Add files using upload-large-folder tool

VLMEvalKit-sudoku/docs/en/Quickstart.md

@@ -0,0 +1,221 @@
+# Quickstart
+
+Before running the evaluation script, you need to **configure** the VLMs and set the model_paths properly.
+
+After that, you can use a single script `run.py` to inference and evaluate multiple VLMs and benchmarks at a same time.
+
+## Step 0. Installation & Setup essential keys
+
+**Installation.**
+
+```bash
+git clone https://github.com/open-compass/VLMEvalKit.git
+cd VLMEvalKit
+pip install -e .
+```
+
+**Setup Keys.**
+
+To infer with API models (GPT-4v, Gemini-Pro-V, etc.) or use LLM APIs as the **judge or choice extractor**, you need to first setup API keys. VLMEvalKit will use an judge **LLM** to extract answer from the output if you set the key, otherwise it uses the **exact matching** mode (find "Yes", "No", "A", "B", "C"... in the output strings). **The exact matching can only be applied to the Yes-or-No tasks and the Multi-choice tasks.**
+- You can place the required keys in `$VLMEvalKit/.env` or directly set them as the environment variable. If you choose to create a `.env` file, its content will look like:
+
+  ```bash
+  # The .env file, place it under $VLMEvalKit
+  # API Keys of Proprietary VLMs
+  # QwenVL APIs
+  DASHSCOPE_API_KEY=
+  # Gemini w. Google Cloud Backends
+  GOOGLE_API_KEY=
+  # OpenAI API
+  OPENAI_API_KEY=
+  OPENAI_API_BASE=
+  # StepAI API
+  STEPAI_API_KEY=
+  # REKA API
+  REKA_API_KEY=
+  # GLMV API
+  GLMV_API_KEY=
+  # CongRong API
+  CW_API_BASE=
+  CW_API_KEY=
+  # SenseNova API
+  SENSENOVA_API_KEY=
+  # Hunyuan-Vision API
+  HUNYUAN_SECRET_KEY=
+  HUNYUAN_SECRET_ID=
+  # LMDeploy API
+  LMDEPLOY_API_BASE=
+  # You can also set a proxy for calling api models during the evaluation stage
+  EVAL_PROXY=
+  ```
+
+- Fill the blanks with your API keys (if necessary). Those API keys will be automatically loaded when doing the inference and evaluation.
+## Step 1. Configuration
+
+**VLM Configuration**: All VLMs are configured in `vlmeval/config.py`. Few legacy VLMs (like MiniGPT-4, LLaVA-v1-7B) requires additional configuration (configuring the code / model_weight root in the config file). During evaluation, you should use the model name specified in `supported_VLM` in `vlmeval/config.py` to select the VLM. Make sure you can successfully infer with the VLM before starting the evaluation with the following command `vlmutil check {MODEL_NAME}`.
+
+## Step 2. Evaluation
+
+**New!!!**  We integrated a new config system to enable more flexible evaluation settings. Check the [Document](/docs/en/ConfigSystem.md) or run `python run.py --help` for more details 🔥🔥🔥
+
+We use `run.py` for evaluation. To use the script, you can use `$VLMEvalKit/run.py` or create a soft-link of the script (to use the script anywhere):
+
+**Arguments**
+
+- `--data (list[str])`: Set the dataset names that are supported in VLMEvalKit (names can be found in the codebase README).
+- `--model (list[str])`: Set the VLM names that are supported in VLMEvalKit (defined in `supported_VLM` in `vlmeval/config.py`).
+- `--mode (str, default to 'all', choices are ['all', 'infer'])`: When `mode` set to "all", will perform both inference and evaluation; when set to "infer", will only perform the inference.
+- `--api-nproc (int, default to 4)`: The number of threads for OpenAI API calling.
+- `--work-dir (str, default to '.')`: The directory to save evaluation results.
+
+**Command for Evaluating Image Benchmarks **
+
+You can run the script with `python` or `torchrun`:
+
+```bash
+# When running with `python`, only one VLM instance is instantiated, and it might use multiple GPUs (depending on its default behavior).
+# That is recommended for evaluating very large VLMs (like IDEFICS-80B-Instruct).
+
+# IDEFICS-80B-Instruct on MMBench_DEV_EN, MME, and SEEDBench_IMG, Inference and Evalution
+python run.py --data MMBench_DEV_EN MME SEEDBench_IMG --model idefics_80b_instruct --verbose
+# IDEFICS-80B-Instruct on MMBench_DEV_EN, MME, and SEEDBench_IMG, Inference only
+python run.py --data MMBench_DEV_EN MME SEEDBench_IMG --model idefics_80b_instruct --verbose --mode infer
+
+# When running with `torchrun`, one VLM instance is instantiated on each GPU. It can speed up the inference.
+# However, that is only suitable for VLMs that consume small amounts of GPU memory.
+
+# IDEFICS-9B-Instruct, Qwen-VL-Chat, mPLUG-Owl2 on MMBench_DEV_EN, MME, and SEEDBench_IMG. On a node with 8 GPU. Inference and Evaluation.
+torchrun --nproc-per-node=8 run.py --data MMBench_DEV_EN MME SEEDBench_IMG --model idefics_80b_instruct qwen_chat mPLUG-Owl2 --verbose
+# Qwen-VL-Chat on MME. On a node with 2 GPU. Inference and Evaluation.
+torchrun --nproc-per-node=2 run.py --data MME --model qwen_chat --verbose
+```
+
+**Command for Evaluating Video Benchmarks**
+
+```bash
+# When running with `python`, only one VLM instance is instantiated, and it might use multiple GPUs (depending on its default behavior).
+# That is recommended for evaluating very large VLMs (like IDEFICS-80B-Instruct).
+
+# IDEFICS2-8B on MMBench-Video, with 8 frames as inputs and vanilla evaluation. On a node with 8 GPUs. MMBench_Video_8frame_nopack is a defined dataset setting in `vlmeval/dataset/video_dataset_config.py`.
+torchrun --nproc-per-node=8 run.py --data MMBench_Video_8frame_nopack --model idefics2_8
+# GPT-4o (API model) on MMBench-Video, with 1 frame per second as inputs and pack evaluation (all questions of a video in a single query).
+python run.py --data MMBench_Video_1fps_pack --model GPT4o
+```
+
+The evaluation results will be printed as logs, besides. **Result Files** will also be generated in the directory `$YOUR_WORKING_DIRECTORY/{model_name}`. Files ending with `.csv` contain the evaluated metrics.
+
+### Frequently Asked Questions
+
+#### Constructing Input Prompt: The `build_prompt()` Function
+If you find that the model's output does not match the expected results when evaluating a specific benchmark, it could be due to the model not constructing the input prompt correctly.
+
+In VLMEvalKit, each `dataset` class includes a function named `build_prompt()`, which is responsible for formatting input questions. Different benchmarks can either customize their own `build_prompt()` function or use the default implementation.
+
+For instance, when handling the default [Multiple-Choice QA](https://github.com/open-compass/VLMEvalKit/blob/43af13e052de6805a8b08cd04aed5e0d74f82ff5/vlmeval/dataset/image_mcq.py#L164), the `ImageMCQDataset.build_prompt()` method combines elements such as `hint`, `question`, and `options` (if present in the dataset) into a complete question format, as shown below:
+
+```
+HINT
+QUESTION
+Options:
+A. Option A
+B. Option B
+···
+Please select the correct answer from the options above.
+```
+
+Additionally, since different models may have varying evaluation requirements, VLMEvalKit also supports customizing the prompt construction method at the model level through `model.build_prompt()`. For an example, you can refer to [InternVL](https://github.com/open-compass/VLMEvalKit/blob/43af13e052de6805a8b08cd04aed5e0d74f82ff5/vlmeval/vlm/internvl_chat.py#L324).
+
+**Note: If both `model.build_prompt()` and `dataset.build_prompt()` are defined, `model.build_prompt()` will take precedence over `dataset.build_prompt()`, effectively overriding it.**
+
+Some models, such as Qwen2VL and InternVL, define extensive prompt-building methods for various types of benchmarks. To provide more flexibility in adapting to different benchmarks, VLMEvalKit allows users to customize the `model.use_custom_prompt()` function within the model. By adding or modifying the `use_custom_prompt()` function, you can decide which benchmarks should utilize the model's custom prompt logic. Below is an example:
+
+```python
+def use_custom_prompt(self, dataset: str) -> bool:
+    from vlmeval.dataset import DATASET_TYPE, DATASET_MODALITY
+    dataset_type = DATASET_TYPE(dataset, default=None)
+    if not self._use_custom_prompt:
+        return False
+    if listinstr(['MMVet'], dataset):
+        return True
+    if dataset_type == 'MCQ':
+        return True
+    if DATASET_MODALITY(dataset) == 'VIDEO':
+        return False
+    return False
+```
+Only when the `use_custom_prompt()` function returns `True` will VLMEvalKit call the model's `build_prompt()` function for the current benchmark.
+With this approach, you can flexibly control which benchmarks use the model's custom prompt logic based on your specific needs, thereby better adapting to different models and tasks.
+
+#### Model Splitting
+
+Currently, VLMEvalKit automatically supports GPU resource allocation and model splitting between processes on the same machine. This feature is supported when the inference backend is `lmdeploy` or `transformers`, with the following behaviors:
+
+- When launching with `python` command, the model is by default allocated to all available GPUs. If you want to specify which GPUs to use, you can use `CUDA_VISIBLE_DEVICES` environment variable.
+- When starting with `torchrun` command, each model instance will be allocated to `N_GPU // N_PROC` GPUs, where `N_PROC` is the number of processes specified by the `--nproc-per-node` parameter in the torchrun command. The value of `N_GPU` is determined as follows:
+  - If `CUDA_VISIBLE_DEVICES` environment variable is not set, `N_GPU` will be the total number of available GPUs.
+  - If `CUDA_VISIBLE_DEVICES` environment variable is set, `N_GPU` will be the number of GPUs specified by the `CUDA_VISIBLE_DEVICES` environment variable, and only the specified GPUs will be utilized.
+Below are specific examples of running evaluation tasks on a machine equipped with 8 GPUs:
+
+```bash
+<!-- Launch two model instances in data parallel, each instance using 4 GPUs -->
+torchrun --nproc-per-node=2 run.py --data MMBench_DEV_EN --model InternVL3-78B
+<!-- Launch one model instance, using all 8 GPUs -->
+python run.py --data MMBench_DEV_EN --model InternVL3-78B
+<!-- Launch three model instances, each instance using 2 GPUs, GPU 0 and 7 are not used -->
+CUDA_VISIBLE_DEVICES=1,2,3,4,5,6 torchrun --nproc-per-node=3 run.py --data MMBench_DEV_EN --model InternVL3-38B
+```
+
+PS: The feature is not compatible with `vllm` backend. When you evaluate a model with `vllm` backend, please use `python` to launch, and all visible GPU devices will be used.
+
+#### Performance Discrepancies
+
+Model performance may vary across different environments. As a result, you might observe discrepancies between your evaluation results and those listed on the official VLMEvalKit leaderboard. These differences could be attributed to variations in versions of libraries such as `transformers`, `cuda`, and `torch`.
+
+Besides, if you encounter unexpected performance, we recommend first reviewing the local generation records (`{model}_{dataset}.xlsx`) or the evaluation records (`{model}_{dataset}_{judge_model}.xlsx`). This may help you better understand the evaluation outcomes and identify potential issues.
+
+## Deploy a local language model as the judge / choice extractor
+The default setting mentioned above uses OpenAI's GPT as the judge LLM. However, you can also deploy a local judge LLM with [LMDeploy](https://github.com/InternLM/lmdeploy).
+
+First install:
+```
+pip install lmdeploy openai
+```
+
+And then deploy a local judge LLM with the single line of code. LMDeploy will automatically download the model from Huggingface. Assuming we use internlm2-chat-1_8b as the judge, port 23333, and the key sk-123456 (the key must start with "sk-" and follow with any number you like):
+```
+lmdeploy serve api_server internlm/internlm2-chat-1_8b --server-port 23333
+```
+
+You need to get the model name registered by LMDeploy with the following python code:
+```
+from openai import OpenAI
+client = OpenAI(
+    api_key='sk-123456',
+    base_url="http://0.0.0.0:23333/v1"
+)
+model_name = client.models.list().data[0].id
+```
+
+Now set some environment variables to tell VLMEvalKit how to use the local judge LLM. As mentioned above, you can also set them in `$VLMEvalKit/.env` file:
+```
+OPENAI_API_KEY=sk-123456
+OPENAI_API_BASE=http://0.0.0.0:23333/v1/chat/completions
+LOCAL_LLM=<model_name you get>
+```
+
+Finally, you can run the commands in step 2 to evaluate your VLM with the local judge LLM.
+
+Note that
+
+- If you hope to deploy the judge LLM in a single GPU and evaluate your VLM on other GPUs because of limited GPU memory, try `CUDA_VISIBLE_DEVICES=x` like
+```
+CUDA_VISIBLE_DEVICES=0 lmdeploy serve api_server internlm/internlm2-chat-1_8b --server-port 23333
+CUDA_VISIBLE_DEVICES=1,2,3 torchrun --nproc-per-node=3 run.py --data HallusionBench  --model qwen_chat --verbose
+```
+- If the local judge LLM is not good enough in following the instructions, the evaluation may fail. Please report such failures (e.g., by issues).
+- It's possible to deploy the judge LLM in different ways, e.g., use a private LLM (not from HuggingFace) or use a quantized LLM. Please refer to the [LMDeploy doc](https://lmdeploy.readthedocs.io/en/latest/serving/api_server.html). You can use any other deployment framework if they support OpenAI API.
+
+
+### Using LMDeploy to Accelerate Evaluation and Inference
+
+You can refer this [doc](/docs/en/EvalByLMDeploy.md)

VLMEvalKit-sudoku/docs/en/_templates/404.html

@@ -0,0 +1,18 @@
+{% extends "layout.html" %}
+
+{% block body %}
+
+<h1>Page Not Found</h1>
+<p>
+  The page you are looking for cannot be found.
+</p>
+<p>
+  If you just switched documentation versions, it is likely that the page you were on is moved. You can look for it in
+  the content table left, or go to <a href="{{ pathto(root_doc) }}">the homepage</a>.
+</p>
+<!-- <p>
+  If you cannot find documentation you want, please <a
+    href="">open an issue</a> to tell us!
+</p> -->
+
+{% endblock %}

VLMEvalKit-sudoku/docs/en/_templates/callable.rst

@@ -0,0 +1,14 @@
+.. role:: hidden
+    :class: hidden-section
+.. currentmodule:: {{ module }}
+
+
+{{ name | underline}}
+
+.. autoclass:: {{ name }}
+    :members:
+    :special-members: __call__
+
+..
+  autogenerated from _templates/callable.rst
+  note it does not have :inherited-members:

VLMEvalKit-sudoku/docs/zh-CN/Makefile

@@ -0,0 +1,20 @@
+# Minimal makefile for Sphinx documentation
+#
+
+# You can set these variables from the command line, and also
+# from the environment for the first two.
+SPHINXOPTS    ?=
+SPHINXBUILD   ?= sphinx-build
+SOURCEDIR     = .
+BUILDDIR      = _build
+
+# Put it first so that "make" without argument is like "make help".
+help:
+	@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
+
+.PHONY: help Makefile
+
+# Catch-all target: route all unknown targets to Sphinx using the new
+# "make mode" option.  $(O) is meant as a shortcut for $(SPHINXOPTS).
+%: Makefile
+	@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)

VLMEvalKit-sudoku/docs/zh-CN/_static/css/readthedocs.css

@@ -0,0 +1,63 @@
+.header-logo {
+  background-image: url("../image/logo.svg");
+  background-size: 275px 80px;
+  height: 80px;
+  width: 275px;
+}
+
+
+@media screen and (min-width: 1100px) {
+  .header-logo {
+    top: -25px;
+  }
+}
+
+pre {
+    white-space: pre;
+}
+
+@media screen and (min-width: 2000px) {
+  .pytorch-content-left {
+    width: 1200px;
+    margin-left: 30px;
+  }
+  article.pytorch-article {
+    max-width: 1200px;
+  }
+  .pytorch-breadcrumbs-wrapper {
+    width: 1200px;
+  }
+  .pytorch-right-menu.scrolling-fixed {
+    position: fixed;
+    top: 45px;
+    left: 1580px;
+  }
+}
+
+
+article.pytorch-article section code {
+  padding: .2em .4em;
+  background-color: #f3f4f7;
+  border-radius: 5px;
+}
+
+/* Disable the change in tables */
+article.pytorch-article section table code {
+  padding: unset;
+  background-color: unset;
+  border-radius: unset;
+}
+
+table.autosummary td {
+  width: 50%
+}
+
+img.align-center {
+  display: block;
+  margin-left: auto;
+  margin-right: auto;
+}
+
+article.pytorch-article p.rubric {
+  font-weight: bold;
+}

