeds: Optional[torch.FloatTensor] = None
+    last_hidden_state: torch.FloatTensor = None
+    hidden_states: Optional[Tuple[torch.FloatTensor, ...]] = None
+    attentions: Optional[Tuple[torch.FloatTensor, ...]] = None
+
+
+@dataclass
+# Copied from transformers.models.clip.modeling_clip.CLIPOutput with CLIP->Siglip
+class SiglipOutput(ModelOutput):
+    """
+    Args:
+        loss (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `return_loss` is `True`):
+            Contrastive loss for image-text similarity.
+        logits_per_image (`torch.FloatTensor` of shape `(image_batch_size, text_batch_size)`):
+            The scaled dot product scores between `image_embeds` and `text_embeds`. This represents the image-text
+            similarity scores.
+        logits_per_text (`torch.FloatTensor` of shape `(text_batch_size, image_batch_size)`):
+            The scaled dot product scores between `text_embeds` and `image_embeds`. This represents the text-image
+            similarity scores.
+        text_embeds (`torch.FloatTensor` of shape `(batch_size, output_dim`):
+            The text embeddings obtained by applying the projection layer to the pooled output of [`SiglipTextModel`].
+        image_embeds (`torch.FloatTensor` of shape `(batch_size, output_dim`):
+            The image embeddings obtained by applying the projection layer to the pooled output of [`SiglipVisionModel`].
+        text_model_output (`BaseModelOutputWithPooling`):
+            The output of the [`SiglipTextModel`].
+        vision_model_output (`BaseModelOutputWithPooling`):
+            The output of the [`SiglipVisionModel`].
+    """
+
+    loss: Optional[torch.FloatTensor] = None
+    logits_per_image: torch.FloatTensor = None
+    logits_per_text: torch.FloatTensor = None
+    text_embeds: torch.FloatTensor = None
+    image_embeds: torch.FloatTensor = None
+    text_model_output: BaseModelOutputWithPooling = None
+    vision_model_output: BaseModelOutputWithPooling = None
+
+    def to_tuple(self) -> Tuple[Any]:
+        return tuple(
+            self[k] if k not in ["text_model_output", "vision_model_output"] else getattr(self, k).to_tuple()
+            for k in self.keys()
+        )
+
+
+class SiglipVisionEmbeddings(nn.Module):
+    def __init__(self, config: SiglipVisionConfig):
+        super().__init__()
+        self.config = config
+        self.embed_dim = config.hidden_size
+        self.image_size = config.image_size
+        self.patch_size = config.patch_size
+
+        self.patch_embedding = nn.Conv2d(
+            in_channels=config.num_channels,
+            out_channels=self.embed_dim,
+            kernel_size=self.patch_size,
+            stride=self.patch_size,
+            padding="valid",
+        )
+
+        self.num_patches = (self.image_size // self.patch_size) ** 2
+        self.num_positions = self.num_patches
+        self.position_embedding = nn.Embedding(self.num_positions, self.embed_dim)
+        self.register_buffer("position_ids", torch.arange(self.num_positions).expand((1, -1)), persistent=False)
+
+    def interpolate_pos_encoding(self, embeddings: torch.Tensor, height: int, width: int) -> torch.Tensor:
+        """
+        This method allows to interpolate the pre-trained position encodings, to be able to use the model on higher resolution
+        images. This method is also adapted to support torch.jit tracing and no class embeddings.
+
+        Adapted from:
+        - https://github.com/facebookresearch/dino/blob/de9ee3df6cf39fac952ab558447af1fa1365362a/vision_transformer.py#L174-L194, and
+        - https://github.com/facebookresearch/dinov2/blob/e1277af2ba9496fbadf7aec6eba56e8d882d1e35/dinov2/models/vision_transformer.py#L179-L211
+        """
+
+        num_patches = embeddings.shape[1]
+        num_positions = self.position_embedding.weight.shape[0]
+
+        # always interpolate when tracing to ensure the exported model works for dynamic input shapes
+        if not torch.jit.is_tracing() and num_patches == num_positions and height == width:
+            return self.position_embedding(self.position_ids)
+
+        patch_pos_embed = self.position_embedding.weight.unsqueeze(0)
+
+        dim = embeddings.shape[-1]
+
+        new_height = height // self.patch_size
+        new_width = width // self.patch_size
+
+        sqrt_num_positions = torch_int(num_positions**0.5)
+        patch_pos_embed = patch_pos_embed.reshape(1, sqrt_num_positions, sqrt_num_positions, dim)
+        patch_pos_embed = patch_pos_embed.permute(0, 3, 1, 2)
+
+        patch_pos_embed = nn.functional.interpolate(
+            patch_pos_embed,
+            size=(new_height, new_width),
+            mode="bicubic",
+            align_corners=False,
+        )
+
+        patch_pos_embed = patch_pos_embed.permute(0, 2, 3, 1).view(1, -1, dim)
+        return patch_pos_embed
+
+    def forward(self, pixel_values: torch.FloatTensor, interpolate_pos_encoding=False) -> torch.Tensor:
+        _, _, height, width = pixel_values.shape
+        target_dtype = self.patch_embedding.weight.dtype
+        # import pdb;pdb.set_trace()
+        patch_embeds = self.patch_embedding(pixel_values.to(dtype=target_dtype))  # shape = [*, width, grid, grid]
+        embeddings = patch_embeds.flatten(2).transpose(1, 2)
+
+        if interpolate_pos_encoding:
+            embeddings = embeddings + self.interpolate_pos_encoding(embeddings, height, width)
+        else:
+            embeddings = embeddings + self.position_embedding(self.position_ids)
+        return embeddings
+
+
+# Copied from transformers.models.clip.modeling_clip.CLIPTextEmbeddings with CLIP->Siglip
+class SiglipTextEmbeddings(nn.Module):
+    def __init__(self, config: SiglipTextConfig):
+        super().__init__()
+        embed_dim = config.hidden_size
+
+        self.token_embedding = nn.Embedding(config.vocab_size, embed_dim)
+        self.position_embedding = nn.Embedding(config.max_position_embeddings, embed_dim)
+
+        # position_ids (1, len position emb) is contiguous in memory and exported when serialized
+        self.register_buffer(
+            "position_ids", torch.arange(config.max_position_embeddings).expand((1, -1)), persistent=False
+        )
+
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+    ) -> torch.Tensor:
+        seq_length = input_ids.shape[-1] if input_ids is not None else inputs_embeds.shape[-2]
+        max_position_embedding = self.position_embedding.weight.shape[0]
+
+        if seq_length > max_position_embedding:
+            raise ValueError(
+                f"Sequence length must be less than max_position_embeddings (got `sequence length`: "
+                f"{seq_length} and max_position_embeddings: {max_position_embedding}"
+            )
+
+        if position_ids is None:
+            position_ids = self.position_ids[:, :seq_length]
+
+        if inputs_embeds is None:
+            inputs_embeds = self.token_embedding(input_ids)
+
+        position_embeddings = self.position_embedding(position_ids)
+        embeddings = inputs_embeds + position_embeddings
+
+        return embeddings
+
+
+class SiglipAttention(nn.Module):
+    """Multi-headed attention from 'Attention Is All You Need' paper"""
+
+    # Copied from transformers.models.clip.modeling_clip.CLIPAttention.__init__
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.embed_dim = config.hidden_size
+        self.num_heads = config.num_attention_heads
+        self.head_dim = self.embed_dim // self.num_heads
+        if self.head_dim * self.num_heads != self.embed_dim:
+            raise ValueError(
+                f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim} and `num_heads`:"
+                f" {self.num_heads})."