VLMEvalKit-sudoku/docs/zh-CN/_static/js/custom.js

@@ -0,0 +1,10 @@
+var collapsedSections = [];
+
+$(document).ready(function () {
+  $('.model-summary').DataTable({
+    "stateSave": false,
+    "lengthChange": false,
+    "pageLength": 20,
+    "order": []
+  });
+});

VLMEvalKit-sudoku/llava/constants.py

@@ -0,0 +1,12 @@
+CONTROLLER_HEART_BEAT_EXPIRATION = 30
+WORKER_HEART_BEAT_INTERVAL = 15
+
+LOGDIR = "."
+
+# Model Constants
+IGNORE_INDEX = -100
+IMAGE_TOKEN_INDEX = -200
+DEFAULT_IMAGE_TOKEN = "<image>"
+DEFAULT_IMAGE_PATCH_TOKEN = "<im_patch>"
+DEFAULT_IM_START_TOKEN = "<im_start>"
+DEFAULT_IM_END_TOKEN = "<im_end>"

VLMEvalKit-sudoku/llava/conversation.py

@@ -0,0 +1,596 @@
+import dataclasses
+from enum import auto, Enum
+from typing import List, Any, Dict, Union, Tuple
+import re
+import base64
+from io import BytesIO
+from PIL import Image
+from transformers import AutoTokenizer
+
+
+class SeparatorStyle(Enum):
+    """Different separator style."""
+
+    SINGLE = auto()
+    TWO = auto()
+    MPT = auto()
+    PLAIN = auto()
+    CHATML = auto()
+    LLAMA_2 = auto()
+    LLAMA_3 = auto()
+    QWEN = auto()
+    GEMMA = auto()
+
+
+@dataclasses.dataclass
+class Conversation:
+    """A class that keeps all conversation history."""
+
+    system: str
+    roles: List[str]
+    messages: List[List[str]]
+    offset: int
+    sep_style: SeparatorStyle = SeparatorStyle.SINGLE
+    sep: str = "###"
+    sep2: str = None
+    version: str = "Unknown"
+
+    tokenizer_id: str = ""
+    tokenizer: Any = None
+    # Stop criteria (the default one is EOS token)
+    stop_str: Union[str, List[str]] = None
+    # Stops generation if meeting any token in this list
+    stop_token_ids: List[int] = None
+
+    skip_next: bool = False
+
+    def get_prompt(self):
+        messages = self.messages
+        if len(messages) > 0 and type(messages[0][1]) is tuple:
+            messages = self.messages.copy()
+            init_role, init_msg = messages[0].copy()
+            init_msg = init_msg[0]
+            if "mmtag" in self.version:
+                init_msg = init_msg.replace("<image>", "").strip()
+                messages[0] = (init_role, init_msg)
+                messages.insert(0, (self.roles[0], "<Image><image></Image>"))
+                messages.insert(1, (self.roles[1], "Received."))
+            elif not init_msg.startswith("<image>"):
+                init_msg = init_msg.replace("<image>", "").strip()
+                messages[0] = (init_role, "<image>\n" + init_msg)
+            else:
+                messages[0] = (init_role, init_msg)
+
+        if self.sep_style == SeparatorStyle.SINGLE:
+            ret = self.system + self.sep
+            for role, message in messages:
+                if message:
+                    if type(message) is tuple:
+                        message, _, _ = message
+                    ret += role + ": " + message + self.sep
+                else:
+                    ret += role + ":"
+
+        elif self.sep_style == SeparatorStyle.TWO:
+            seps = [self.sep, self.sep2]
+            ret = self.system + seps[0]
+            for i, (role, message) in enumerate(messages):
+                if message:
+                    if type(message) is tuple:
+                        message, _, _ = message
+                    ret += role + ": " + message + seps[i % 2]
+                else:
+                    ret += role + ":"
+
+        elif self.sep_style == SeparatorStyle.CHATML:
+            ret = "" if self.system == "" else self.system + self.sep + "\n"
+            for role, message in messages:
+                if message:
+                    if type(message) is tuple:
+                        message, images, _ = message
+                        message = "<image>" * len(images) + message
+                    ret += role + "\n" + message + self.sep + "\n"
+                else:
+                    ret += role + "\n"
+            return ret
+
+        elif self.sep_style == SeparatorStyle.LLAMA_3:
+            if self.tokenizer is None:
+                raise ValueError("Llama 3 tokenizer is not available. Make sure you have the necessary permissions.")
+            chat_template_messages = [{"role": "system", "content": self.system}]
+            for role, message in messages:
+                if message:
+                    if type(message) is tuple:
+                        message, images = message
+                        message = "<image>" * len(images) + message
+                    chat_template_messages.append({"role": role, "content": message})
+
+            # print(chat_template_messages)
+            return self.tokenizer.apply_chat_template(chat_template_messages, tokenize=False, add_generation_prompt=True)
+            # ret = "" if self.system == "" else self.system + self.sep + "\n"
+            # for role, message in messages:
+            #     if message:
+            #         if type(message) is tuple:
+            #             message, images = message
+            #             message = "<image>" * len(images) + message
+            #         ret += role + "\n" + message + self.sep + "\n"
+            #     else:
+            #         ret += role + "\n"
+            # return ret
+
+        elif self.sep_style == SeparatorStyle.MPT:
+            ret = self.system + self.sep
+            for role, message in messages:
+                if message:
+                    if type(message) is tuple:
+                        message, _, _ = message
+                    ret += role + message + self.sep
+                else:
+                    ret += role
+
+        elif self.sep_style == SeparatorStyle.GEMMA:
+            ret = ""
+            for i, (role, message) in enumerate(messages):
+                assert role == self.roles[i % 2], "Conversation should alternate user/assistant/user/assistant/..."
+                if message:
+                    if type(message) is tuple:
+                        message, _, _ = message
+                    ret += role + message + self.sep
+                else:
+                    ret += role
+
+        elif self.sep_style == SeparatorStyle.LLAMA_2:
+            wrap_sys = lambda msg: f"<<SYS>>\n{msg}\n<</SYS>>\n\n" if len(msg) > 0 else msg
+            wrap_inst = lambda msg: f"[INST] {msg} [/INST]"
+            ret = ""
+
+            for i, (role, message) in enumerate(messages):
+                if i == 0:
+                    assert message, "first message should not be none"
+                    assert role == self.roles[0], "first message should come from user"
+                if message:
+                    if type(message) is tuple:
+                        message, _, _ = message
+                    if i == 0:
+                        message = wrap_sys(self.system) + message
+                    if i % 2 == 0:
+                        message = wrap_inst(message)
+                        ret += self.sep + message
+                    else:
+                        ret += " " + message + " " + self.sep2
+                else:
+                    ret += ""
+            ret = ret.lstrip(self.sep)
+
+        elif self.sep_style == SeparatorStyle.PLAIN:
+            seps = [self.sep, self.sep2]
+            ret = self.system
+            for i, (role, message) in enumerate(messages):
+                if message:
+                    if type(message) is tuple:
+                        message, _, _ = message
+                    ret += message + seps[i % 2]
+                else:
+                    ret += ""
+        else:
+            raise ValueError(f"Invalid style: {self.sep_style}")
+
+        return ret
+
+    def append_message(self, role, message):
+        self.messages.append([role, message])
+
+    def process_image(self, image, image_process_mode, return_pil=False, image_format="PNG"):
+        if image_process_mode == "Pad":
+
+            def expand2square(pil_img, background_color=(122, 116, 104)):
+                width, height = pil_img.size
+                if width == height:
+                    return pil_img
+                elif width > height:
+                    result = Image.new(pil_img.mode, (width, width), background_color)
+                    result.paste(pil_img, (0, (width - height) // 2))
+                    return result
+                else:
+                    result = Image.new(pil_img.mode, (height, height), background_color)
+                    result.paste(pil_img, ((height - width) // 2, 0))
+                    return result
+
+            image = expand2square(image)
+        elif image_process_mode in ["Default", "Crop"]:
+            pass
+        elif image_process_mode == "Resize":
+            image = image.resize((336, 336))
+        else:
+            raise ValueError(f"Invalid image_process_mode: {image_process_mode}")
+
+        if type(image) is not Image.Image:
+            image = Image.open(image).convert("RGB")
+
+        max_hw, min_hw = max(image.size), min(image.size)
+        aspect_ratio = max_hw / min_hw
+        max_len, min_len = 672, 448
+        shortest_edge = int(min(max_len / aspect_ratio, min_len, min_hw))
+        longest_edge = int(shortest_edge * aspect_ratio)
+        W, H = image.size
+        if H > W:
+            H, W = longest_edge, shortest_edge
+        else:
+            H, W = shortest_edge, longest_edge
+        image = image.resize((W, H))
+        if return_pil:
+            return image
+        else:
+            buffered = BytesIO()
+            image.save(buffered, format=image_format)
+            img_b64_str = base64.b64encode(buffered.getvalue()).decode()
+            return img_b64_str
+
+    def get_images(self, return_pil=False, return_path=False):
+        images = []
+        for i, (role, msg) in enumerate(self.messages[self.offset :]):
+            if i % 2 == 0:
+                if type(msg) is tuple:
+                    msg, image, image_process_mode = msg
+                    if type(image) != list:
+                        image = [image]
+                    for img in image:
+                        if not return_path and self.is_image_file(img):
+                            img = self.process_image(img, image_process_mode, return_pil=return_pil)
+                        else:
+                            images.append(img)
+        return images
+
+    def is_image_file(self, filename):
+        image_extensions = [".png", ".jpg", ".jpeg", ".gif", ".bmp", ".tiff", ".webp"]
+        return any(filename.lower().endswith(ext) for ext in image_extensions)
+
+    def is_video_file(self, filename):
+        video_extensions = [".mp4", ".mov", ".avi", ".mkv", ".wmv", ".flv", ".mpeg", ".mpg"]
+        return any(filename.lower().endswith(ext) for ext in video_extensions)
+
+    def to_gradio_chatbot(self):
+        ret = []
+        for i, (role, msg) in enumerate(self.messages[self.offset :]):
+            if i % 2 == 0:
+                if type(msg) is tuple:
+                    msg, image, image_process_mode = msg
+                    if type(image) != list:
+                        image = [image]
+                    if len(image) == 1:
+                        msg = "<image>\n" + msg.replace("<image>", "").strip()
+                    else:
+                        msg = re.sub(r"(<image>)\n(?=<image>)", r"\1 ", msg)
+
+                    img_str_list = []                         
+                    for img in image:
+                        if self.is_image_file(img):
+                            img_b64_str = self.process_image(img, "Default", return_pil=False, image_format="JPEG")
+                            img_str = f'<img src="data:image/jpeg;base64,{img_b64_str}" style="max-width: 256px; max-height: 256px; width: auto; height: auto; object-fit: contain;"/>'
+                            img_str_list.append(img_str)
+                        elif self.is_video_file(img):
+                            ret.append(((img,), None))
+
+                    msg = msg.strip()
+                    img_place_holder = ""
+                    for img_str in img_str_list:
+                        img_place_holder += f"{img_str}\n\n"
+
+                    if len(img_str_list) > 0:
+                        msg = f"{img_place_holder}\n\n{msg}"
+
+                    if len(msg) > 0:
+                        ret.append([msg, None])
+                else:
+                    ret.append([msg, None])
+            else:
+                ret[-1][-1] = msg
+        return ret
+
+    def copy(self):
+        return Conversation(system=self.system, roles=self.roles, messages=[[x, y] for x, y in self.messages], offset=self.offset, sep_style=self.sep_style, sep=self.sep, sep2=self.sep2, version=self.version)
+
+    def dict(self):
+        if len(self.get_images()) > 0:
+            return {
+                "system": self.system,
+                "roles": self.roles,
+                "messages": [[x, y[0] if type(y) is tuple else y] for x, y in self.messages],
+                "offset": self.offset,
+                "sep": self.sep,
+                "sep2": self.sep2,
+            }
+        return {
+            "system": self.system,
+            "roles": self.roles,
+            "messages": self.messages,
+            "offset": self.offset,
+            "sep": self.sep,
+            "sep2": self.sep2,
+        }
+
+
+conv_vicuna_v0 = Conversation(
+    system="A chat between a curious human and an artificial intelligence assistant. " "The assistant gives helpful, detailed, and polite answers to the human's questions.",
+    roles=("Human", "Assistant"),
+    messages=[
+        ["Human", "What are the key differences between renewable and non-renewable energy sources?"],
+        [
+            "Assistant",
+            "Renewable energy sources are those that can be replenished naturally in a relatively "
+            "short amount of time, such as solar, wind, hydro, geothermal, and biomass. "
+            "Non-renewable energy sources, on the other hand, are finite and will eventually be "
+            "depleted, such as coal, oil, and natural gas. Here are some key differences between "
+            "renewable and non-renewable energy sources:\n"
+            "1. Availability: Renewable energy sources are virtually inexhaustible, while non-renewable "
+            "energy sources are finite and will eventually run out.\n"
+            "2. Environmental impact: Renewable energy sources have a much lower environmental impact "
+            "than non-renewable sources, which can lead to air and water pollution, greenhouse gas emissions, "
+            "and other negative effects.\n"
+            "3. Cost: Renewable energy sources can be more expensive to initially set up, but they typically "
+            "have lower operational costs than non-renewable sources.\n"
+            "4. Reliability: Renewable energy sources are often more reliable and can be used in more remote "
+            "locations than non-renewable sources.\n"
+            "5. Flexibility: Renewable energy sources are often more flexible and can be adapted to different "
+            "situations and needs, while non-renewable sources are more rigid and inflexible.\n"
+            "6. Sustainability: Renewable energy sources are more sustainable over the long term, while "
+            "non-renewable sources are not, and their depletion can lead to economic and social instability.\n",
+        ],
+    ],
+    offset=2,
+    sep_style=SeparatorStyle.SINGLE,
+    sep="###",
+)
+
+conv_vicuna_v1 = Conversation(
+    system="A chat between a curious user and an artificial intelligence assistant. " "The assistant gives helpful, detailed, and polite answers to the user's questions.",
+    roles=("USER", "ASSISTANT"),
+    version="v1",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.TWO,
+    sep=" ",
+    sep2="</s>",
+)
+
+conv_llama_2 = Conversation(
+    system="""You are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe.  Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature.
+
+If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.""",
+    roles=("USER", "ASSISTANT"),
+    version="llama_v2",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.LLAMA_2,
+    sep="<s>",
+    sep2="</s>",
+)
+
+conv_llava_llama_2 = Conversation(
+    system="You are a helpful language and vision assistant. " "You are able to understand the visual content that the user provides, " "and assist the user with a variety of tasks using natural language.",
+    roles=("USER", "ASSISTANT"),
+    version="llama_v2",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.LLAMA_2,
+    sep="<s>",
+    sep2="</s>",
+)
+
+def safe_load_tokenizer(tokenizer_id):
+    try:
+        return AutoTokenizer.from_pretrained(tokenizer_id)
+    except Exception:
+        return None
+
+conv_llava_llama_3 = Conversation(
+    system="You are a helpful language and vision assistant. " "You are able to understand the visual content that the user provides, " "and assist the user with a variety of tasks using natural language.",
+    roles=("user", "assistant"),
+    version="llama_v3",
+    messages=[],
+    offset=0,
+    sep="<|eot_id|>",
+    sep_style=SeparatorStyle.LLAMA_3,
+    tokenizer_id="meta-llama/Meta-Llama-3-8B-Instruct",
+    tokenizer=safe_load_tokenizer("meta-llama/Meta-Llama-3-8B-Instruct"),
+    stop_token_ids=[128009],
+)
+
+conv_mistral_instruct = Conversation(
+    system="",
+    roles=("USER", "ASSISTANT"),
+    version="llama_v2",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.LLAMA_2,
+    sep="",
+    sep2="</s>",
+)
+
+conv_llava_llama_2_simple = Conversation(
+    system="Answer the questions about the visual content that the user provides.",
+    roles=("USER", "ASSISTANT"),
+    version="llama_v2",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.LLAMA_2,
+    sep="<s>",
+    sep2="</s>",
+)
+
+conv_llava_llama_2_mmtag = Conversation(
+    system="Answer the questions about the visual content that the user provides." "The visual content will be provided with the following format: <Image>visual content</Image>.",
+    roles=("USER", "ASSISTANT"),
+    version="llama_v2_mmtag",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.LLAMA_2,
+    sep="<s>",
+    sep2="</s>",
+)
+
+conv_mpt = Conversation(
+    system="""<|im_start|>system
+A conversation between a user and an LLM-based AI assistant. The assistant gives helpful and honest answers.""",
+    roles=("<|im_start|>user\n", "<|im_start|>assistant\n"),
+    version="mpt",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.MPT,
+    sep="<|im_end|>",
+)
+
+conv_qwen = Conversation(
+    system="""<|im_start|>system
+You are a helpful assistant.""",
+    roles=("<|im_start|>user", "<|im_start|>assistant"),
+    version="qwen",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.CHATML,
+    sep="<|im_end|>",
+)
+
+conv_qwen3 = Conversation(
+    system="",
+    roles=("<|im_start|>user", "<|im_start|>assistant"),
+    version="qwen3",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.CHATML,
+    sep="<|im_end|>",
+)
+
+conv_gemma_instruct = Conversation(system="", roles=("<start_of_turn>user\n", "<start_of_turn>model\n"), version="gemma", messages=[], offset=0, sep_style=SeparatorStyle.GEMMA, sep="<end_of_turn>\n")
+
+conv_llava_plain = Conversation(
+    system="",
+    roles=("", ""),
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.PLAIN,
+    sep="\n",
+)
+
+conv_llava_v0 = Conversation(
+    system="A chat between a curious human and an artificial intelligence assistant. " "The assistant gives helpful, detailed, and polite answers to the human's questions.",
+    roles=("Human", "Assistant"),
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.SINGLE,
+    sep="###",
+)
+
+conv_llava_v0_mmtag = Conversation(
+    system="A chat between a curious user and an artificial intelligence assistant. "
+    "The assistant is able to understand the visual content that the user provides, and assist the user with a variety of tasks using natural language."
+    "The visual content will be provided with the following format: <Image>visual content</Image>.",
+    roles=("Human", "Assistant"),
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.SINGLE,
+    sep="###",
+    version="v0_mmtag",
+)
+
+conv_llava_v1 = Conversation(
+    system="A chat between a curious human and an artificial intelligence assistant. " "The assistant gives helpful, detailed, and polite answers to the human's questions.",
+    roles=("USER", "ASSISTANT"),
+    version="v1",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.TWO,
+    sep=" ",
+    sep2="</s>",
+)
+
+conv_llava_v1_mmtag = Conversation(
+    system="A chat between a curious user and an artificial intelligence assistant. "
+    "The assistant is able to understand the visual content that the user provides, and assist the user with a variety of tasks using natural language."
+    "The visual content will be provided with the following format: <Image>visual content</Image>.",
+    roles=("USER", "ASSISTANT"),
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.TWO,
+    sep=" ",
+    sep2="</s>",
+    version="v1_mmtag",
+)
+
+conv_mistral_orca = Conversation(
+    system="""<|im_start|>system
+You are MistralOrca, a large language model trained by Alignment Lab AI. Write out your reasoning step-by-step to be sure you get the right answers!""",
+    roles=("<|im_start|>user\n", "<|im_start|>assistant\n"),
+    version="mpt",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.MPT,
+    sep="<|im_end|>",
+)
+
+conv_mistral_zephyr = Conversation(
+    system="""<|system|>
+You are a helpful AI assistant.""",
+    roles=("<|user|>\n", "<|assistant|>\n"),
+    version="mpt",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.MPT,
+    sep="</s>",
+)
+
+conv_mistral_direct = Conversation(
+    system="""<|im_start|>system
+Answer the questions.""",
+    roles=("<|im_start|>user\n", "<|im_start|>assistant\n"),
+    version="mpt",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.MPT,
+    sep="<|im_end|>",
+)
+
+conv_chatml_direct = Conversation(
+    system="""<|im_start|>system
+Answer the questions.""",
+    roles=("<|im_start|>user\n", "<|im_start|>assistant\n"),
+    version="mpt",
+    messages=[],
+    offset=0,
+    sep_style=SeparatorStyle.MPT,
+    sep="<|im_end|>",
+)
+
+default_conversation = conv_vicuna_v0
+conv_templates = {
+    "default": conv_vicuna_v0,
+    "v0": conv_vicuna_v0,
+    "v1": conv_vicuna_v1,
+    "vicuna_v1": conv_vicuna_v1,
+    "llama_2": conv_llama_2,
+    "mistral_instruct": conv_mistral_instruct,
+    "mistral_orca": conv_mistral_orca,
+    "mistral_zephyr": conv_mistral_zephyr,
+    "mistral_direct": conv_mistral_direct,
+    "plain": conv_llava_plain,
+    "v0_plain": conv_llava_plain,
+    "chatml_direct": conv_chatml_direct,
+    "llava_v0": conv_llava_v0,
+    "llava_v0_mmtag": conv_llava_v0_mmtag,
+    "llava_v1": conv_llava_v1,
+    "llava_v1_mmtag": conv_llava_v1_mmtag,
+    "llava_llama_2": conv_llava_llama_2,
+    "llava_llama_3": conv_llava_llama_3,
+    "llava_llama_2_simple": conv_llava_llama_2_simple,
+    "llava_llama_2_mmtag": conv_llava_llama_2_mmtag,
+    "llava_mistral_instruct": conv_mistral_instruct,
+    "mpt": conv_mpt,
+    "qwen_1_5": conv_qwen,
+    "qwen_2": conv_qwen,
+    "qwen3": conv_qwen3,
+    "gemma_instruct": conv_gemma_instruct,
+}
+
+
+if __name__ == "__main__":
+    print(default_conversation.get_prompt())

VLMEvalKit-sudoku/llava/eval/eval_docvqa.py

@@ -0,0 +1,63 @@
+import os
+import argparse
+import json
+import re
+
+from llava.eval.m4c_evaluator import STVQAANLSEvaluator
+
+def get_args():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--annotation-file', type=str)
+    parser.add_argument('--result-file', type=str)
+    parser.add_argument('--result-dir', type=str)
+    parser.add_argument('--mid_result', type=str)
+    parser.add_argument('--output_result', type=str)
+    return parser.parse_args()
+
+
+def eval_single(annotation_file, result_file):
+    experiment_name = os.path.splitext(os.path.basename(result_file))[0]
+    print(experiment_name)
+    # annotations = json.load(open(annotation_file))['data']
+    annotations = [
+        json.loads(q) for q in open(os.path.expanduser(annotation_file), "r")
+    ]
+    annotations = {(annotation['question_id'], annotation['question'].lower()): annotation for annotation in annotations}
+    results = [json.loads(line) for line in open(result_file)]
+
+    pred_list = []
+    mid_list = []
+    for result in results:
+        annotation = annotations[(result['question_id'], result['prompt'].lower())]
+        pred_list.append({
+            "pred_answer": result['text'],
+            "gt_answers": [annotation['answer']],
+        })
+        mid_list.append(result)
+        mid_list[-1]["gt_answers"] = annotation['answer']
+
+    evaluator = STVQAANLSEvaluator()
+    acc = evaluator.eval_pred_list(pred_list)
+    acc = 100. * acc
+    print('Samples: {}\nAccuracy: {:.2f}%\n'.format(len(pred_list), acc))
+    return len(pred_list), acc, mid_list
+
+
+if __name__ == "__main__":
+    args = get_args()
+
+    if args.result_file is not None:
+        samples, acc, mid_result = eval_single(args.annotation_file, args.result_file)
+
+    if args.result_dir is not None:
+        for result_file in sorted(os.listdir(args.result_dir)):
+            if not result_file.endswith('.jsonl'):
+                print(f'Skipping {result_file}')
+                continue
+            samples, acc, mid_result = eval_single(args.annotation_file, os.path.join(args.result_dir, result_file))
+
+    # with open(args.mid_result, 'w') as f:
+    #     json.dump(mid_result, f, indent=2)
+
+    # with open(args.output_result, 'w') as f:
+    #     json.dump({'samples': samples, 'acc': acc}, f, indent=2)

VLMEvalKit-sudoku/llava/eval/eval_gpt_review_bench.py

@@ -0,0 +1,121 @@
+import argparse
+import json
+import os
+
+import openai
+import time
+
+NUM_SECONDS_TO_SLEEP = 0.5
+
+
+def get_eval(content: str, max_tokens: int):
+    while True:
+        try:
+            response = openai.ChatCompletion.create(
+                model='gpt-4-0314',
+                messages=[{
+                    'role': 'system',
+                    'content': 'You are a helpful and precise assistant for checking the quality of the answer.'
+                }, {
+                    'role': 'user',
+                    'content': content,
+                }],
+                temperature=0.2,  # TODO: figure out which temperature is best for evaluation
+                max_tokens=max_tokens,
+            )
+            break
+        except openai.error.RateLimitError:
+            pass
+        except Exception as e:
+            print(e)
+        time.sleep(NUM_SECONDS_TO_SLEEP)
+
+    return response['choices'][0]['message']['content']
+
+
+def parse_score(review):
+    try:
+        score_pair = review.split('\n')[0]
+        score_pair = score_pair.replace(',', ' ')
+        sp = score_pair.split(' ')
+        if len(sp) == 2:
+            return [float(sp[0]), float(sp[1])]
+        else:
+            print('error', review)
+            return [-1, -1]
+    except Exception as e:
+        print(e)
+        print('error', review)
+        return [-1, -1]
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(description='ChatGPT-based QA evaluation.')
+    parser.add_argument('-q', '--question')
+    parser.add_argument('-c', '--context')
+    parser.add_argument('-a', '--answer-list', nargs='+', default=[])
+    parser.add_argument('-r', '--rule')
+    parser.add_argument('-o', '--output')
+    parser.add_argument('--max-tokens', type=int, default=1024, help='maximum number of tokens produced in the output')
+    args = parser.parse_args()
+
+    f_q = open(os.path.expanduser(args.question))
+    f_ans1 = open(os.path.expanduser(args.answer_list[0]))
+    f_ans2 = open(os.path.expanduser(args.answer_list[1]))
+    rule_dict = json.load(open(os.path.expanduser(args.rule), 'r'))
+
+    if os.path.isfile(os.path.expanduser(args.output)):
+        cur_reviews = [json.loads(line) for line in open(os.path.expanduser(args.output))]
+    else:
+        cur_reviews = []
+
+    review_file = open(f'{args.output}', 'a')
+
+    context_list = [json.loads(line) for line in open(os.path.expanduser(args.context))]
+    image_to_context = {context['image']: context for context in context_list}
+
+    handles = []
+    idx = 0
+    for ques_js, ans1_js, ans2_js in zip(f_q, f_ans1, f_ans2):
+        ques = json.loads(ques_js)
+        ans1 = json.loads(ans1_js)
+        ans2 = json.loads(ans2_js)
+
+        inst = image_to_context[ques['image']]
+
+        if isinstance(inst['caption'], list):
+            cap_str = '\n'.join(inst['caption'])
+        else:
+            cap_str = inst['caption']
+
+        category = 'llava_bench_' + json.loads(ques_js)['category']
+        if category in rule_dict:
+            rule = rule_dict[category]
+        else:
+            assert False, f"Visual QA category not found in rule file: {category}."
+        prompt = rule['prompt']
+        role = rule['role']
+        content = (f'[Context]\n{cap_str}\n\n'
+                   f'[Question]\n{ques["text"]}\n\n'
+                   f'[{role} 1]\n{ans1["text"]}\n\n[End of {role} 1]\n\n'
+                   f'[{role} 2]\n{ans2["text"]}\n\n[End of {role} 2]\n\n'
+                   f'[System]\n{prompt}\n\n')
+        cur_js = {
+            'id': idx+1,
+            'question_id': ques['question_id'],
+            'answer1_id': ans1.get('answer_id', ans1['question_id']),
+            'answer2_id': ans2.get('answer_id', ans2['answer_id']),
+            'category': category
+        }
+        if idx >= len(cur_reviews):
+            review = get_eval(content, args.max_tokens)
+            scores = parse_score(review)
+            cur_js['content'] = review
+            cur_js['tuple'] = scores
+            review_file.write(json.dumps(cur_js) + '\n')
+            review_file.flush()
+        else:
+            print(f'Skipping {idx} as we already have it.')
+        idx += 1
+        print(idx)
+    review_file.close()

VLMEvalKit-sudoku/llava/eval/eval_gpt_review_visual.py

@@ -0,0 +1,118 @@
+import argparse
+import json
+import os
+
+import openai
+import time
+
+NUM_SECONDS_TO_SLEEP = 0.5
+
+
+def get_eval(content: str, max_tokens: int):
+    while True:
+        try:
+            response = openai.ChatCompletion.create(
+                model='gpt-4-0314',
+                messages=[{
+                    'role': 'system',
+                    'content': 'You are a helpful and precise assistant for checking the quality of the answer.'
+                }, {
+                    'role': 'user',
+                    'content': content,
+                }],
+                temperature=0.2,  # TODO: figure out which temperature is best for evaluation
+                max_tokens=max_tokens,
+            )
+            break
+        except openai.error.RateLimitError:
+            pass
+        except Exception as e:
+            print(e)
+        time.sleep(NUM_SECONDS_TO_SLEEP)
+
+    return response['choices'][0]['message']['content']
+
+
+def parse_score(review):
+    try:
+        score_pair = review.split('\n')[0]
+        score_pair = score_pair.replace(',', ' ')
+        sp = score_pair.split(' ')
+        if len(sp) == 2:
+            return [float(sp[0]), float(sp[1])]
+        else:
+            print('error', review)
+            return [-1, -1]
+    except Exception as e:
+        print(e)
+        print('error', review)
+        return [-1, -1]
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(description='ChatGPT-based QA evaluation.')
+    parser.add_argument('-q', '--question')
+    parser.add_argument('-c', '--context')
+    parser.add_argument('-a', '--answer-list', nargs='+', default=[])
+    parser.add_argument('-r', '--rule')
+    parser.add_argument('-o', '--output')
+    parser.add_argument('--max-tokens', type=int, default=1024, help='maximum number of tokens produced in the output')
+    args = parser.parse_args()
+
+    f_q = open(os.path.expanduser(args.question))
+    f_ans1 = open(os.path.expanduser(args.answer_list[0]))
+    f_ans2 = open(os.path.expanduser(args.answer_list[1]))
+    rule_dict = json.load(open(os.path.expanduser(args.rule), 'r'))
+
+    if os.path.isfile(os.path.expanduser(args.output)):
+        cur_reviews = [json.loads(line) for line in open(os.path.expanduser(args.output))]
+    else:
+        cur_reviews = []
+
+    review_file = open(f'{args.output}', 'a')
+
+    context_list = [json.loads(line) for line in open(os.path.expanduser(args.context))]
+    image_to_context = {context['image']: context for context in context_list}
+
+    handles = []
+    idx = 0
+    for ques_js, ans1_js, ans2_js in zip(f_q, f_ans1, f_ans2):
+        ques = json.loads(ques_js)
+        ans1 = json.loads(ans1_js)
+        ans2 = json.loads(ans2_js)
+
+        inst = image_to_context[ques['image']]
+        cap_str = '\n'.join(inst['captions'])
+        box_str = '\n'.join([f'{instance["category"]}: {instance["bbox"]}' for instance in inst['instances']])
+
+        category = json.loads(ques_js)['category']
+        if category in rule_dict:
+            rule = rule_dict[category]
+        else:
+            assert False, f"Visual QA category not found in rule file: {category}."
+        prompt = rule['prompt']
+        role = rule['role']
+        content = (f'[Context]\n{cap_str}\n\n{box_str}\n\n'
+                   f'[Question]\n{ques["text"]}\n\n'
+                   f'[{role} 1]\n{ans1["text"]}\n\n[End of {role} 1]\n\n'
+                   f'[{role} 2]\n{ans2["text"]}\n\n[End of {role} 2]\n\n'
+                   f'[System]\n{prompt}\n\n')
+        cur_js = {
+            'id': idx+1,
+            'question_id': ques['question_id'],
+            'answer1_id': ans1.get('answer_id', ans1['question_id']),
+            'answer2_id': ans2.get('answer_id', ans2['answer_id']),
+            'category': category
+        }
+        if idx >= len(cur_reviews):
+            review = get_eval(content, args.max_tokens)
+            scores = parse_score(review)
+            cur_js['content'] = review
+            cur_js['tuple'] = scores
+            review_file.write(json.dumps(cur_js) + '\n')
+            review_file.flush()
+        else:
+            print(f'Skipping {idx} as we already have it.')
+        idx += 1
+        print(idx)
+    review_file.close()