+            )
+        self.scale = self.head_dim**-0.5
+        self.dropout = config.attention_dropout
+
+        self.k_proj = nn.Linear(self.embed_dim, self.embed_dim)
+        self.v_proj = nn.Linear(self.embed_dim, self.embed_dim)
+        self.q_proj = nn.Linear(self.embed_dim, self.embed_dim)
+        self.out_proj = nn.Linear(self.embed_dim, self.embed_dim)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = False,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor]]:
+        """Input shape: Batch x Time x Channel"""
+
+        batch_size, q_len, _ = hidden_states.size()
+
+        query_states = self.q_proj(hidden_states)
+        key_states = self.k_proj(hidden_states)
+        value_states = self.v_proj(hidden_states)
+
+        query_states = query_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        key_states = key_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        value_states = value_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+
+        k_v_seq_len = key_states.shape[-2]
+        attn_weights = torch.matmul(query_states, key_states.transpose(2, 3)) * self.scale
+
+        if attn_weights.size() != (batch_size, self.num_heads, q_len, k_v_seq_len):
+            raise ValueError(
+                f"Attention weights should be of size {(batch_size, self.num_heads, q_len, k_v_seq_len)}, but is"
+                f" {attn_weights.size()}"
+            )
+
+        if attention_mask is not None:
+            if attention_mask.size() != (batch_size, 1, q_len, k_v_seq_len):
+                raise ValueError(
+                    f"Attention mask should be of size {(batch_size, 1, q_len, k_v_seq_len)}, but is {attention_mask.size()}"
+                )
+            attn_weights = attn_weights + attention_mask
+
+        # upcast attention to fp32
+        attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_states.dtype)
+        attn_weights = nn.functional.dropout(attn_weights, p=self.dropout, training=self.training)
+        attn_output = torch.matmul(attn_weights, value_states)
+
+        if attn_output.size() != (batch_size, self.num_heads, q_len, self.head_dim):
+            raise ValueError(
+                f"`attn_output` should be of size {(batch_size, self.num_heads, q_len, self.head_dim)}, but is"
+                f" {attn_output.size()}"
+            )
+
+        attn_output = attn_output.transpose(1, 2).contiguous()
+        attn_output = attn_output.reshape(batch_size, q_len, self.embed_dim)
+
+        attn_output = self.out_proj(attn_output)
+
+        return attn_output, attn_weights
+
+
+class SiglipFlashAttention2(SiglipAttention):
+    """
+    SiglipAttention flash attention module. This module inherits from `SiglipAttention` as the weights of the module stays
+    untouched. The only required change would be on the forward pass where it needs to correctly call the public API of
+    flash attention and deal with padding tokens in case the input contains any of them.
+    """
+
+    is_causal = False
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+
+        # TODO: Should be removed once Flash Attention for RoCm is bumped to 2.1.
+        # flash_attn<2.1 generates top-left aligned causal mask, while what is needed here is bottom-right alignement, that was made default for flash_attn>=2.1. This attribute is used to handle this difference. Reference: https://github.com/Dao-AILab/flash-attention/releases/tag/v2.1.0.
+        # Beware that with flash_attn<2.1, using q_seqlen != k_seqlen (except for the case q_seqlen == 1) produces a wrong mask (top-left).
+        self._flash_attn_uses_top_left_mask = not is_flash_attn_greater_or_equal_2_10()
+
+    # Adapted from transformers.models.llama.modeling_llama.LlamaFlashAttention2.forward
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.LongTensor] = None,
+        output_attentions: bool = False,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        output_attentions = False
+
+        batch_size, q_len, _ = hidden_states.size()
+
+        query_states = self.q_proj(hidden_states)
+        key_states = self.k_proj(hidden_states)
+        value_states = self.v_proj(hidden_states)
+
+        # Flash attention requires the input to have the shape
+        # batch_size x seq_length x head_dim x hidden_dim
+        # therefore we just need to keep the original shape
+        query_states = query_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        key_states = key_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        value_states = value_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+
+        # TODO: These transpose are quite inefficient but Flash Attention requires the layout [batch_size, sequence_length, num_heads, head_dim]. We would need to refactor the KV cache
+        # to be able to avoid many of these transpose/reshape/view.
+        query_states = query_states.transpose(1, 2)
+        key_states = key_states.transpose(1, 2)
+        value_states = value_states.transpose(1, 2)
+
+        dropout_rate = self.dropout if self.training else 0.0
+
+        # In PEFT, usually we cast the layer norms in float32 for training stability reasons
+        # therefore the input hidden states gets silently casted in float32. Hence, we need
+        # cast them back in the correct dtype just to be sure everything works as expected.
+        # This might slowdown training & inference so it is recommended to not cast the LayerNorms
+        # in fp32.
+
+        input_dtype = query_states.dtype
+        if input_dtype == torch.float32:
+            if torch.is_autocast_enabled():
+                target_dtype = torch.get_autocast_gpu_dtype()
+            # Handle the case where the model is quantized
+            elif hasattr(self.config, "_pre_quantization_dtype"):
+                target_dtype = self.config._pre_quantization_dtype
+            else:
+                target_dtype = self.q_proj.weight.dtype
+
+            logger.warning_once(
+                f"The input hidden states seems to be silently casted in float32, this might be related to"
+                f" the fact you have upcasted embedding or layer norm layers in float32. We will cast back the input in"
+                f" {target_dtype}."
+            )
+
+            query_states = query_states.to(target_dtype)
+            key_states = key_states.to(target_dtype)
+            value_states = value_states.to(target_dtype)
+
+        attn_output = _flash_attention_forward(
+            query_states,
+            key_states,
+            value_states,
+            attention_mask,
+            q_len,
+            dropout=dropout_rate,
+            is_causal=self.is_causal,
+            use_top_left_mask=self._flash_attn_uses_top_left_mask,
+        )
+
+        attn_output = attn_output.reshape(batch_size, q_len, self.embed_dim).contiguous()
+        attn_output = self.out_proj(attn_output)
+
+        if not output_attentions:
+            attn_weights = None
+
+        return attn_output, attn_weights
+
+
+class SiglipSdpaAttention(SiglipAttention):
+    """
+    Siglip attention module using torch.nn.functional.scaled_dot_product_attention. This module inherits from
+    `SiglipAttention` as the weights of the module stays untouched. The only changes are on the forward pass to adapt to
+    SDPA API.
+    """
+
+    is_causal = False
+
+    # Adapted from SiglipAttention.forward and transformers.models.llama.modeling_llama.LlamaSdpaAttention.forward
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = False,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor]]:
+        if output_attentions:
+            # TODO: Improve this warning with e.g. `model.config.attn_implementation = "manual"` once this is implemented.
+            logger.warning_once(
+                "SiglipModel is using SiglipSdpaAttention, but `torch.nn.functional.scaled_dot_product_attention` does not support `output_attentions=True`. Falling back to the manual attention implementation, "
+                'but specifying the manual implementation will be required from Transformers version v5.0.0 onwards. This warning can be removed using the argument `attn_implementation="eager"` when loading the model.'
+            )
+            return super().forward(
+                hidden_states=hidden_states,
+                attention_mask=attention_mask,
+                output_attentions=output_attentions,
+            )
+
+        batch_size, q_len, _ = hidden_states.size()
+
+        query_states = self.q_proj(hidden_states)
+        key_states = self.k_proj(hidden_states)
+        value_states = self.v_proj(hidden_states)
+
+        query_states = query_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        key_states = key_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        value_states = value_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+
+        # SDPA with memory-efficient backend is currently (torch==2.1.2) bugged with non-contiguous inputs with custom attn_mask,
+        # Reference: https://github.com/pytorch/pytorch/issues/112577.
+        if query_states.device.type == "cuda" and attention_mask is not None:
+            query_states = query_states.contiguous()
+            key_states = key_states.contiguous()
+            value_states = value_states.contiguous()
+
+        # We dispatch to SDPA's Flash Attention or Efficient kernels via this `is_causal` if statement instead of an inline conditional assignment
+        # in SDPA to support both torch.compile's dynamic shapes and full graph options. An inline conditional prevents dynamic shapes from compiling.
+        is_causal = True if self.is_causal and q_len > 1 else False
+
+        attn_output = torch.nn.functional.scaled_dot_product_attention(
+            query_states,
+            key_states,
+            value_states,
+            attn_mask=attention_mask,
+            dropout_p=self.dropout if self.training else 0.0,
+            is_causal=is_causal,
+        )
+
+        attn_output = attn_output.transpose(1, 2).contiguous()
+        attn_output = attn_output.view(batch_size, q_len, self.embed_dim)
+
+        attn_output = self.out_proj(attn_output)
+
+        return attn_output, None
+
+
+SIGLIP_ATTENTION_CLASSES = {
+    "eager": SiglipAttention,
+    "flash_attention_2": SiglipFlashAttention2,
+    "sdpa": SiglipSdpaAttention,
+}
+
+
+# Copied from transformers.models.clip.modeling_clip.CLIPMLP with CLIP->Siglip
+class SiglipMLP(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.activation_fn = ACT2FN[config.hidden_act]
+        self.fc1 = nn.Linear(config.hidden_size, config.intermediate_size)
+        self.fc2 = nn.Linear(config.intermediate_size, config.hidden_size)
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        hidden_states = self.fc1(hidden_states)
+        hidden_states = self.activation_fn(hidden_states)
+        hidden_states = self.fc2(hidden_states)
+        return hidden_states
+
+
+class SiglipEncoderLayer(nn.Module):
+    def __init__(self, config: SiglipConfig):
+        super().__init__()
+        self.embed_dim = config.hidden_size
+        self.self_attn = SIGLIP_ATTENTION_CLASSES[config._attn_implementation](config=config)
+        self.layer_norm1 = nn.LayerNorm(self.embed_dim, eps=config.layer_norm_eps)
+        self.mlp = SiglipMLP(config)
+        self.layer_norm2 = nn.LayerNorm(self.embed_dim, eps=config.layer_norm_eps)
+
+    # Ignore copy
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: torch.Tensor,
+        output_attentions: Optional[bool] = False,
+    ) -> Tuple[torch.FloatTensor]:
+        """
+        Args:
+            hidden_states (`torch.FloatTensor`):
+                Input to the layer of shape `(batch, seq_len, embed_dim)`.