VLMEvalKit-sudoku/llava/eval/eval_science_qa.py

@@ -0,0 +1,114 @@
+import argparse
+import json
+import os
+import re
+import random
+
+
+def get_args():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--base-dir', type=str)
+    parser.add_argument('--result-file', type=str)
+    parser.add_argument('--output-file', type=str)
+    parser.add_argument('--output-result', type=str)
+    parser.add_argument('--split', type=str, default='test')
+    parser.add_argument('--options', type=list, default=["A", "B", "C", "D", "E"])
+    return parser.parse_args()
+
+
+def convert_caps(results):
+    fakecaps = []
+    for result in results:
+        image_id = result['question_id']
+        caption = result['text']
+        fakecaps.append({"image_id": int(image_id), "caption": caption})
+    return fakecaps
+
+
+def get_pred_idx(prediction, choices, options):
+    """
+    Get the index (e.g. 2) from the prediction (e.g. 'C')
+    """
+    if prediction in options[:len(choices)]:
+        return options.index(prediction)
+    else:
+        return -1
+        return random.choice(range(len(choices)))
+
+
+if __name__ == "__main__":
+    args = get_args()
+
+    base_dir = args.base_dir
+    split_indices = json.load(open(os.path.join(base_dir, "pid_splits.json")))[args.split]
+    problems = json.load(open(os.path.join(base_dir, "problems.json")))
+    predictions = [json.loads(line) for line in open(args.result_file)]
+    predictions = {pred['question_id']: pred for pred in predictions}
+    split_problems = {idx: problems[idx] for idx in split_indices}
+
+    results = {'correct': [], 'incorrect': []}
+    sqa_results = {}
+    sqa_results['acc'] = None
+    sqa_results['correct'] = None
+    sqa_results['count'] = None
+    sqa_results['results'] = {}
+    sqa_results['outputs'] = {}
+
+    for prob_id, prob in split_problems.items():
+        if prob_id not in predictions:
+            pred = {'text': 'FAILED', 'prompt': 'Unknown'}
+            pred_text = 'FAILED'
+        else:
+            pred = predictions[prob_id]
+            pred_text = pred['text']
+
+        if pred_text in args.options:
+            answer = pred_text
+        elif len(pred_text) >= 3 and pred_text[0] in args.options and pred_text[1:3] == ". ":
+            answer = pred_text[0]
+        else:
+            pattern = re.compile(r'The answer is ([A-Z]).')
+            res = pattern.findall(pred_text)
+            if len(res) == 1:
+                answer = res[0]  # 'A', 'B', ...
+            else:
+                answer = "FAILED"
+
+        pred_idx = get_pred_idx(answer, prob['choices'], args.options)
+
+        analysis = {
+            'question_id': prob_id,
+            'parsed_ans': answer,
+            'ground_truth': args.options[prob['answer']],
+            'question': pred['prompt'],
+            'pred': pred_text,
+            'is_multimodal': '<image>' in pred['prompt'],
+        }
+
+        sqa_results['results'][prob_id] = get_pred_idx(answer, prob['choices'], args.options)
+        sqa_results['outputs'][prob_id] = pred_text
+
+        if pred_idx == prob['answer']:
+            results['correct'].append(analysis)
+        else:
+            results['incorrect'].append(analysis)
+
+    correct = len(results['correct'])
+    total = len(results['correct']) + len(results['incorrect'])
+
+    ###### IMG ######
+    multimodal_correct = len([x for x in results['correct'] if x['is_multimodal']])
+    multimodal_incorrect = len([x for x in results['incorrect'] if x['is_multimodal']])
+    multimodal_total = multimodal_correct + multimodal_incorrect
+    ###### IMG ######
+
+    print(f'Total: {total}, Correct: {correct}, Accuracy: {correct / total * 100:.2f}%, IMG-Accuracy: {multimodal_correct / multimodal_total * 100:.2f}%')
+
+    sqa_results['acc'] = correct / total * 100
+    sqa_results['correct'] = correct
+    sqa_results['count'] = total
+
+    with open(args.output_file, 'w') as f:
+        json.dump(results, f, indent=2)
+    with open(args.output_result, 'w') as f:
+        json.dump(sqa_results, f, indent=2)

VLMEvalKit-sudoku/llava/eval/model_vqa_loader.py

@@ -0,0 +1,256 @@
+import argparse
+import torch
+import os
+import json
+from tqdm import tqdm
+import shortuuid
+
+from llava.constants import IMAGE_TOKEN_INDEX, DEFAULT_IMAGE_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN
+from llava.conversation import conv_templates, SeparatorStyle
+from llava.model.builder import load_pretrained_model
+from llava.utils import disable_torch_init
+from llava.mm_utils import tokenizer_image_token, process_images, get_model_name_from_path, expand2square, KeywordsStoppingCriteria
+from llava.constants import IGNORE_INDEX, DEFAULT_IMAGE_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN, IMAGE_TOKEN_INDEX
+
+from torch.utils.data import Dataset, DataLoader
+from typing import Dict, Optional, Sequence, List
+import transformers
+import re
+from PIL import Image
+import math
+from llava.slice_process import slice_image_minicpm, split_image, resize_image_keep_ratio
+
+
+def split_list(lst, n):
+    """Split a list into n (roughly) equal-sized chunks"""
+    chunk_size = math.ceil(len(lst) / n)  # integer division
+    return [lst[i:i+chunk_size] for i in range(0, len(lst), chunk_size)]
+
+
+def get_chunk(lst, n, k):
+    chunks = split_list(lst, n)
+    return chunks[k]
+
+def preprocess_qwen(sources, tokenizer: transformers.PreTrainedTokenizer, has_image: bool = False, max_len=2048, system_message: str = "You are a helpful assistant.") -> Dict:
+    roles = {"human": "<|im_start|>user", "gpt": "<|im_start|>assistant"}
+
+    im_start, im_end = tokenizer.additional_special_tokens_ids
+    nl_tokens = tokenizer("\n").input_ids
+    _system = tokenizer("system").input_ids + nl_tokens
+    _user = tokenizer("user").input_ids + nl_tokens
+    _assistant = tokenizer("assistant").input_ids + nl_tokens
+
+    # Apply prompt templates
+    input_ids, targets = [], []
+
+    source = sources
+    if roles[source[0]["from"]] != roles["human"]:
+        source = source[1:]
+
+    input_id, target = [], []
+    system = [im_start] + _system + tokenizer(system_message).input_ids + [im_end] + nl_tokens
+    input_id += system
+    target += [im_start] + [IGNORE_INDEX] * (len(system) - 3) + [im_end] + nl_tokens
+    assert len(input_id) == len(target)
+    for j, sentence in enumerate(source):
+        role = roles[sentence["from"]]
+        if has_image and sentence["value"] is not None and "<image>" in sentence["value"]:
+            num_image = len(re.findall(DEFAULT_IMAGE_TOKEN, sentence["value"]))
+            texts = sentence["value"].split('<image>')
+            _input_id = tokenizer(role).input_ids + nl_tokens 
+            for i,text in enumerate(texts):
+                _input_id += tokenizer(text).input_ids 
+                if i<len(texts)-1:
+                    _input_id += [IMAGE_TOKEN_INDEX] + nl_tokens
+            _input_id += [im_end] + nl_tokens
+            assert sum([i==IMAGE_TOKEN_INDEX for i in _input_id])==num_image
+        else:
+            if sentence["value"] is None:
+                _input_id = tokenizer(role).input_ids + nl_tokens
+            else:
+                _input_id = tokenizer(role).input_ids + nl_tokens + tokenizer(sentence["value"]).input_ids + [im_end] + nl_tokens
+        input_id += _input_id
+        if role == "<|im_start|>user":
+            _target = [im_start] + [IGNORE_INDEX] * (len(_input_id) - 3) + [im_end] + nl_tokens
+        elif role == "<|im_start|>assistant":
+            _target = [im_start] + [IGNORE_INDEX] * len(tokenizer(role).input_ids) + _input_id[len(tokenizer(role).input_ids) + 1 : -2] + [im_end] + nl_tokens
+        else:
+            raise NotImplementedError
+        target += _target
+
+    input_ids.append(input_id)
+    targets.append(target)
+    input_ids = torch.tensor(input_ids, dtype=torch.long)
+    targets = torch.tensor(targets, dtype=torch.long)
+    return input_ids
+
+# Custom dataset class
+class CustomDataset(Dataset):
+    def __init__(self, questions, image_folder, tokenizer, image_processor, model_config):
+        self.questions = questions
+        self.image_folder = image_folder
+        self.tokenizer = tokenizer
+        self.image_processor = image_processor
+        self.model_config = model_config
+
+    def __getitem__(self, index):
+        line = self.questions[index]
+        image_file = line["image"]
+        qs = line["text"]
+        processor = self.image_processor
+        if self.model_config.mm_use_im_start_end:
+            qs = DEFAULT_IM_START_TOKEN + DEFAULT_IMAGE_TOKEN + DEFAULT_IM_END_TOKEN + '\n' + qs
+        else:
+            qs = DEFAULT_IMAGE_TOKEN + '\n' + qs
+
+        conv = conv_templates[args.conv_mode].copy()
+        conv.append_message(conv.roles[0], qs)
+        conv.append_message(conv.roles[1], None)
+        prompt = conv.get_prompt()
+
+        image = Image.open(os.path.join(self.image_folder, image_file)).convert('RGB')
+        # image_tensor = process_images([image], self.image_processor, self.model_config)[0]
+
+        # 2x2切片
+        # image = expand2square(image, tuple(int(x*255) for x in processor.image_mean))
+        # sub_images = split_image(image, scale=672, grid=(2, 2))
+        # sub_images.append(image)
+        # image = sub_images
+        # image = processor.preprocess(image, return_tensors='pt')['pixel_values'] # bs, 3, h, w
+        # image_tensor = image.flatten(0, 1)
+
+        # adapt
+        # image, _, _, _ = slice_image_minicpm(
+        #     image, max_slice_nums=7, scale_resolution=336, patch_size=14, never_split=False)
+        # image = processor.preprocess(image, do_resize=False, do_center_crop=False, 
+        #                             do_rescale=True, do_normalize=True, return_tensors='pt')['pixel_values'][0]
+        # image_tensor = image
+
+        image = resize_image_keep_ratio(image, max_size=1024)
+
+        source_image, patches, best_grid, ind_tokens = slice_image_minicpm(
+            image, max_slice_nums=7, scale_resolution=336, patch_size=14, never_split=False)
+
+        if best_grid is None: #说明没有切片
+            source_tensors = processor.preprocess(source_image, do_resize=False, do_center_crop=False, 
+                                                    do_rescale=True, do_normalize=True, 
+                                                    return_tensors='pt')['pixel_values'] # 1, 3, abs_h, abs_w
+            crop_size = processor.crop_size
+            patch_tensors = torch.zeros(1, 3, crop_size['height'], crop_size['width'])
+        else:
+            source_tensors = processor.preprocess(source_image, do_resize=False, do_center_crop=False, 
+                                                    do_rescale=True, do_normalize=True, 
+                                                    return_tensors='pt')['pixel_values'] # 1, 3, abs_h, abs_w
+            patch_tensors = processor.preprocess(patches, do_resize=False, do_center_crop=False, 
+                                                    do_rescale=True, do_normalize=True, 
+                                                    return_tensors='pt')['pixel_values'] # num_slice, 3, s_h, s_w
+        image_tensor = source_tensors[0] # 3, h, w
+        patch_images = patch_tensors # bs, 3, h, w
+
+        input_ids = tokenizer_image_token(prompt, self.tokenizer, IMAGE_TOKEN_INDEX, return_tensors='pt')
+
+        return input_ids, image_tensor, image.size, patch_images, ind_tokens
+
+    def __len__(self):
+        return len(self.questions)
+
+
+def collate_fn(batch):
+    input_ids, image_tensors, image_sizes, patch_images, ind_tokens = zip(*batch)
+    input_ids = torch.stack(input_ids, dim=0)
+    image_tensors = torch.stack(image_tensors, dim=0)
+    return input_ids, image_tensors, image_sizes, patch_images, ind_tokens
+
+
+# DataLoader
+def create_data_loader(questions, image_folder, tokenizer, image_processor, model_config, batch_size=1, num_workers=4):
+    assert batch_size == 1, "batch_size must be 1"
+    dataset = CustomDataset(questions, image_folder, tokenizer, image_processor, model_config)
+    data_loader = DataLoader(dataset, batch_size=batch_size, num_workers=num_workers, shuffle=False, collate_fn=collate_fn)
+    return data_loader
+
+
+def eval_model(args):
+    # Model
+    disable_torch_init()
+    model_path = os.path.expanduser(args.model_path)
+    model_name = get_model_name_from_path(model_path)
+    tokenizer, model, image_processor, context_len = load_pretrained_model(model_path, args.model_base, model_name, _args=args)
+
+    questions = [json.loads(q) for q in open(os.path.expanduser(args.question_file), "r")]
+    questions = get_chunk(questions, args.num_chunks, args.chunk_idx)
+    answers_file = os.path.expanduser(args.answers_file)
+    os.makedirs(os.path.dirname(answers_file), exist_ok=True)
+    ans_file = open(answers_file, "w")
+
+    if 'plain' in model_name and 'finetune' not in model_name.lower() and 'mmtag' not in args.conv_mode:
+        args.conv_mode = args.conv_mode + '_mmtag'
+        print(f'It seems that this is a plain model, but it is not using a mmtag prompt, auto switching to {args.conv_mode}.')
+
+    data_loader = create_data_loader(questions, args.image_folder, tokenizer, image_processor, model.config)
+
+    for (input_ids, image_tensor, image_sizes, patch_images, ind_tokens), line in tqdm(zip(data_loader, questions), total=len(questions)):
+        idx = line["question_id"]
+        cur_prompt = line["text"]
+
+        input_ids = input_ids.to(device='cuda', non_blocking=True)
+
+        image_tensor = [image_tensor[0].to(dtype=torch.float16, device='cuda', non_blocking=True)]
+        patch_images = [item.to(dtype=torch.float16, device='cuda', non_blocking=True) for item in patch_images]
+        
+        args.conv_mode = "qwen_1_5"
+
+        conv = conv_templates[args.conv_mode].copy()
+        stop_str = conv.sep if conv.sep_style != SeparatorStyle.TWO else conv.sep2
+        keywords = [stop_str]
+        stopping_criteria = KeywordsStoppingCriteria(keywords, tokenizer, input_ids)
+
+
+        with torch.inference_mode():
+            output_ids = model.generate(
+                input_ids,
+                images=image_tensor,
+                image_sizes=image_sizes,
+                patch_images=patch_images,
+                ind_tokens=ind_tokens,
+                do_sample=True if args.temperature > 0 else False,
+                temperature=args.temperature,
+                top_p=args.top_p,
+                num_beams=args.num_beams,
+                max_new_tokens=args.max_new_tokens,
+                use_cache=True)
+
+        outputs = tokenizer.batch_decode(output_ids, skip_special_tokens=True)[0]
+        outputs = outputs.strip()
+        if outputs.endswith(stop_str):
+            outputs = outputs[:-len(stop_str)]
+        outputs = outputs.strip()
+
+        ans_id = shortuuid.uuid()
+        ans_file.write(json.dumps({"question_id": idx,
+                                   "prompt": cur_prompt,
+                                   "text": outputs,
+                                   "answer_id": ans_id,
+                                   "model_id": model_name,
+                                   "metadata": {}}) + "\n")
+        # ans_file.flush()
+    ans_file.close()
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--model-path", type=str, default="facebook/opt-350m")
+    parser.add_argument("--model-base", type=str, default=None)
+    parser.add_argument("--image-folder", type=str, default="")
+    parser.add_argument("--question-file", type=str, default="tables/question.jsonl")
+    parser.add_argument("--answers-file", type=str, default="answer.jsonl")
+    parser.add_argument("--conv-mode", type=str, default="llava_v1")
+    parser.add_argument("--num-chunks", type=int, default=1)
+    parser.add_argument("--chunk-idx", type=int, default=0)
+    parser.add_argument("--temperature", type=float, default=0.2)
+    parser.add_argument("--top_p", type=float, default=None)
+    parser.add_argument("--num_beams", type=int, default=1)
+    parser.add_argument("--max_new_tokens", type=int, default=128)
+    parser.add_argument("--fted_encoder", type=bool, default=True)
+    args = parser.parse_args()
+
+    eval_model(args)