+            attention_mask (`torch.FloatTensor`):
+                Attention mask of shape `(batch, 1, q_len, k_v_seq_len)` where padding elements are indicated by very large negative values.
+            output_attentions (`bool`, *optional*, defaults to `False`):
+                Whether or not to return the attentions tensors of all attention layers. See `attentions` under
+                returned tensors for more detail.
+        """
+        residual = hidden_states
+
+        hidden_states = self.layer_norm1(hidden_states)
+        hidden_states, attn_weights = self.self_attn(
+            hidden_states=hidden_states,
+            attention_mask=attention_mask,
+            output_attentions=output_attentions,
+        )
+        hidden_states = residual + hidden_states
+
+        residual = hidden_states
+        hidden_states = self.layer_norm2(hidden_states)
+        hidden_states = self.mlp(hidden_states)
+        hidden_states = residual + hidden_states
+
+        outputs = (hidden_states,)
+
+        if output_attentions:
+            outputs += (attn_weights,)
+
+        return outputs
+
+
+class SiglipPreTrainedModel(PreTrainedModel):
+    """
+    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
+    models.
+    """
+
+    config_class = SiglipConfig
+    base_model_prefix = "siglip"
+    supports_gradient_checkpointing = True
+
+    _no_split_modules = [
+        "SiglipTextEmbeddings",
+        "SiglipEncoderLayer",
+        "SiglipVisionEmbeddings",
+        "SiglipEncoderLayer",
+        "SiglipMultiheadAttentionPoolingHead",
+    ]
+    _supports_flash_attn_2 = True
+    _supports_sdpa = True
+
+    def _init_weights(self, module):
+        """Initialize the weights"""
+        if isinstance(module, SiglipVisionEmbeddings):
+            width = (
+                self.config.vision_config.hidden_size
+                if isinstance(self.config, SiglipConfig)
+                else self.config.hidden_size
+            )
+            nn.init.normal_(module.position_embedding.weight, std=1 / np.sqrt(width))
+        elif isinstance(module, nn.Embedding):
+            default_flax_embed_init(module.weight)
+        elif isinstance(module, SiglipAttention):
+            nn.init.xavier_uniform_(module.q_proj.weight)
+            nn.init.xavier_uniform_(module.k_proj.weight)
+            nn.init.xavier_uniform_(module.v_proj.weight)
+            nn.init.xavier_uniform_(module.out_proj.weight)
+            nn.init.zeros_(module.q_proj.bias)
+            nn.init.zeros_(module.k_proj.bias)
+            nn.init.zeros_(module.v_proj.bias)
+            nn.init.zeros_(module.out_proj.bias)
+        elif isinstance(module, SiglipMLP):
+            nn.init.xavier_uniform_(module.fc1.weight)
+            nn.init.xavier_uniform_(module.fc2.weight)
+            nn.init.normal_(module.fc1.bias, std=1e-6)
+            nn.init.normal_(module.fc2.bias, std=1e-6)
+        elif isinstance(module, SiglipMultiheadAttentionPoolingHead):
+            nn.init.xavier_uniform_(module.probe.data)
+            nn.init.xavier_uniform_(module.attention.in_proj_weight.data)
+            nn.init.zeros_(module.attention.in_proj_bias.data)
+        elif isinstance(module, SiglipModel):
+            logit_scale_init = torch.log(torch.tensor(1.0))
+            module.logit_scale.data.fill_(logit_scale_init)
+            module.logit_bias.data.zero_()
+        elif isinstance(module, SiglipForImageClassification):
+            nn.init.normal_(
+                module.classifier.weight,
+                std=self.config.vision_config.hidden_size**-0.5 * self.config.initializer_factor,
+            )
+        elif isinstance(module, (nn.Linear, nn.Conv2d)):
+            lecun_normal_(module.weight)
+            if module.bias is not None:
+                nn.init.zeros_(module.bias)
+        elif isinstance(module, nn.LayerNorm):
+            module.bias.data.zero_()
+            module.weight.data.fill_(1.0)
+
+
+SIGLIP_START_DOCSTRING = r"""
+    This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the
+    library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
+    etc.)
+
+    This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
+    Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
+    and behavior.
+
+    Parameters:
+        config ([`SiglipConfig`]): Model configuration class with all the parameters of the model.
+            Initializing with a config file does not load the weights associated with the model, only the
+            configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights.
+"""
+
+SIGLIP_TEXT_INPUTS_DOCSTRING = r"""
+    Args:
+        input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
+            Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
+            it.
+
+            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
+            [`PreTrainedTokenizer.__call__`] for details.
+
+            [What are input IDs?](../glossary#input-ids)
+        attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
+
+            - 1 for tokens that are **not masked**,
+            - 0 for tokens that are **masked**.
+
+            [What are attention masks?](../glossary#attention-mask)
+        position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
+            config.max_position_embeddings - 1]`.
+
+            [What are position IDs?](../glossary#position-ids)
+        output_attentions (`bool`, *optional*):
+            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
+            tensors for more detail.
+        output_hidden_states (`bool`, *optional*):
+            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
+            more detail.
+        return_dict (`bool`, *optional*):
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+"""
+
+SIGLIP_VISION_INPUTS_DOCSTRING = r"""
+    Args:
+        pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)`):
+            Pixel values. Padding will be ignored by default should you provide it. Pixel values can be obtained using
+            [`AutoImageProcessor`]. See [`CLIPImageProcessor.__call__`] for details.
+        output_attentions (`bool`, *optional*):
+            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
+            tensors for more detail.
+        output_hidden_states (`bool`, *optional*):
+            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
+            more detail.
+        interpolate_pos_encoding (`bool`, *optional*, defaults to `False`):
+            Whether to interpolate the pre-trained position encodings.