VLMEvalKit-sudoku/llava/eval/qa_baseline_gpt35.py

@@ -0,0 +1,74 @@
+"""Generate answers with GPT-3.5"""
+# Note: you need to be using OpenAI Python v0.27.0 for the code below to work
+import argparse
+import json
+import os
+import time
+import concurrent.futures
+
+import openai
+import tqdm
+import shortuuid
+
+MODEL = 'gpt-3.5-turbo'
+MODEL_ID = 'gpt-3.5-turbo:20230327'
+
+def get_answer(question_id: int, question: str, max_tokens: int):
+    ans = {
+        'answer_id': shortuuid.uuid(),
+        'question_id': question_id,
+        'model_id': MODEL_ID,
+    }
+    for _ in range(3):
+        try:
+            response = openai.ChatCompletion.create(
+                model=MODEL,
+                messages=[{
+                    'role': 'system',
+                    'content': 'You are a helpful assistant.'
+                }, {
+                    'role': 'user',
+                    'content': question,
+                }],
+                max_tokens=max_tokens,
+            )
+            ans['text'] = response['choices'][0]['message']['content']
+            return ans
+        except Exception as e:
+            print('[ERROR]', e)
+            ans['text'] = '#ERROR#'
+            time.sleep(1)
+    return ans
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(description='ChatGPT answer generation.')
+    parser.add_argument('-q', '--question')
+    parser.add_argument('-o', '--output')
+    parser.add_argument('--max-tokens', type=int, default=1024, help='maximum number of tokens produced in the output')
+    args = parser.parse_args()
+
+    questions_dict = {}
+    with open(os.path.expanduser(args.question)) as f:
+        for line in f:
+            if not line:
+                continue
+            q = json.loads(line)
+            questions_dict[q['question_id']] = q['text']
+
+    answers = []
+
+    with concurrent.futures.ThreadPoolExecutor(max_workers=32) as executor:
+        futures = []
+        for qid, question in questions_dict.items():
+            future = executor.submit(get_answer, qid, question, args.max_tokens)
+            futures.append(future)
+
+        for future in tqdm.tqdm(concurrent.futures.as_completed(futures), total=len(futures)):
+            answers.append(future.result())
+
+    answers.sort(key=lambda x: x['question_id'])
+
+    with open(os.path.expanduser(args.output), 'w') as f:
+        table = [json.dumps(ans) for ans in answers]
+        f.write('\n'.join(table))

VLMEvalKit-sudoku/llava/model/__init__.py

@@ -0,0 +1,17 @@
+import os
+
+AVAILABLE_MODELS = {
+    "llava_llama": "LlavaLlamaForCausalLM, LlavaConfig",
+    "llava_qwen": "LlavaQwenForCausalLM, LlavaQwenConfig",
+    "llava_mistral": "LlavaMistralForCausalLM, LlavaMistralConfig",
+    "llava_mixtral": "LlavaMixtralForCausalLM, LlavaMixtralConfig",
+    "llava_qwen3": "LlavaQwen3ForCausalLM, LlavaQwen3Config"
+    # "llava_qwen_moe": "LlavaQwenMoeForCausalLM, LlavaQwenMoeConfig",    
+    # Add other models as needed
+}
+
+for model_name, model_classes in AVAILABLE_MODELS.items():
+    try:
+        exec(f"from .language_model.{model_name} import {model_classes}")
+    except Exception as e:
+        print(f"Failed to import {model_name} from llava.language_model.{model_name}. Error: {e}")

VLMEvalKit-sudoku/llava/model/builder.py

@@ -0,0 +1,211 @@
+#    Copyright 2023 Haotian Liu
+#
+#    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.
+
+
+import os
+import warnings
+import shutil
+
+from transformers import AutoTokenizer, AutoModelForCausalLM, AutoConfig, BitsAndBytesConfig
+import torch
+from llava.model import *
+from llava.constants import DEFAULT_IMAGE_PATCH_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN
+from llava.utils import rank0_print
+
+def load_pretrained_model(model_path, model_base, model_name, load_8bit=False, load_4bit=False, 
+                          device_map="auto", device="cuda", use_flash_attn=False, _args=None, overwrite_config=None, **kwargs):
+    kwargs = {"device_map": device_map, **kwargs}
+
+    if device != "cuda":
+        kwargs['device_map'] = {"": device}
+
+    if load_8bit:
+        kwargs['load_in_8bit'] = True
+    elif load_4bit:
+        kwargs['load_in_4bit'] = True
+        kwargs['quantization_config'] = BitsAndBytesConfig(
+            load_in_4bit=True,
+            bnb_4bit_compute_dtype=torch.float16,
+            bnb_4bit_use_double_quant=True,
+            bnb_4bit_quant_type='nf4'
+        )
+    else:
+        kwargs['torch_dtype'] = torch.float16
+
+    if use_flash_attn:
+        kwargs['attn_implementation'] = 'flash_attention_2'
+
+    if 'llava' in model_name.lower():
+        # Load LLaVA model
+        if 'lora' in model_name.lower() and model_base is None:
+            warnings.warn('There is `lora` in model name but no `model_base` is provided. If you are loading a LoRA model, please provide the `model_base` argument. Detailed instruction: https://github.com/haotian-liu/LLaVA#launch-a-model-worker-lora-weights-unmerged.')
+        if 'lora' in model_name.lower() and model_base is not None:
+            from llava.model.language_model.llava_llama import LlavaConfig
+            lora_cfg_pretrained = LlavaConfig.from_pretrained(model_path)
+            tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=False)
+            print('Loading LLaVA from base model...')
+            model = LlavaLlamaForCausalLM.from_pretrained(model_base, low_cpu_mem_usage=True, config=lora_cfg_pretrained, **kwargs)
+            token_num, tokem_dim = model.lm_head.out_features, model.lm_head.in_features
+            if model.lm_head.weight.shape[0] != token_num:
+                model.lm_head.weight = torch.nn.Parameter(torch.empty(token_num, tokem_dim, device=model.device, dtype=model.dtype))
+                model.model.embed_tokens.weight = torch.nn.Parameter(torch.empty(token_num, tokem_dim, device=model.device, dtype=model.dtype))
+
+            print('Loading additional LLaVA weights...')
+            if os.path.exists(os.path.join(model_path, 'non_lora_trainables.bin')):
+                non_lora_trainables = torch.load(os.path.join(model_path, 'non_lora_trainables.bin'), map_location='cpu')
+            else:
+                # this is probably from HF Hub
+                from huggingface_hub import hf_hub_download
+                def load_from_hf(repo_id, filename, subfolder=None):
+                    cache_file = hf_hub_download(
+                        repo_id=repo_id,
+                        filename=filename,
+                        subfolder=subfolder)
+                    return torch.load(cache_file, map_location='cpu')
+                non_lora_trainables = load_from_hf(model_path, 'non_lora_trainables.bin')
+            non_lora_trainables = {(k[11:] if k.startswith('base_model.') else k): v for k, v in non_lora_trainables.items()}
+            if any(k.startswith('model.model.') for k in non_lora_trainables):
+                non_lora_trainables = {(k[6:] if k.startswith('model.') else k): v for k, v in non_lora_trainables.items()}
+            model.load_state_dict(non_lora_trainables, strict=False)
+
+            from peft import PeftModel
+            print('Loading LoRA weights...')
+            model = PeftModel.from_pretrained(model, model_path)
+            print('Merging LoRA weights...')
+            model = model.merge_and_unload()
+            print('Model is loaded...')
+        elif model_base is not None:
+            # this may be mm projector only
+            print('Loading LLaVA from base model...')
+            if 'mpt' in model_name.lower():
+                if not os.path.isfile(os.path.join(model_path, 'configuration_mpt.py')):
+                    shutil.copyfile(os.path.join(model_base, 'configuration_mpt.py'), os.path.join(model_path, 'configuration_mpt.py'))
+                tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=True)
+                cfg_pretrained = AutoConfig.from_pretrained(model_path, trust_remote_code=True)
+                model = LlavaMptForCausalLM.from_pretrained(model_base, low_cpu_mem_usage=True, config=cfg_pretrained, **kwargs)
+            else:
+                tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=False)
+                cfg_pretrained = AutoConfig.from_pretrained(model_path)
+                model = LlavaLlamaForCausalLM.from_pretrained(model_base, low_cpu_mem_usage=True, config=cfg_pretrained, **kwargs)
+
+            mm_projector_weights = torch.load(os.path.join(model_path, 'mm_projector.bin'), map_location='cpu')
+            mm_projector_weights = {k: v.to(torch.float16) for k, v in mm_projector_weights.items()}
+            model.load_state_dict(mm_projector_weights, strict=False)
+        else:
+            if 'mpt' in model_name.lower():
+                tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=True)
+                model = LlavaMptForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, **kwargs)
+            elif 'mistral' in model_name.lower():
+                tokenizer = AutoTokenizer.from_pretrained(model_path)
+                model = LlavaMistralForCausalLM.from_pretrained(
+                    model_path,
+                    low_cpu_mem_usage=True,
+                    **kwargs
+                )
+            elif 'qwen3' in model_name.lower():
+                tokenizer = AutoTokenizer.from_pretrained(model_path)
+                from llava.model.language_model.llava_qwen3 import LlavaQwen3ForCausalLM
+                model = LlavaQwen3ForCausalLM.from_pretrained(
+                    model_path,
+                    low_cpu_mem_usage=True,
+                    **kwargs
+                )
+                model.to(torch.bfloat16)
+            elif "qwen" in model_name.lower() or "quyen" in model_name.lower():
+                attn_implementation="flash_attention_2"
+                tokenizer = AutoTokenizer.from_pretrained(model_path)
+                if "moe" in model_name.lower() or "A14B" in model_name.lower():
+                    from llava.model.language_model.llava_qwen_moe import LlavaQwenMoeConfig
+                    if overwrite_config is not None:
+                        llava_cfg = LlavaQwenMoeConfig.from_pretrained(model_path)
+                        rank0_print(f"Overwriting config with {overwrite_config}")
+                        for k, v in overwrite_config.items():
+                            setattr(llava_cfg, k, v)
+                        model = LlavaQwenMoeForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, attn_implementation=attn_implementation, config=llava_cfg, **kwargs)
+                    else:
+                        model = LlavaQwenMoeForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, attn_implementation=attn_implementation, **kwargs)
+
+                else:
+                    from llava.model.language_model.llava_qwen import LlavaQwenConfig
+                    if overwrite_config is not None:
+                        llava_cfg = LlavaQwenConfig.from_pretrained(model_path)
+                        rank0_print(f"Overwriting config with {overwrite_config}")
+                        for k, v in overwrite_config.items():
+                            setattr(llava_cfg, k, v)
+                        model = LlavaQwenForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, config=llava_cfg, **kwargs)
+                    else:
+                        model = LlavaQwenForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, **kwargs)
+
+                model.to(torch.bfloat16)
+            else:
+                tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False)
+                model = LlavaLlamaForCausalLM.from_pretrained(
+                    model_path,
+                    low_cpu_mem_usage=True,
+                    **kwargs
+                )
+                model.to(torch.bfloat16)
+    else:
+        # Load language model
+        if model_base is not None:
+            # PEFT model
+            from peft import PeftModel
+            tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=False)
+            model = AutoModelForCausalLM.from_pretrained(model_base, low_cpu_mem_usage=True, **kwargs)
+            print(f"Loading LoRA weights from {model_path}")
+            model = PeftModel.from_pretrained(model, model_path)
+            print(f"Merging weights")
+            model = model.merge_and_unload()
+            print('Convert to FP16...')
+            model.to(torch.float16)
+        else:
+            use_fast = False
+            if 'mpt' in model_name.lower():
+                tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=True)
+                model = AutoModelForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, trust_remote_code=True, **kwargs)
+            else:
+                tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False)
+                model = AutoModelForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, **kwargs)
+
+    image_processor = None
+
+    if 'llava' in model_name.lower():
+        mm_use_im_start_end = getattr(model.config, "mm_use_im_start_end", False)
+        mm_use_im_patch_token = getattr(model.config, "mm_use_im_patch_token", True)
+        if mm_use_im_patch_token:
+            tokenizer.add_tokens([DEFAULT_IMAGE_PATCH_TOKEN], special_tokens=True)
+        if mm_use_im_start_end:
+            tokenizer.add_tokens([DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN], special_tokens=True)
+        model.resize_token_embeddings(len(tokenizer))
+
+        vision_tower = model.get_vision_tower()
+        if not vision_tower.is_loaded:
+            if _args is not None and _args.fted_encoder == True:
+                vision_tower.load_model(device_map=device_map, model_path=model_path)
+            else:
+                vision_tower.load_model(device_map=device_map)
+        if device_map != 'auto':
+            vision_tower.to(device=device_map, dtype=torch.float16)
+        image_processor = vision_tower.image_processor
+
+    if hasattr(model.config, "max_sequence_length"):
+        context_len = model.config.max_sequence_length
+    elif hasattr(model.config, "max_position_embeddings"):
+        context_len = model.config.max_position_embeddings
+    elif hasattr(model.config, "tokenizer_model_max_length"):
+        context_len = model.config.tokenizer_model_max_length
+    else:
+        context_len = 2048
+
+    return tokenizer, model, image_processor, context_len

VLMEvalKit-sudoku/llava/model/multimodal_encoder/attn_res.py

@@ -0,0 +1,103 @@
+
+
+import torch
+import torch.nn as nn
+import re
+from functools import partial
+import numpy as np
+from torch.nn.init import trunc_normal_
+from torch.nn import functional as F
+import math
+from transformers.activations import ACT2FN
+from transformers.utils import (
+    is_flash_attn_2_available,
+    is_flash_attn_greater_or_equal_2_10,
+)
+if is_flash_attn_2_available():
+    from transformers.modeling_flash_attention_utils import _flash_attention_forward
+
+class TokenPacker(nn.Module):
+    is_causal = False
+    def __init__(
+            self,
+            embed_dim=1152,
+            intermediate_size=4304,
+            num_heads=16,
+            norm_layer=partial(nn.LayerNorm, eps=1e-6)
+    ):
+        super().__init__()
+        self.embed_dim = embed_dim
+        self.num_heads = num_heads
+        gamma_init_eps = 1e-5
+        layer_norm_eps = 1e-6
+        self.head_dim = self.embed_dim // self.num_heads
+        self.scale = self.head_dim**-0.5
+        
+        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)
+
+        self.ln_q = nn.LayerNorm(embed_dim, eps=layer_norm_eps)
+        self.ln_kv = nn.LayerNorm(embed_dim, eps=layer_norm_eps)
+        self.gamma1 = nn.Parameter(gamma_init_eps * torch.ones(embed_dim), requires_grad=True)
+
+        self.ln_ffn = nn.LayerNorm(embed_dim, eps=layer_norm_eps)
+        self.ffn = nn.Sequential(
+            nn.Linear(self.embed_dim, intermediate_size),
+            ACT2FN['gelu_pytorch_tanh'],
+            nn.Linear(intermediate_size, self.embed_dim),
+        )
+        self.gamma2 = nn.Parameter(gamma_init_eps * torch.ones(embed_dim), requires_grad=True)
+        self._flash_attn_uses_top_left_mask = not is_flash_attn_greater_or_equal_2_10()
+    
+    def _attn(self, query, key, value, attn_mask):
+        B, N_q, _ = query.shape
+        B, N_k, _ = key.shape
+        query_states = self.q_proj(query)
+        key_states = self.k_proj(key)
+        value_states = self.v_proj(value)
+
+        query_states = query_states.view(B, N_q, self.num_heads, self.head_dim)
+        key_states = key_states.view(B, N_k, self.num_heads, self.head_dim)
+        value_states = value_states.view(B, N_k, self.num_heads, self.head_dim)
+        attn_output = _flash_attention_forward(
+            query_states,
+            key_states,
+            value_states,
+            attn_mask,
+            N_q,
+            is_causal=self.is_causal,
+            use_top_left_mask=self._flash_attn_uses_top_left_mask,
+        )
+        attn_output = attn_output.reshape(B, N_q, self.embed_dim).contiguous()
+        attn_output = self.out_proj(attn_output)
+        return attn_output
+
+    def forward(self, x, x_res, query=None, key=None, value=None, attn_mask=None):
+        ### x_res是((h w) 4 c)
+        ### x是((h w) 1 c) 均值
+        if query is not None:
+            query = self.ln_q(query)
+        else:
+            query = self.ln_q(x)[:, None, :]
+        if key is not None:
+            key = key
+        else:
+            key = self.ln_kv(x_res)
+        if value is not None:
+            value = value 
+        else:
+            value = key
+
+        out = self._attn(
+            query,
+            key,
+            value,
+            attn_mask=attn_mask)[0]
+        x_res = out
+
+        x = x + self.gamma1 * x_res # qkv norm，算attn，然后乘以gamma1，然后add 残差
+        x = x + self.gamma2 * self.ffn(self.ln_ffn(x))
+        # 这个x这个avg是直通分支，通过gamma学习引入残差分量
+        return x, key, value

VLMEvalKit-sudoku/llava/model/multimodal_encoder/clip_encoder.py

@@ -0,0 +1,223 @@
+import torch
+import torch.nn as nn
+import os
+from safetensors import safe_open
+from llava.utils import rank0_print
+from transformers import CLIPVisionModel, CLIPImageProcessor, CLIPVisionConfig
+from llava.model.multimodal_encoder.adapt_clip_vision_model import AdaptCLIPVisionModel
+try:
+    from s2wrapper import forward as multiscale_forward
+except:
+    pass
+
+def load_vision_tower_values(model_path, device):
+    """
+    在给定的路径下查找所有 `.safetensors` 文件，加载它们，并返回 key 中包含 `vision_tower` 的权重值。
+
+    参数:
+    - model_path (str): Hugging Face 模型文件夹的路径。
+
+    返回:
+    - vision_tower_values (dict): 包含所有 `vision_tower` 相关的键和值的字典。
+    """
+    # 找到路径中的所有 `.safetensors` 文件
+    safetensor_files = [f for f in os.listdir(model_path) if f.endswith('.safetensors')]
+    
+    vision_tower_values = {}
+
+    # 遍历每个 `.safetensors` 文件
+    for safetensor_file in safetensor_files:
+        safetensor_path = os.path.join(model_path, safetensor_file)
+        
+        # 使用 safetensors 库打开并读取文件内容
+        with safe_open(safetensor_path, framework="pt", device=str(device)) as f:
+            for key in f.keys():
+                # 如果 key 中包含 `vision_tower`，将其加入结果字典
+                if 'vision_tower' in key:
+                    key_new = key.replace('model.vision_tower.vision_tower.', '')
+                    vision_tower_values[key_new] = f.get_tensor(key)
+
+    return vision_tower_values
+class CLIPVisionTower(nn.Module):
+    def __init__(self, vision_tower, args, delay_load=False):
+        super().__init__()
+
+        self.is_loaded = False
+
+        self.vision_tower_name = vision_tower
+        self.select_layer = args.mm_vision_select_layer
+        self.select_feature = getattr(args, "mm_vision_select_feature", "patch")
+
+        if not delay_load:
+            rank0_print(f"Loading vision tower: {vision_tower}")
+            self.load_model()
+        elif getattr(args, "unfreeze_mm_vision_tower", False):
+            # TODO: better detector is needed.
+            rank0_print(f"The checkpoint seems to contain `vision_tower` weights: `unfreeze_mm_vision_tower`: True.")
+            self.load_model()
+        elif hasattr(args, "mm_tunable_parts") and "mm_vision_tower" in args.mm_tunable_parts:
+            rank0_print(f"The checkpoint seems to contain `vision_tower` weights: `mm_tunable_parts` contains `mm_vision_tower`.")
+            self.load_model()
+        else:
+            self.cfg_only = CLIPVisionConfig.from_pretrained(self.vision_tower_name)
+
+    def load_model(self, device_map=None, model_path=None):
+        if self.is_loaded:
+            rank0_print("{} is already loaded, `load_model` called again, skipping.".format(self.vision_tower_name))
+            return
+
+        self.image_processor = CLIPImageProcessor.from_pretrained(self.vision_tower_name)
+        #self.vision_tower = CLIPVisionModel.from_pretrained(self.vision_tower_name, device_map=device_map)
+        
+        print('---------init adapt_vision_model---------')
+        self.vision_tower = AdaptCLIPVisionModel.from_pretrained(self.vision_tower_name, device_map=device_map)
+        if model_path is None:
+            print('---------from frozen ckpt---------')
+        else:
+            print('---------from ft ckpt---------')
+            vision_tower_values = load_vision_tower_values(model_path, self.vision_tower.device)
+            load_info = self.vision_tower.load_state_dict(vision_tower_values, strict=False)
+            print(f'load info: {load_info}')
+
+        self.vision_tower.requires_grad_(False)
+
+        self.is_loaded = True
+
+    def feature_select(self, image_forward_outs):
+        select_feature_type = self.select_feature
+
+        if self.select_feature in ["slicefour_patch", "slicefour_cls_patch"]:
+            select_every_k_layer = len(image_forward_outs.hidden_states) // 4
+            image_features = torch.cat([image_forward_outs.hidden_states[i] for i in range(select_every_k_layer + self.select_layer, len(image_forward_outs.hidden_states), select_every_k_layer)], dim=-1)
+            select_feature_type = select_feature_type.replace("slicefour_", "")
+        elif self.select_feature in ["slice_m25811_f6_patch", "slice_m25811_f6_cls_patch"]:
+            select_layers = [-2, -5, -8, -11, 6]
+            image_features = torch.cat([image_forward_outs.hidden_states[i] for i in select_layers], dim=-1)
+            select_feature_type = select_feature_type.replace("slice_m25811_f6_", "")
+        else:
+            image_features = image_forward_outs.hidden_states[self.select_layer]
+
+        if select_feature_type == "patch":
+            image_features = image_features[:, 1:]
+        elif select_feature_type == "cls_patch":
+            image_features = image_features
+        else:
+            raise ValueError(f"Unexpected select feature: {select_feature_type}")
+        return image_features
+
+    def forward(self, images, patch_sizes):
+        tgt_sizes = torch.tensor(patch_sizes, dtype=torch.long, device=images[0].device)
+
+        #FIXME the pooled_output here is incorrect for post_layernorm on padded features
+        image_forward_outs = self.vision_tower(images, tgt_sizes=tgt_sizes, output_hidden_states=True)
+        features = self.feature_select(image_forward_outs).to(images[0].dtype)
+
+        image_features = [] #list torch.Size([1, 1024, 25, 22])
+        for i in range(len(features)):
+            h, w = patch_sizes[i]
+            feature = features[i][:h * w, :].unsqueeze(0)
+            # feature = feature.permute(0, 2, 1)  #torch.Size([1, 1024, 25*22])
+            # feature = feature.unflatten(2, [h, w])  #torch.Size([1, 1024, 25, 22])
+            image_features.append(feature)
+
+        return image_features
+
+    def forward_uhd_v2(self, images, tgt_sizes): 
+        #FIXME the pooled_output here is incorrect for post_layernorm on padded features
+        image_forward_outs = self.vision_tower(images, tgt_sizes=tgt_sizes, output_hidden_states=True)
+        image_features = self.feature_select(image_forward_outs).to(images[0].dtype)
+        return image_features
+
+    @property
+    def dummy_feature(self):
+        return torch.zeros(1, self.hidden_size, device=self.device, dtype=self.dtype)
+
+    @property
+    def dtype(self):
+        return self.vision_tower.dtype
+
+    @property
+    def device(self):
+        return self.vision_tower.device
+
+    @property
+    def config(self):
+        if self.is_loaded:
+            return self.vision_tower.config
+        else:
+            return self.cfg_only
+
+    @property
+    def hidden_size(self):
+        _hidden_size = self.config.hidden_size
+        if "slicefour" in self.select_feature:
+            _hidden_size *= 4
+        if "slice_m25811_f6" in self.select_feature:
+            _hidden_size *= 5
+        return _hidden_size
+
+    @property
+    def num_patches_per_side(self):
+        return self.config.image_size // self.config.patch_size
+
+    @property
+    def num_patches(self):
+        _num_patches = (self.config.image_size // self.config.patch_size) ** 2
+        if "cls_patch" in self.select_feature:
+            _num_patches += 1
+        return _num_patches
+
+    @property
+    def image_size(self):
+        return self.config.image_size
+
+
+class CLIPVisionTowerS2(CLIPVisionTower):
+    def __init__(self, vision_tower, args, delay_load=False):
+
+        self.s2_scales = getattr(args, "s2_scales", "336,672,1008")
+        self.s2_scales = list(map(int, self.s2_scales.split(",")))
+        self.s2_scales.sort()
+        self.s2_split_size = self.s2_scales[0]
+        self.s2_image_size = self.s2_scales[-1]
+
+        super().__init__(vision_tower, args, delay_load)
+
+        # change resize/crop size in preprocessing to the largest image size in s2_scale
+        if not delay_load or getattr(args, "unfreeze_mm_vision_tower", False):
+            self.image_processor.size["shortest_edge"] = self.s2_image_size
+            self.image_processor.crop_size["height"] = self.image_processor.crop_size["width"] = self.s2_image_size
+
+    def load_model(self, device_map=None):
+        if self.is_loaded:
+            rank0_print("{} is already loaded, `load_model` called again, skipping.".format(self.vision_tower_name))
+            return
+
+        self.image_processor = CLIPImageProcessor.from_pretrained(self.vision_tower_name)
+        self.vision_tower = CLIPVisionModel.from_pretrained(self.vision_tower_name, device_map=device_map)
+        self.vision_tower.requires_grad_(False)
+
+        self.image_processor.size["shortest_edge"] = self.s2_image_size
+        self.image_processor.crop_size["height"] = self.image_processor.crop_size["width"] = self.s2_image_size
+
+        self.is_loaded = True
+
+    def forward_feature(self, images):
+        image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype), output_hidden_states=True)
+        image_features = self.feature_select(image_forward_outs).to(images.dtype)
+        return image_features
+
+    def forward(self, images):
+        if type(images) is list:
+            image_features = []
+            for image in images:
+                image_feature = multiscale_forward(self.forward_feature, image.unsqueeze(0), img_sizes=self.s2_scales, max_split_size=self.s2_split_size, split_forward=True)
+                image_features.append(image_feature)
+        else:
+            image_features = multiscale_forward(self.forward_feature, images, img_sizes=self.s2_scales, max_split_size=self.s2_split_size, split_forward=True)
+
+        return image_features
+
+    @property
+    def hidden_size(self):
+        return self.config.hidden_size * len(self.s2_scales)