+        return_dict (`bool`, *optional*):
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+"""
+
+SIGLIP_INPUTS_DOCSTRING = r"""
+    Args:
+        input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
+            Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
+            it.
+
+            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
+            [`PreTrainedTokenizer.__call__`] for details.
+
+            [What are input IDs?](../glossary#input-ids)
+        attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
+
+            - 1 for tokens that are **not masked**,
+            - 0 for tokens that are **masked**.
+
+            [What are attention masks?](../glossary#attention-mask)
+        position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
+            config.max_position_embeddings - 1]`.
+
+            [What are position IDs?](../glossary#position-ids)
+        pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)`):
+            Pixel values. Padding will be ignored by default should you provide it. Pixel values can be obtained using
+            [`AutoImageProcessor`]. See [`CLIPImageProcessor.__call__`] for details.
+        return_loss (`bool`, *optional*):
+            Whether or not to return the contrastive loss.
+        output_attentions (`bool`, *optional*):
+            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
+            tensors for more detail.
+        output_hidden_states (`bool`, *optional*):
+            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
+            more detail.
+        interpolate_pos_encoding (`bool`, *optional*, defaults to `False`):
+            Whether to interpolate the pre-trained position encodings.
+        return_dict (`bool`, *optional*):
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+"""
+
+
+# Copied from transformers.models.altclip.modeling_altclip.AltCLIPEncoder with AltCLIP->Siglip
+class SiglipEncoder(nn.Module):
+    """
+    Transformer encoder consisting of `config.num_hidden_layers` self attention layers. Each layer is a
+    [`SiglipEncoderLayer`].
+
+    Args:
+        config: SiglipConfig
+    """
+
+    def __init__(self, config: SiglipConfig):
+        super().__init__()
+        self.config = config
+        self.layers = nn.ModuleList([SiglipEncoderLayer(config) for _ in range(config.num_hidden_layers)])
+        self.gradient_checkpointing = False
+
+    # Ignore copy
+    def forward(
+        self,
+        inputs_embeds,
+        attention_mask: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutput]:
+        r"""
+        Args:
+            inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
+                Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation.
+                This is useful if you want more control over how to convert `input_ids` indices into associated vectors
+                than the model's internal embedding lookup matrix.
+            attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
+                Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
+
+                - 1 for tokens that are **not masked**,
+                - 0 for tokens that are **masked**.
+
+                [What are attention masks?](../glossary#attention-mask)
+            output_attentions (`bool`, *optional*):
+                Whether or not to return the attentions tensors of all attention layers. See `attentions` under
+                returned tensors for more detail.
+            output_hidden_states (`bool`, *optional*):
+                Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
+                for more detail.
+            return_dict (`bool`, *optional*):
+                Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+        """
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        encoder_states = () if output_hidden_states else None
+        all_attentions = () if output_attentions else None
+
+        hidden_states = inputs_embeds
+        for encoder_layer in self.layers:
+            if output_hidden_states:
+                encoder_states = encoder_states + (hidden_states,)
+            if self.gradient_checkpointing and self.training:
+                layer_outputs = self._gradient_checkpointing_func(
+                    encoder_layer.__call__,
+                    hidden_states,
+                    attention_mask,
+                    output_attentions,
+                )
+            else:
+                layer_outputs = encoder_layer(
+                    hidden_states,
+                    attention_mask,
+                    output_attentions=output_attentions,
+                )
+
+            hidden_states = layer_outputs[0]
+
+            if output_attentions:
+                all_attentions = all_attentions + (layer_outputs[1],)
+
+        if output_hidden_states:
+            encoder_states = encoder_states + (hidden_states,)
+
+        if not return_dict:
+            return tuple(v for v in [hidden_states, encoder_states, all_attentions] if v is not None)
+        return BaseModelOutput(
+            last_hidden_state=hidden_states, hidden_states=encoder_states, attentions=all_attentions
+        )
+
+
+class SiglipTextTransformer(nn.Module):
+    def __init__(self, config: SiglipTextConfig):
+        super().__init__()
+        self.config = config
+        embed_dim = config.hidden_size
+        self.embeddings = SiglipTextEmbeddings(config)
+        self.encoder = SiglipEncoder(config)
+        self.final_layer_norm = nn.LayerNorm(embed_dim, eps=config.layer_norm_eps)
+
+        self.head = nn.Linear(embed_dim, embed_dim)
+        self._use_flash_attention_2 = config._attn_implementation == "flash_attention_2"
+
+    @add_start_docstrings_to_model_forward(SIGLIP_TEXT_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=SiglipTextConfig)
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        Returns:
+
+        """
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        if input_ids is None:
+            raise ValueError("You have to specify input_ids")
+
+        input_shape = input_ids.size()
+        input_ids = input_ids.view(-1, input_shape[-1])
+
+        hidden_states = self.embeddings(input_ids=input_ids, position_ids=position_ids)
+
+        # note: SigLIP's text model does not use a causal mask, unlike the original CLIP model.
+        # expand attention_mask
+        if attention_mask is not None and not self._use_flash_attention_2:
+            # [batch_size, seq_len] -> [batch_size, 1, tgt_seq_len, src_seq_len]
+            attention_mask = _prepare_4d_attention_mask(attention_mask, hidden_states.dtype)
+
+        encoder_outputs = self.encoder(
+            inputs_embeds=hidden_states,
+            attention_mask=attention_mask,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        last_hidden_state = encoder_outputs[0]
+        last_hidden_state = self.final_layer_norm(last_hidden_state)
+
+        # Assuming "sticky" EOS tokenization, last token is always EOS.