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/factory.py

@@ -0,0 +1,528 @@
+import json
+import logging
+import os
+import pathlib
+import re
+from copy import deepcopy
+from pathlib import Path
+from typing import Optional, Tuple, Union, Dict, Any
+import torch
+
+try:
+    import deepspeed
+except ImportError:
+    deepspeed = None
+
+from .constants import OPENAI_DATASET_MEAN, OPENAI_DATASET_STD
+from .model import CLIP, CustomCLIP, convert_weights_to_lp, convert_to_custom_text_state_dict, get_cast_dtype
+from .openai import load_openai_model
+from .pretrained import is_pretrained_cfg, get_pretrained_cfg, download_pretrained, list_pretrained_tags_by_model
+from .transform import image_transform
+from .tokenizer import HFTokenizer, tokenize
+from .utils import resize_clip_pos_embed, resize_evaclip_pos_embed, resize_visual_pos_embed, resize_eva_pos_embed
+
+
+_MODEL_CONFIG_PATHS = [Path(__file__).parent / f"model_configs/"]
+_MODEL_CONFIGS = {}  # directory (model_name: config) of model architecture configs
+
+
+def _natural_key(string_):
+    return [int(s) if s.isdigit() else s for s in re.split(r"(\d+)", string_.lower())]
+
+
+def _rescan_model_configs():
+    global _MODEL_CONFIGS
+
+    config_ext = (".json",)
+    config_files = []
+    for config_path in _MODEL_CONFIG_PATHS:
+        if config_path.is_file() and config_path.suffix in config_ext:
+            config_files.append(config_path)
+        elif config_path.is_dir():
+            for ext in config_ext:
+                config_files.extend(config_path.glob(f"*{ext}"))
+
+    for cf in config_files:
+        with open(cf, "r", encoding="utf8") as f:
+            model_cfg = json.load(f)
+            if all(a in model_cfg for a in ("embed_dim", "vision_cfg", "text_cfg")):
+                _MODEL_CONFIGS[cf.stem] = model_cfg
+
+    _MODEL_CONFIGS = dict(sorted(_MODEL_CONFIGS.items(), key=lambda x: _natural_key(x[0])))
+
+
+_rescan_model_configs()  # initial populate of model config registry
+
+
+def list_models():
+    """enumerate available model architectures based on config files"""
+    return list(_MODEL_CONFIGS.keys())
+
+
+def add_model_config(path):
+    """add model config path or file and update registry"""
+    if not isinstance(path, Path):
+        path = Path(path)
+    _MODEL_CONFIG_PATHS.append(path)
+    _rescan_model_configs()
+
+
+def get_model_config(model_name):
+    if model_name in _MODEL_CONFIGS:
+        return deepcopy(_MODEL_CONFIGS[model_name])
+    else:
+        return None
+
+
+def get_tokenizer(model_name):
+    config = get_model_config(model_name)
+    tokenizer = HFTokenizer(config["text_cfg"]["hf_tokenizer_name"]) if "hf_tokenizer_name" in config["text_cfg"] else tokenize
+    return tokenizer
+
+
+# loading openai CLIP weights when is_openai=True for training
+def load_state_dict(checkpoint_path: str, map_location: str = "cpu", model_key: str = "model|module|state_dict", is_openai: bool = False, skip_list: list = []):
+    if is_openai:
+        model = torch.jit.load(checkpoint_path, map_location="cpu").eval()
+        state_dict = model.state_dict()
+        for key in ["input_resolution", "context_length", "vocab_size"]:
+            state_dict.pop(key, None)
+    else:
+        checkpoint = torch.load(checkpoint_path, map_location=map_location)
+        for mk in model_key.split("|"):
+            if isinstance(checkpoint, dict) and mk in checkpoint:
+                state_dict = checkpoint[mk]
+                break
+            else:
+                state_dict = checkpoint
+        if next(iter(state_dict.items()))[0].startswith("module"):
+            state_dict = {k[7:]: v for k, v in state_dict.items()}
+
+    for k in skip_list:
+        if k in list(state_dict.keys()):
+            logging.info(f"Removing key {k} from pretrained checkpoint")
+            del state_dict[k]
+
+    if os.getenv("RoPE") == "1":
+        for k in list(state_dict.keys()):
+            if "freqs_cos" in k or "freqs_sin" in k:
+                del state_dict[k]
+    return state_dict
+
+
+def load_checkpoint(model, checkpoint_path, model_key="model|module|state_dict", strict=True):
+    state_dict = load_state_dict(checkpoint_path, model_key=model_key, is_openai=False)
+    # detect old format and make compatible with new format
+    if "positional_embedding" in state_dict and not hasattr(model, "positional_embedding"):
+        state_dict = convert_to_custom_text_state_dict(state_dict)
+    if "text.logit_scale" in state_dict and hasattr(model, "logit_scale"):
+        state_dict["logit_scale"] = state_dict["text.logit_scale"]
+        del state_dict["text.logit_scale"]
+
+    # resize_clip_pos_embed for CLIP and open CLIP
+    if "visual.positional_embedding" in state_dict:
+        resize_clip_pos_embed(state_dict, model)
+    # specified to eva_vit_model
+    elif "visual.pos_embed" in state_dict:
+        resize_evaclip_pos_embed(state_dict, model)
+
+    # resize_clip_pos_embed(state_dict, model)
+    incompatible_keys = model.load_state_dict(state_dict, strict=strict)
+    logging.info(f"incompatible_keys.missing_keys: {incompatible_keys.missing_keys}")
+    return incompatible_keys
+
+
+def load_clip_visual_state_dict(checkpoint_path: str, map_location: str = "cpu", is_openai: bool = False, skip_list: list = []):
+    state_dict = load_state_dict(checkpoint_path, map_location=map_location, is_openai=is_openai, skip_list=skip_list)
+
+    for k in list(state_dict.keys()):
+        if not k.startswith("visual."):
+            del state_dict[k]
+    for k in list(state_dict.keys()):
+        if k.startswith("visual."):
+            new_k = k[7:]
+            state_dict[new_k] = state_dict[k]
+            del state_dict[k]
+    return state_dict
+
+
+def load_clip_text_state_dict(checkpoint_path: str, map_location: str = "cpu", is_openai: bool = False, skip_list: list = []):
+    state_dict = load_state_dict(checkpoint_path, map_location=map_location, is_openai=is_openai, skip_list=skip_list)
+
+    for k in list(state_dict.keys()):
+        if k.startswith("visual."):
+            del state_dict[k]
+    return state_dict
+
+
+def get_pretrained_tag(pretrained_model):
+    pretrained_model = pretrained_model.lower()
+    if "laion" in pretrained_model or "open_clip" in pretrained_model:
+        return "open_clip"
+    elif "openai" in pretrained_model:
+        return "clip"
+    elif "eva" in pretrained_model and "clip" in pretrained_model:
+        return "eva_clip"
+    else:
+        return "other"
+
+
+def load_zero_partitions(model, state_dict, is_deepspeed_zero3_enabled, pretrained_model_path, ignore_mismatched_sizes=False):
+    """
+    adept from pytorch lightning and transformers
+    with deepspeed.zero.Init():
+        model = MyModel()
+    state_dict = torch.load(model_path, map_location="cpu")
+    load_zero_partitions(model, prefix="")
+    """
+
+    # because zero3 puts placeholders in model params, this context
+    # manager gathers (unpartitions) the params of the current layer, then loads from
+    # the state dict and then re-partitions them again
+    model_state_dict = model.state_dict()
+    expected_keys = list(model_state_dict.keys())
+    loaded_keys = list(state_dict.keys())
+    missing_keys = list(set(expected_keys) - set(loaded_keys))
+    unexpected_keys = list(set(loaded_keys) - set(expected_keys))
+
+    # Mistmatched keys contains tuples key/shape1/shape2 of weights in the checkpoint that have a shape not
+    # matching the weights in the model.
+    mismatched_keys = []
+    if ignore_mismatched_sizes:
+        for checkpoint_key in loaded_keys:
+            model_key = checkpoint_key
+
+            if model_key in model_state_dict and state_dict[checkpoint_key].shape != model_state_dict[model_key].shape:
+                mismatched_keys.append((checkpoint_key, state_dict[checkpoint_key].shape, model_state_dict[model_key].shape))
+                del state_dict[checkpoint_key]
+    # copy state_dict so _load_from_state_dict can modify it
+    metadata = getattr(state_dict, "_metadata", None)
+    state_dict = state_dict.copy()
+    if metadata is not None:
+        state_dict._metadata = metadata
+
+    error_msgs = []
+
+    # PyTorch's `_load_from_state_dict` does not copy parameters in a module's descendants
+    # so we need to apply the function recursively.
+    def load(module, prefix=""):
+        local_metadata = {} if metadata is None else metadata.get(prefix[:-1], {})
+        args = (state_dict, prefix, local_metadata, True, [], [], error_msgs)
+        if is_deepspeed_zero3_enabled:
+            # because zero3 puts placeholders in model params, this context
+            # manager gathers (unpartitions) the params of the current layer, then loads from
+            # the state dict and then re-partitions them again
+            with deepspeed.zero.GatheredParameters(list(module.parameters(recurse=False)), modifier_rank=0):
+                if torch.distributed.get_rank() == 0:
+                    module._load_from_state_dict(*args)
+        else:
+            module._load_from_state_dict(*args)
+
+        for name, child in module._modules.items():
+            if child is not None:
+                load(child, prefix + name + ".")
+
+    # Make sure we are able to load base models as well as derived models (with heads)
+    start_prefix = ""
+    model_to_load = model
+    load(model_to_load, prefix=start_prefix)
+    del state_dict
+    if len(error_msgs) > 0:
+        error_msg = "\n\t".join(error_msgs)
+        if "size mismatch" in error_msg:
+            error_msg += "\n\tYou may consider adding `ignore_mismatched_sizes=True` in the model `from_pretrained` method."
+        raise RuntimeError(f"Error(s) in loading state_dict for {model.__class__.__name__}:\n\t{error_msg}")
+    if len(unexpected_keys) > 0:
+        logging.warning(
+            f"Some weights of the model checkpoint at {pretrained_model_path} were not used when"
+            f" initializing {model.__class__.__name__}: {unexpected_keys}\n- This IS expected if you are"
+            f" initializing {model.__class__.__name__} from the checkpoint of a model trained on another task or"
+            " with another architecture (e.g. initializing a BertForSequenceClassification model from a"
+            " BertForPreTraining model).\n- This IS NOT expected if you are initializing"
+            f" {model.__class__.__name__} from the checkpoint of a model that you expect to be exactly identical"
+            " (initializing a BertForSequenceClassification model from a BertForSequenceClassification model)."
+        )
+    else:
+        logging.info(f"All model checkpoint weights were used when initializing {model.__class__.__name__}.\n")
+    if len(missing_keys) > 0:
+        logging.warning(
+            f"Some weights of {model.__class__.__name__} were not initialized from the model checkpoint at"
+            f" {pretrained_model_path} and are newly initialized: {missing_keys}\nYou should probably"
+            " TRAIN this model on a down-stream task to be able to use it for predictions and inference."
+        )
+    elif len(mismatched_keys) == 0:
+        logging.info(
+            f"All the weights of {model.__class__.__name__} were initialized from the model checkpoint at"
+            f" {pretrained_model_path}.\nIf your task is similar to the task the model of the checkpoint"
+            f" was trained on, you can already use {model.__class__.__name__} for predictions without further"
+            " training."
+        )
+    if len(mismatched_keys) > 0:
+        mismatched_warning = "\n".join([f"- {key}: found shape {shape1} in the checkpoint and {shape2} in the model instantiated" for key, shape1, shape2 in mismatched_keys])
+        logging.warning(
+            f"Some weights of {model.__class__.__name__} were not initialized from the model checkpoint at"
+            f" {pretrained_model_path} and are newly initialized because the shapes did not"
+            f" match:\n{mismatched_warning}\nYou should probably TRAIN this model on a down-stream task to be able"
+            " to use it for predictions and inference."
+        )
+
+
+def load_pretrained_checkpoint(model, visual_checkpoint_path, text_checkpoint_path, strict=True, visual_model=None, text_model=None, model_key="model|module|state_dict", skip_list=[]):
+    visual_tag = get_pretrained_tag(visual_model)
+    text_tag = get_pretrained_tag(text_model)
+
+    logging.info(f"num of model state_dict keys: {len(model.state_dict().keys())}")
+    visual_incompatible_keys, text_incompatible_keys = None, None
+    if visual_checkpoint_path:
+        if visual_tag == "eva_clip" or visual_tag == "open_clip":
+            visual_state_dict = load_clip_visual_state_dict(visual_checkpoint_path, is_openai=False, skip_list=skip_list)
+        elif visual_tag == "clip":
+            visual_state_dict = load_clip_visual_state_dict(visual_checkpoint_path, is_openai=True, skip_list=skip_list)
+        else:
+            visual_state_dict = load_state_dict(visual_checkpoint_path, model_key=model_key, is_openai=False, skip_list=skip_list)
+
+        # resize_clip_pos_embed for CLIP and open CLIP
+        if "positional_embedding" in visual_state_dict:
+            resize_visual_pos_embed(visual_state_dict, model)
+        # specified to EVA model
+        elif "pos_embed" in visual_state_dict:
+            resize_eva_pos_embed(visual_state_dict, model)
+
+        visual_incompatible_keys = model.visual.load_state_dict(visual_state_dict, strict=strict)
+        logging.info(f"num of loaded visual_state_dict keys: {len(visual_state_dict.keys())}")
+        logging.info(f"visual_incompatible_keys.missing_keys: {visual_incompatible_keys.missing_keys}")
+
+    if text_checkpoint_path:
+        if text_tag == "eva_clip" or text_tag == "open_clip":
+            text_state_dict = load_clip_text_state_dict(text_checkpoint_path, is_openai=False, skip_list=skip_list)
+        elif text_tag == "clip":
+            text_state_dict = load_clip_text_state_dict(text_checkpoint_path, is_openai=True, skip_list=skip_list)
+        else:
+            text_state_dict = load_state_dict(visual_checkpoint_path, model_key=model_key, is_openai=False, skip_list=skip_list)
+
+        text_incompatible_keys = model.text.load_state_dict(text_state_dict, strict=strict)
+
+        logging.info(f"num of loaded text_state_dict keys: {len(text_state_dict.keys())}")
+        logging.info(f"text_incompatible_keys.missing_keys: {text_incompatible_keys.missing_keys}")
+
+    return visual_incompatible_keys, text_incompatible_keys
+
+
+def create_model(
+    model_name: str,
+    pretrained: Optional[str] = None,
+    precision: str = "fp32",
+    device: Union[str, torch.device] = "cpu",
+    jit: bool = False,
+    force_quick_gelu: bool = False,
+    force_custom_clip: bool = False,
+    force_patch_dropout: Optional[float] = None,
+    pretrained_image: str = "",
+    pretrained_text: str = "",
+    pretrained_hf: bool = True,
+    pretrained_visual_model: str = None,
+    pretrained_text_model: str = None,
+    cache_dir: Optional[str] = None,
+    skip_list: list = [],
+):
+    model_name = model_name.replace("/", "-")  # for callers using old naming with / in ViT names
+    if isinstance(device, str):
+        device = torch.device(device)
+
+    if pretrained and pretrained.lower() == "openai":
+        logging.info(f"Loading pretrained {model_name} from OpenAI.")
+        model = load_openai_model(
+            model_name,
+            precision=precision,
+            device=device,
+            jit=jit,
+            cache_dir=cache_dir,
+        )
+    else:
+        model_cfg = get_model_config(model_name)
+        if model_cfg is not None:
+            logging.info(f"Loaded {model_name} model config.")
+        else:
+            logging.error(f"Model config for {model_name} not found; available models {list_models()}.")
+            raise RuntimeError(f"Model config for {model_name} not found.")
+
+        if "rope" in model_cfg.get("vision_cfg", {}):
+            if model_cfg["vision_cfg"]["rope"]:
+                os.environ["RoPE"] = "1"
+        else:
+            os.environ["RoPE"] = "0"
+
+        if force_quick_gelu:
+            # override for use of QuickGELU on non-OpenAI transformer models
+            model_cfg["quick_gelu"] = True
+
+        if force_patch_dropout is not None:
+            # override the default patch dropout value
+            model_cfg["vision_cfg"]["patch_dropout"] = force_patch_dropout
+
+        cast_dtype = get_cast_dtype(precision)
+        custom_clip = model_cfg.pop("custom_text", False) or force_custom_clip or ("hf_model_name" in model_cfg["text_cfg"])
+
+        if custom_clip:
+            if "hf_model_name" in model_cfg.get("text_cfg", {}):
+                model_cfg["text_cfg"]["hf_model_pretrained"] = pretrained_hf
+            model = CustomCLIP(**model_cfg, cast_dtype=cast_dtype)
+        else:
+            model = CLIP(**model_cfg, cast_dtype=cast_dtype)
+
+        pretrained_cfg = {}
+        if pretrained:
+            checkpoint_path = ""
+            pretrained_cfg = get_pretrained_cfg(model_name, pretrained)
+            if pretrained_cfg:
+                checkpoint_path = download_pretrained(pretrained_cfg, cache_dir=cache_dir)
+            elif os.path.exists(pretrained):
+                checkpoint_path = pretrained
+
+            if checkpoint_path:
+                logging.info(f"Loading pretrained {model_name} weights ({pretrained}).")
+                load_checkpoint(model, checkpoint_path, model_key="model|module|state_dict", strict=False)
+            else:
+                error_str = f"Pretrained weights ({pretrained}) not found for model {model_name}." f"Available pretrained tags ({list_pretrained_tags_by_model(model_name)}."
+                logging.warning(error_str)
+                raise RuntimeError(error_str)
+        else:
+            visual_checkpoint_path = ""
+            text_checkpoint_path = ""
+
+            if pretrained_image:
+                pretrained_visual_model = pretrained_visual_model.replace("/", "-")  # for callers using old naming with / in ViT names
+                pretrained_image_cfg = get_pretrained_cfg(pretrained_visual_model, pretrained_image)
+                if "timm_model_name" in model_cfg.get("vision_cfg", {}):
+                    # pretrained weight loading for timm models set via vision_cfg
+                    model_cfg["vision_cfg"]["timm_model_pretrained"] = True
+                elif pretrained_image_cfg:
+                    visual_checkpoint_path = download_pretrained(pretrained_image_cfg, cache_dir=cache_dir)
+                elif os.path.exists(pretrained_image):
+                    visual_checkpoint_path = pretrained_image
+                else:
+                    logging.warning(f"Pretrained weights ({visual_checkpoint_path}) not found for model {model_name}.visual.")
+                    raise RuntimeError(f"Pretrained weights ({visual_checkpoint_path}) not found for model {model_name}.visual.")
+
+            if pretrained_text:
+                pretrained_text_model = pretrained_text_model.replace("/", "-")  # for callers using old naming with / in ViT names
+                pretrained_text_cfg = get_pretrained_cfg(pretrained_text_model, pretrained_text)
+                if pretrained_image_cfg:
+                    text_checkpoint_path = download_pretrained(pretrained_text_cfg, cache_dir=cache_dir)
+                elif os.path.exists(pretrained_text):
+                    text_checkpoint_path = pretrained_text
+                else:
+                    logging.warning(f"Pretrained weights ({text_checkpoint_path}) not found for model {model_name}.text.")
+                    raise RuntimeError(f"Pretrained weights ({text_checkpoint_path}) not found for model {model_name}.text.")
+
+            if visual_checkpoint_path:
+                logging.info(f"Loading pretrained {model_name}.visual weights ({visual_checkpoint_path}).")
+            if text_checkpoint_path:
+                logging.info(f"Loading pretrained {model_name}.text weights ({text_checkpoint_path}).")
+
+            if visual_checkpoint_path or text_checkpoint_path:
+                load_pretrained_checkpoint(model, visual_checkpoint_path, text_checkpoint_path, strict=False, visual_model=pretrained_visual_model, text_model=pretrained_text_model, model_key="model|module|state_dict", skip_list=skip_list)
+
+        if "fp16" in precision or "bf16" in precision:
+            logging.info(f"convert precision to {precision}")
+            model = model.to(torch.bfloat16) if "bf16" in precision else model.to(torch.float16)
+
+        # model.to(device=device)
+
+        # set image / mean metadata from pretrained_cfg if available, or use default
+        model.visual.image_mean = pretrained_cfg.get("mean", None) or OPENAI_DATASET_MEAN
+        model.visual.image_std = pretrained_cfg.get("std", None) or OPENAI_DATASET_STD
+
+        if jit:
+            model = torch.jit.script(model)
+
+    return model
+
+
+def create_model_and_transforms(
+    model_name: str,
+    pretrained: Optional[str] = None,
+    precision: str = "fp32",
+    device: Union[str, torch.device] = "cpu",
+    jit: bool = False,
+    force_quick_gelu: bool = False,
+    force_custom_clip: bool = False,
+    force_patch_dropout: Optional[float] = None,
+    pretrained_image: str = "",
+    pretrained_text: str = "",
+    pretrained_hf: bool = True,
+    pretrained_visual_model: str = None,
+    pretrained_text_model: str = None,
+    image_mean: Optional[Tuple[float, ...]] = None,
+    image_std: Optional[Tuple[float, ...]] = None,
+    cache_dir: Optional[str] = None,
+    skip_list: list = [],
+):
+    model = create_model(
+        model_name,
+        pretrained,
+        precision=precision,
+        device=device,
+        jit=jit,
+        force_quick_gelu=force_quick_gelu,
+        force_custom_clip=force_custom_clip,
+        force_patch_dropout=force_patch_dropout,
+        pretrained_image=pretrained_image,
+        pretrained_text=pretrained_text,
+        pretrained_hf=pretrained_hf,
+        pretrained_visual_model=pretrained_visual_model,
+        pretrained_text_model=pretrained_text_model,
+        cache_dir=cache_dir,
+        skip_list=skip_list,
+    )
+
+    image_mean = image_mean or getattr(model.visual, "image_mean", None)
+    image_std = image_std or getattr(model.visual, "image_std", None)
+    preprocess_train = image_transform(model.visual.image_size, is_train=True, mean=image_mean, std=image_std)
+    preprocess_val = image_transform(model.visual.image_size, is_train=False, mean=image_mean, std=image_std)
+
+    return model, preprocess_train, preprocess_val
+
+
+def create_model_from_pretrained(
+    model_name: str,
+    pretrained: str,
+    precision: str = "fp32",
+    device: Union[str, torch.device] = "cpu",
+    jit: bool = False,
+    force_quick_gelu: bool = False,
+    force_custom_clip: bool = False,
+    force_patch_dropout: Optional[float] = None,
+    return_transform: bool = True,
+    image_mean: Optional[Tuple[float, ...]] = None,
+    image_std: Optional[Tuple[float, ...]] = None,
+    cache_dir: Optional[str] = None,
+    is_frozen: bool = False,
+):
+    if not is_pretrained_cfg(model_name, pretrained) and not os.path.exists(pretrained):
+        raise RuntimeError(f"{pretrained} is not a valid pretrained cfg or checkpoint for {model_name}." f" Use open_clip.list_pretrained() to find one.")
+
+    model = create_model(
+        model_name,
+        pretrained,
+        precision=precision,
+        device=device,
+        jit=jit,
+        force_quick_gelu=force_quick_gelu,
+        force_custom_clip=force_custom_clip,
+        force_patch_dropout=force_patch_dropout,
+        cache_dir=cache_dir,
+    )
+
+    if is_frozen:
+        for param in model.parameters():
+            param.requires_grad = False
+
+    if not return_transform:
+        return model
+
+    image_mean = image_mean or getattr(model.visual, "image_mean", None)
+    image_std = image_std or getattr(model.visual, "image_std", None)
+    preprocess = image_transform(model.visual.image_size, is_train=False, mean=image_mean, std=image_std)
+
+    return model, preprocess