+        pooled_output = last_hidden_state[:, -1, :]
+        pooled_output = self.head(pooled_output)
+
+        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,
+        )
+
+
+@add_start_docstrings(
+    """The text model from SigLIP without any head or projection on top.""",
+    SIGLIP_START_DOCSTRING,
+)
+class SiglipTextModel(SiglipPreTrainedModel):
+    config_class = SiglipTextConfig
+
+    def __init__(self, config: SiglipTextConfig):
+        super().__init__(config)
+        self.text_model = SiglipTextTransformer(config)
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_input_embeddings(self) -> nn.Module:
+        return self.text_model.embeddings.token_embedding
+
+    def set_input_embeddings(self, value):
+        self.text_model.embeddings.token_embedding = value
+
+    @add_start_docstrings_to_model_forward(SIGLIP_TEXT_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=SiglipTextConfig)
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        Returns:
+
+        Examples:
+
+        ```python
+        >>> from transformers import AutoTokenizer, SiglipTextModel
+
+        >>> model = SiglipTextModel.from_pretrained("google/siglip-base-patch16-224")
+        >>> tokenizer = AutoTokenizer.from_pretrained("google/siglip-base-patch16-224")
+
+        >>> # important: make sure to set padding="max_length" as that's how the model was trained
+        >>> inputs = tokenizer(["a photo of a cat", "a photo of a dog"], padding="max_length", return_tensors="pt")
+
+        >>> outputs = model(**inputs)
+        >>> last_hidden_state = outputs.last_hidden_state
+        >>> pooled_output = outputs.pooler_output  # pooled (EOS token) states
+        ```"""
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        return self.text_model(
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+
+class SiglipVisionTransformer(nn.Module):
+    def __init__(self, config: SiglipVisionConfig):
+        super().__init__()
+        self.config = config
+        embed_dim = config.hidden_size
+
+        self.embeddings = SiglipVisionEmbeddings(config)
+        self.encoder = SiglipEncoder(config)
+        self.post_layernorm = nn.LayerNorm(embed_dim, eps=config.layer_norm_eps)
+        self.use_head = True if not hasattr(config, "vision_use_head") else config.vision_use_head
+        if self.use_head:
+            self.head = SiglipMultiheadAttentionPoolingHead(config)
+
+    @add_start_docstrings_to_model_forward(SIGLIP_VISION_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=SiglipVisionConfig)
+    def forward(
+        self,
+        pixel_values,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        interpolate_pos_encoding: Optional[bool] = False,
+    ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        Returns:
+
+        """
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        hidden_states = self.embeddings(pixel_values, interpolate_pos_encoding=interpolate_pos_encoding)
+
+        encoder_outputs = self.encoder(
+            inputs_embeds=hidden_states,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        last_hidden_state = encoder_outputs[0]
+        last_hidden_state = self.post_layernorm(last_hidden_state)
+
+        pooler_output = self.head(last_hidden_state) if self.use_head else None
+        if not return_dict:
+            return (last_hidden_state, pooler_output) + encoder_outputs[1:]
+
+        return BaseModelOutputWithPooling(
+            last_hidden_state=last_hidden_state,
+            pooler_output=pooler_output,
+            hidden_states=encoder_outputs.hidden_states,
+            attentions=encoder_outputs.attentions,
+        )
+
+
+class SiglipMultiheadAttentionPoolingHead(nn.Module):
+    """Multihead Attention Pooling."""
+
+    def __init__(self, config: SiglipVisionConfig):
+        super().__init__()
+
+        self.probe = nn.Parameter(torch.randn(1, 1, config.hidden_size))
+        self.attention = torch.nn.MultiheadAttention(config.hidden_size, config.num_attention_heads, batch_first=True)
+        self.layernorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
+        self.mlp = SiglipMLP(config)
+
+    def forward(self, hidden_state):
+        batch_size = hidden_state.shape[0]
+        probe = self.probe.repeat(batch_size, 1, 1)
+
+        hidden_state = self.attention(probe, hidden_state, hidden_state)[0]
+
+        residual = hidden_state
+        hidden_state = self.layernorm(hidden_state)
+        hidden_state = residual + self.mlp(hidden_state)
+
+        return hidden_state[:, 0]
+
+
+@add_start_docstrings(
+    """The vision model from SigLIP without any head or projection on top.""",
+    SIGLIP_START_DOCSTRING,
+)
+class SiglipVisionModel(SiglipPreTrainedModel):
+    config_class = SiglipVisionConfig
+    main_input_name = "pixel_values"
+
+    def __init__(self, config: SiglipVisionConfig):
+        super().__init__(config)
+
+        self.vision_model = SiglipVisionTransformer(config)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_input_embeddings(self) -> nn.Module:
+        return self.vision_model.embeddings.patch_embedding
+
+    @add_start_docstrings_to_model_forward(SIGLIP_VISION_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=SiglipVisionConfig)
+    def forward(
+        self,
+        pixel_values,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        interpolate_pos_encoding: bool = False,
+    ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        Returns:
+
+        Examples:
+
+        ```python
+        >>> from PIL import Image
+        >>> import requests
+        >>> from transformers import AutoProcessor, SiglipVisionModel
+
+        >>> model = SiglipVisionModel.from_pretrained("google/siglip-base-patch16-224")
+        >>> processor = AutoProcessor.from_pretrained("google/siglip-base-patch16-224")
+
+        >>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
+        >>> image = Image.open(requests.get(url, stream=True).raw)
+
+        >>> inputs = processor(images=image, return_tensors="pt")
+
+        >>> outputs = model(**inputs)
+        >>> last_hidden_state = outputs.last_hidden_state
+        >>> pooled_output = outputs.pooler_output  # pooled features
+        ```"""
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        return self.vision_model(
+            pixel_values=pixel_values,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+            interpolate_pos_encoding=interpolate_pos_encoding,
+        )
+
+
+@add_start_docstrings(SIGLIP_START_DOCSTRING)
+class SiglipModel(SiglipPreTrainedModel):
+    config_class = SiglipConfig
+
+    def __init__(self, config: SiglipConfig):
+        super().__init__(config)
+
+        if not isinstance(config.text_config, SiglipTextConfig):
+            raise TypeError(
+                "config.text_config is expected to be of type SiglipTextConfig but is of type"
+                f" {type(config.text_config)}."
+            )
+
+        if not isinstance(config.vision_config, SiglipVisionConfig):
+            raise TypeError(
+                "config.vision_config is expected to be of type SiglipVisionConfig but is of type"
+                f" {type(config.vision_config)}."
+            )
+
+        text_config = config.text_config
+        vision_config = config.vision_config
+
+        # First, initialize the text and vision models with proper attention implementation
+        text_model = SiglipTextModel._from_config(text_config)
+        vision_model = SiglipVisionModel._from_config(vision_config)
+
+        # Second, get the text and vision submodules (for backward compatibility)
+        self.text_model = text_model.text_model
+        self.vision_model = vision_model.vision_model
+
+        self.logit_scale = nn.Parameter(torch.randn(1))
+        self.logit_bias = nn.Parameter(torch.randn(1))
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings_to_model_forward(SIGLIP_TEXT_INPUTS_DOCSTRING)
+    def get_text_features(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> torch.FloatTensor:
+        r"""
+        Returns:
+            text_features (`torch.FloatTensor` of shape `(batch_size, output_dim`): The text embeddings obtained by
+            applying the projection layer to the pooled output of [`SiglipTextModel`].
+
+        Examples:
+
+        ```python
+        >>> from transformers import AutoTokenizer, AutoModel
+        >>> import torch
+
+        >>> model = AutoModel.from_pretrained("google/siglip-base-patch16-224")
+        >>> tokenizer = AutoTokenizer.from_pretrained("google/siglip-base-patch16-224")
+
+        >>> # important: make sure to set padding="max_length" as that's how the model was trained
+        >>> inputs = tokenizer(["a photo of a cat", "a photo of a dog"], padding="max_length", return_tensors="pt")
+        >>> with torch.no_grad():
+        ...     text_features = model.get_text_features(**inputs)
+        ```"""
+        # Use SigLIP model's config for some fields (if specified) instead of those of vision & text components.
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        text_outputs = self.text_model(
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        pooled_output = text_outputs[1]
+
+        return pooled_output
+
+    @add_start_docstrings_to_model_forward(SIGLIP_VISION_INPUTS_DOCSTRING)
+    def get_image_features(
+        self,
+        pixel_values: Optional[torch.FloatTensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        interpolate_pos_encoding: bool = False,
+    ) -> torch.FloatTensor:
+        r"""
+        Returns:
+            image_features (`torch.FloatTensor` of shape `(batch_size, output_dim`): The image embeddings obtained by
+            applying the projection layer to the pooled output of [`SiglipVisionModel`].
+
+        Examples:
+
+        ```python
+        >>> from PIL import Image
+        >>> import requests
+        >>> from transformers import AutoProcessor, AutoModel
+        >>> import torch
+
+        >>> model = AutoModel.from_pretrained("google/siglip-base-patch16-224")
+        >>> processor = AutoProcessor.from_pretrained("google/siglip-base-patch16-224")
+
+        >>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
+        >>> image = Image.open(requests.get(url, stream=True).raw)
+
+        >>> inputs = processor(images=image, return_tensors="pt")
+
+        >>> with torch.no_grad():
+        ...     image_features = model.get_image_features(**inputs)
+        ```"""
+        # Use SiglipModel's config for some fields (if specified) instead of those of vision & text components.
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        vision_outputs = self.vision_model(
+            pixel_values=pixel_values,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+            interpolate_pos_encoding=interpolate_pos_encoding,
+        )
+
+        pooled_output = vision_outputs[1]
+
+        return pooled_output
+
+    @add_start_docstrings_to_model_forward(SIGLIP_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=SiglipOutput, config_class=SiglipConfig)
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        pixel_values: Optional[torch.FloatTensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        return_loss: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        interpolate_pos_encoding: bool = False,
+    ) -> Union[Tuple, SiglipOutput]:
+        r"""
+        Returns:
+
+        Examples:
+
+        ```python
+        >>> from PIL import Image
+        >>> import requests
+        >>> from transformers import AutoProcessor, AutoModel
+        >>> import torch
+
+        >>> model = AutoModel.from_pretrained("google/siglip-base-patch16-224")
+        >>> processor = AutoProcessor.from_pretrained("google/siglip-base-patch16-224")
+
+        >>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
+        >>> image = Image.open(requests.get(url, stream=True).raw)
+
+        >>> texts = ["a photo of 2 cats", "a photo of 2 dogs"]
+        >>> # important: we pass `padding=max_length` since the model was trained with this
+        >>> inputs = processor(text=texts, images=image, padding="max_length", return_tensors="pt")
+
+        >>> with torch.no_grad():
+        ...     outputs = model(**inputs)
+
+        >>> logits_per_image = outputs.logits_per_image
+        >>> probs = torch.sigmoid(logits_per_image) # these are the probabilities
+        >>> print(f"{probs[0][0]:.1%} that image 0 is '{texts[0]}'")
+        31.9% that image 0 is 'a photo of 2 cats'
+        ```"""
+        # Use SigLIP model's config for some fields (if specified) instead of those of vision & text components.
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        vision_outputs = self.vision_model(
+            pixel_values=pixel_values,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+            interpolate_pos_encoding=interpolate_pos_encoding,
+        )
+
+        text_outputs = self.text_model(
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        image_embeds = vision_outputs[1]
+        text_embeds = text_outputs[1]
+
+        # normalized features
+        image_embeds = image_embeds / image_embeds.norm(p=2, dim=-1, keepdim=True)
+        text_embeds = text_embeds / text_embeds.norm(p=2, dim=-1, keepdim=True)
+
+        # cosine similarity as logits
+        logits_per_text = (
+            torch.matmul(text_embeds, image_embeds.t().to(text_embeds.device)) * self.logit_scale.exp()
+            + self.logit_bias
+        )
+        logits_per_image = logits_per_text.t()
+
+        loss = None
+        if return_loss:
+            # Adapted from https://github.com/google-research/big_vision/blob/01edb81a4716f93a48be43b3a4af14e29cdb3a7f/big_vision/trainers/proj/image_text/siglip.py#L287
+            eye = torch.eye(logits_per_text.size(0), device=logits_per_text.device)
+            m1_diag1 = -torch.ones_like(logits_per_text) + 2 * eye
+            loglik = torch.nn.functional.logsigmoid(m1_diag1 * logits_per_text)
+            nll = -torch.sum(loglik, dim=-1)
+            loss = nll.mean()
+
+        if not return_dict:
+            output = (logits_per_image, logits_per_text, text_embeds, image_embeds, text_outputs, vision_outputs)
+            return ((loss,) + output) if loss is not None else output
+
+        return SiglipOutput(
+            loss=loss,
+            logits_per_image=logits_per_image,
+            logits_per_text=logits_per_text,
+            text_embeds=text_embeds,
+            image_embeds=image_embeds,
+            text_model_output=text_outputs,
+            vision_model_output=vision_outputs,
+        )
+
+
+@add_start_docstrings(
+    """
+    SigLIP vision encoder with an image classification head on top (a linear layer on top of the pooled final hidden states of
+    the patch tokens) e.g. for ImageNet.
+    """,
+    SIGLIP_START_DOCSTRING,
+)
+class SiglipForImageClassification(SiglipPreTrainedModel):
+    main_input_name = "pixel_values"
+
+    def __init__(self, config: SiglipConfig) -> None:
+        super().__init__(config)
+
+        self.num_labels = config.num_labels
+
+        # Create the vision model with proper attention
+        # and take only vision_model submodule (for backward compatibility)
+        vision_model = SiglipVisionModel._from_config(config.vision_config)
+        self.vision_model = vision_model.vision_model
+
+        # Classifier head
+        self.classifier = (
+            nn.Linear(config.vision_config.hidden_size, config.num_labels) if config.num_labels > 0 else nn.Identity()
+        )
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings_to_model_forward(SIGLIP_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=ImageClassifierOutput, config_class=_CONFIG_FOR_DOC)
+    def forward(
+        self,
+        pixel_values: Optional[torch.Tensor] = None,
+        labels: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        interpolate_pos_encoding: bool = False,
+    ) -> Union[tuple, ImageClassifierOutput]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+            Labels for computing the image classification/regression loss. Indices should be in `[0, ...,
+            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
+            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
+
+        Returns:
+
+        Examples:
+
+        ```python
+        >>> from transformers import AutoImageProcessor, SiglipForImageClassification
+        >>> import torch
+        >>> from PIL import Image
+        >>> import requests
+
+        >>> torch.manual_seed(3)  # doctest: +IGNORE_RESULT
+        >>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
+        >>> image = Image.open(requests.get(url, stream=True).raw)
+
+        >>> # note: we are loading a `SiglipModel` from the hub here,
+        >>> # so the head will be randomly initialized, hence the predictions will be random if seed is not set above.
+        >>> image_processor = AutoImageProcessor.from_pretrained("google/siglip-base-patch16-224")
+        >>> model = SiglipForImageClassification.from_pretrained("google/siglip-base-patch16-224")
+
+        >>> inputs = image_processor(images=image, return_tensors="pt")
+        >>> outputs = model(**inputs)
+        >>> logits = outputs.logits
+        >>> # model predicts one of the two classes
+        >>> predicted_class_idx = logits.argmax(-1).item()
+        >>> print("Predicted class:", model.config.id2label[predicted_class_idx])
+        Predicted class: LABEL_1
+        ```"""
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        outputs = self.vision_model(
+            pixel_values,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+            interpolate_pos_encoding=interpolate_pos_encoding,
+        )
+
+        sequence_output = outputs[0]
+
+        # average pool the patch tokens
+        sequence_output = torch.mean(sequence_output, dim=1)
+        # apply classifier
+        logits = self.classifier(sequence_output)
+
+        loss = None
+        if labels is not None:
+            # move labels to correct device to enable model parallelism
+            labels = labels.to(logits.device)
+            if self.config.problem_type is None:
+                if self.num_labels == 1:
+                    self.config.problem_type = "regression"
+                elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int):
+                    self.config.problem_type = "single_label_classification"
+                else:
+                    self.config.problem_type = "multi_label_classification"
+
+            if self.config.problem_type == "regression":
+                loss_fct = MSELoss()
+                if self.num_labels == 1:
+                    loss = loss_fct(logits.squeeze(), labels.squeeze())
+                else:
+                    loss = loss_fct(logits, labels)
+            elif self.config.problem_type == "single_label_classification":
+                loss_fct = CrossEntropyLoss()
+                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
+            elif self.config.problem_type == "multi_label_classification":
+                loss_fct = BCEWithLogitsLoss()
+                loss = loss_fct(logits, labels)
+
+        if not return_dict:
+            output = (logits,) + outputs[2:]
+            return ((loss,) + output) if loss is not None else output
+
+        return ImageClassifierOutput(
+            loss=loss,
+            logits=logits,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+
+__all__ = [
+    "SiglipModel",
+    "SiglipPreTrainedModel",
+    "SiglipTextModel",
+    "SiglipVisionModel",
+    "SiglipForImageClassification",
+]

preprocessor_config.json

@@ -0,0 +1,24 @@
+{
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "SiglipImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "processor_class": "SiglipProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 384,
+    "width": 384
+  }
+}

special_tokens_map.json

@@ -0,0 +1,34 @@
+{
+  "additional_special_tokens": [
+    {
+      "content": "<|endofchunk|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": "<s>",
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<|pad|>"
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da321834704c685bd2e638824b76f3beb841556558a6df5b8a43bf2ad31f7b94
+size 11423209

tokenizer_config.json

@@ -0,0 +1,253 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "128245": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151669": {
+      "content": "<|endofchunk|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151670": {
+      "content": "<|pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|endofchunk|>",
+    "<s>",
+    "<|pad|>"
+  ],
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|pad|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}