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/loss.py

@@ -0,0 +1,123 @@
+import math
+import torch
+import torch.nn as nn
+from torch.nn import functional as F
+
+try:
+    import torch.distributed.nn
+    from torch import distributed as dist
+
+    has_distributed = True
+except ImportError:
+    has_distributed = False
+
+try:
+    import horovod.torch as hvd
+except ImportError:
+    hvd = None
+
+from timm.loss import LabelSmoothingCrossEntropy
+
+
+def gather_features(image_features, text_features, local_loss=False, gather_with_grad=False, rank=0, world_size=1, use_horovod=False):
+    assert has_distributed, "torch.distributed did not import correctly, please use a PyTorch version with support."
+    if use_horovod:
+        assert hvd is not None, "Please install horovod"
+        if gather_with_grad:
+            all_image_features = hvd.allgather(image_features)
+            all_text_features = hvd.allgather(text_features)
+        else:
+            with torch.no_grad():
+                all_image_features = hvd.allgather(image_features)
+                all_text_features = hvd.allgather(text_features)
+            if not local_loss:
+                # ensure grads for local rank when all_* features don't have a gradient
+                gathered_image_features = list(all_image_features.chunk(world_size, dim=0))
+                gathered_text_features = list(all_text_features.chunk(world_size, dim=0))
+                gathered_image_features[rank] = image_features
+                gathered_text_features[rank] = text_features
+                all_image_features = torch.cat(gathered_image_features, dim=0)
+                all_text_features = torch.cat(gathered_text_features, dim=0)
+    else:
+        # We gather tensors from all gpus
+        if gather_with_grad:
+            all_image_features = torch.cat(torch.distributed.nn.all_gather(image_features), dim=0)
+            all_text_features = torch.cat(torch.distributed.nn.all_gather(text_features), dim=0)
+            # all_image_features = torch.cat(torch.distributed.nn.all_gather(image_features, async_op=True), dim=0)
+            # all_text_features = torch.cat(torch.distributed.nn.all_gather(text_features, async_op=True), dim=0)
+        else:
+            gathered_image_features = [torch.zeros_like(image_features) for _ in range(world_size)]
+            gathered_text_features = [torch.zeros_like(text_features) for _ in range(world_size)]
+            dist.all_gather(gathered_image_features, image_features)
+            dist.all_gather(gathered_text_features, text_features)
+            if not local_loss:
+                # ensure grads for local rank when all_* features don't have a gradient
+                gathered_image_features[rank] = image_features
+                gathered_text_features[rank] = text_features
+            all_image_features = torch.cat(gathered_image_features, dim=0)
+            all_text_features = torch.cat(gathered_text_features, dim=0)
+
+    return all_image_features, all_text_features
+
+
+class ClipLoss(nn.Module):
+
+    def __init__(
+        self,
+        local_loss=False,
+        gather_with_grad=False,
+        cache_labels=False,
+        rank=0,
+        world_size=1,
+        use_horovod=False,
+        smoothing=0.0,
+    ):
+        super().__init__()
+        self.local_loss = local_loss
+        self.gather_with_grad = gather_with_grad
+        self.cache_labels = cache_labels
+        self.rank = rank
+        self.world_size = world_size
+        self.use_horovod = use_horovod
+        self.label_smoothing_cross_entropy = LabelSmoothingCrossEntropy(smoothing=smoothing) if smoothing > 0 else None
+
+        # cache state
+        self.prev_num_logits = 0
+        self.labels = {}
+
+    def forward(self, image_features, text_features, logit_scale=1.0):
+        device = image_features.device
+        if self.world_size > 1:
+            all_image_features, all_text_features = gather_features(image_features, text_features, self.local_loss, self.gather_with_grad, self.rank, self.world_size, self.use_horovod)
+
+            if self.local_loss:
+                logits_per_image = logit_scale * image_features @ all_text_features.T
+                logits_per_text = logit_scale * text_features @ all_image_features.T
+            else:
+                logits_per_image = logit_scale * all_image_features @ all_text_features.T
+                logits_per_text = logits_per_image.T
+        else:
+            logits_per_image = logit_scale * image_features @ text_features.T
+            logits_per_text = logit_scale * text_features @ image_features.T
+        # calculated ground-truth and cache if enabled
+        num_logits = logits_per_image.shape[0]
+        if self.prev_num_logits != num_logits or device not in self.labels:
+            labels = torch.arange(num_logits, device=device, dtype=torch.long)
+            if self.world_size > 1 and self.local_loss:
+                labels = labels + num_logits * self.rank
+            if self.cache_labels:
+                self.labels[device] = labels
+                self.prev_num_logits = num_logits
+        else:
+            labels = self.labels[device]
+
+        if self.label_smoothing_cross_entropy:
+            total_loss = (self.label_smoothing_cross_entropy(logits_per_image, labels) + self.label_smoothing_cross_entropy(logits_per_text, labels)) / 2
+        else:
+            total_loss = (F.cross_entropy(logits_per_image, labels) + F.cross_entropy(logits_per_text, labels)) / 2
+
+        acc = None
+        i2t_acc = (logits_per_image.argmax(-1) == labels).sum() / len(logits_per_image)
+        t2i_acc = (logits_per_text.argmax(-1) == labels).sum() / len(logits_per_text)
+        acc = {"i2t": i2t_acc, "t2i": t2i_acc}
+        return total_loss, acc

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/model_configs/EVA-CLIP-18B.json

@@ -0,0 +1,27 @@
+{
+    "embed_dim": 1536,
+    "vision_cfg": {
+        "image_size": 224,
+        "layers": 48,
+        "width": 5120,
+        "head_width": 128,
+        "mlp_ratio": 5,
+        "patch_size": 14,
+        "eva_model_name": "eva-clip-18b-14-x",
+        "drop_path_rate": 0,
+        "qkv_bias": false,
+        "xattn": true,
+        "postnorm": true,
+        "fusedLN": false,
+        "use_rms_norm": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 1280,
+        "heads": 20,
+        "layers": 32,
+        "xattn": false,
+        "fusedLN": false
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/model_configs/EVA-CLIP-8B-plus.json

@@ -0,0 +1,27 @@
+{
+    "embed_dim": 1280,
+    "vision_cfg": {
+        "image_size": 448,
+        "layers": 32,
+        "width": 4096,
+        "head_width": 128,
+        "mlp_ratio": 5,
+        "patch_size": 14,
+        "eva_model_name": "eva-clip-8b-14-plus-x",
+        "drop_path_rate": 0,
+        "qkv_bias": false,
+        "xattn": true,
+        "postnorm": false,
+        "fusedLN": false,
+        "use_rms_norm": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 1280,
+        "heads": 20,
+        "layers": 32,
+        "xattn": false,
+        "fusedLN": false
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/model_configs/EVA01-CLIP-g-14.json

@@ -0,0 +1,24 @@
+{
+    "embed_dim": 1024,
+    "vision_cfg": {
+        "image_size": 224,
+        "layers": 40,
+        "width": 1408,
+        "head_width": 88,
+        "mlp_ratio": 4.3637,
+        "patch_size": 14,
+        "eva_model_name": "eva-clip-g-14-x",
+        "drop_path_rate": 0.4,
+        "xattn": true,
+        "fusedLN": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 768,
+        "heads": 12,
+        "layers": 12,
+        "xattn": false,
+        "fusedLN": true
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/model_configs/EVA02-CLIP-L-14-336.json

@@ -0,0 +1,29 @@
+{
+    "embed_dim": 768,
+    "vision_cfg": {
+        "image_size": 336,
+        "layers": 24,
+        "width": 1024,
+        "drop_path_rate": 0,
+        "head_width": 64,
+        "mlp_ratio": 2.6667,
+        "patch_size": 14,
+        "eva_model_name": "eva-clip-l-14-336",
+        "xattn": true,
+        "fusedLN": true,
+        "rope": true,
+        "pt_hw_seq_len": 16,
+        "intp_freq": true,
+        "naiveswiglu": true,
+        "subln": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 768,
+        "heads": 12,
+        "layers": 12,
+        "xattn": false,
+        "fusedLN": true
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/model_configs/EVA02-CLIP-bigE-14.json

@@ -0,0 +1,25 @@
+{
+    "embed_dim": 1024,
+    "vision_cfg": {
+        "image_size": 224,
+        "layers": 64,
+        "width": 1792,
+        "head_width": 112,
+        "mlp_ratio": 8.571428571428571,
+        "patch_size": 14,
+        "eva_model_name": "eva-clip-4b-14-x",
+        "drop_path_rate": 0,
+        "xattn": true,
+        "postnorm": true,
+        "fusedLN": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 1024,
+        "heads": 16,
+        "layers": 24,
+        "xattn": false,
+        "fusedLN": true
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/model_configs/Internal-EVA02-CLIP-10B-14-448.json

@@ -0,0 +1,25 @@
+{
+    "embed_dim": 1024,
+    "vision_cfg": {
+        "image_size": 448,
+        "layers": 77,
+        "width": 2304,
+        "head_width": 144,
+        "mlp_ratio": 10.9722,
+        "patch_size": 14,
+        "eva_model_name": "eva-clip-10b-14-x",
+        "drop_path_rate": 0,
+        "xattn": true,
+        "postnorm": false,
+        "fusedLN": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 1280,
+        "heads": 20,
+        "layers": 32,
+        "xattn": false,
+        "fusedLN": true
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/model_configs/Internal-EVA02-CLIP-10B-14.json

@@ -0,0 +1,25 @@
+{
+    "embed_dim": 1024,
+    "vision_cfg": {
+        "image_size": 224,
+        "layers": 77,
+        "width": 2304,
+        "head_width": 144,
+        "mlp_ratio": 10.9722,
+        "patch_size": 14,
+        "eva_model_name": "eva-clip-10b-14-x",
+        "drop_path_rate": 0,
+        "xattn": true,
+        "postnorm": false,
+        "fusedLN": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 1280,
+        "heads": 20,
+        "layers": 32,
+        "xattn": false,
+        "fusedLN": true
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/openai.py

@@ -0,0 +1,144 @@
+""" OpenAI pretrained model functions
+
+Adapted from https://github.com/openai/CLIP. Originally MIT License, Copyright (c) 2021 OpenAI.
+"""
+
+import os
+import warnings
+from typing import List, Optional, Union
+
+import torch
+
+from .model import build_model_from_openai_state_dict, convert_weights_to_lp, get_cast_dtype
+from .pretrained import get_pretrained_url, list_pretrained_models_by_tag, download_pretrained_from_url
+
+__all__ = ["list_openai_models", "load_openai_model"]
+
+
+def list_openai_models() -> List[str]:
+    """Returns the names of available CLIP models"""
+    return list_pretrained_models_by_tag("openai")
+
+
+def load_openai_model(
+    name: str,
+    precision: Optional[str] = None,
+    device: Optional[Union[str, torch.device]] = None,
+    jit: bool = True,
+    cache_dir: Optional[str] = None,
+):
+    """Load a CLIP model
+
+    Parameters
+    ----------
+    name : str
+        A model name listed by `clip.available_models()`, or the path to a model checkpoint containing the state_dict
+    precision: str
+        Model precision, if None defaults to 'fp32' if device == 'cpu' else 'fp16'.
+    device : Union[str, torch.device]
+        The device to put the loaded model
+    jit : bool
+        Whether to load the optimized JIT model (default) or more hackable non-JIT model.
+    cache_dir : Optional[str]
+        The directory to cache the downloaded model weights
+
+    Returns
+    -------
+    model : torch.nn.Module
+        The CLIP model
+    preprocess : Callable[[PIL.Image], torch.Tensor]
+        A torchvision transform that converts a PIL image into a tensor that the returned model can take as its input
+    """
+    if device is None:
+        device = "cuda" if torch.cuda.is_available() else "cpu"
+    if precision is None:
+        precision = "fp32" if device == "cpu" else "fp16"
+
+    if get_pretrained_url(name, "openai"):
+        model_path = download_pretrained_from_url(get_pretrained_url(name, "openai"), cache_dir=cache_dir)
+    elif os.path.isfile(name):
+        model_path = name
+    else:
+        raise RuntimeError(f"Model {name} not found; available models = {list_openai_models()}")
+
+    try:
+        # loading JIT archive
+        model = torch.jit.load(model_path, map_location=device if jit else "cpu").eval()
+        state_dict = None
+    except RuntimeError:
+        # loading saved state dict
+        if jit:
+            warnings.warn(f"File {model_path} is not a JIT archive. Loading as a state dict instead")
+            jit = False
+        state_dict = torch.load(model_path, map_location="cpu")
+
+    if not jit:
+        # Build a non-jit model from the OpenAI jitted model state dict
+        cast_dtype = get_cast_dtype(precision)
+        try:
+            model = build_model_from_openai_state_dict(state_dict or model.state_dict(), cast_dtype=cast_dtype)
+        except KeyError:
+            sd = {k[7:]: v for k, v in state_dict["state_dict"].items()}
+            model = build_model_from_openai_state_dict(sd, cast_dtype=cast_dtype)
+
+        # model from OpenAI state dict is in manually cast fp16 mode, must be converted for AMP/fp32/bf16 use
+        model = model.to(device)
+        if precision.startswith("amp") or precision == "fp32":
+            model.float()
+        elif precision == "bf16":
+            convert_weights_to_lp(model, dtype=torch.bfloat16)
+
+        return model
+
+    # patch the device names
+    device_holder = torch.jit.trace(lambda: torch.ones([]).to(torch.device(device)), example_inputs=[])
+    device_node = [n for n in device_holder.graph.findAllNodes("prim::Constant") if "Device" in repr(n)][-1]
+
+    def patch_device(module):
+        try:
+            graphs = [module.graph] if hasattr(module, "graph") else []
+        except RuntimeError:
+            graphs = []
+
+        if hasattr(module, "forward1"):
+            graphs.append(module.forward1.graph)
+
+        for graph in graphs:
+            for node in graph.findAllNodes("prim::Constant"):
+                if "value" in node.attributeNames() and str(node["value"]).startswith("cuda"):
+                    node.copyAttributes(device_node)
+
+    model.apply(patch_device)
+    patch_device(model.encode_image)
+    patch_device(model.encode_text)
+
+    # patch dtype to float32 (typically for CPU)
+    if precision == "fp32":
+        float_holder = torch.jit.trace(lambda: torch.ones([]).float(), example_inputs=[])
+        float_input = list(float_holder.graph.findNode("aten::to").inputs())[1]
+        float_node = float_input.node()
+
+        def patch_float(module):
+            try:
+                graphs = [module.graph] if hasattr(module, "graph") else []
+            except RuntimeError:
+                graphs = []
+
+            if hasattr(module, "forward1"):
+                graphs.append(module.forward1.graph)
+
+            for graph in graphs:
+                for node in graph.findAllNodes("aten::to"):
+                    inputs = list(node.inputs())
+                    for i in [1, 2]:  # dtype can be the second or third argument to aten::to()
+                        if inputs[i].node()["value"] == 5:
+                            inputs[i].node().copyAttributes(float_node)
+
+        model.apply(patch_float)
+        patch_float(model.encode_image)
+        patch_float(model.encode_text)
+        model.float()
+
+    # ensure image_size attr available at consistent location for both jit and non-jit
+    model.visual.image_size = model.input_resolution.item()
+    return model

VLMEvalKit-sudoku/llava/model/multimodal_encoder/dev_eva_clip/eva_clip/timm_model.py

@@ -0,0 +1,114 @@
+""" timm model adapter
+
+Wraps timm (https://github.com/rwightman/pytorch-image-models) models for use as a vision tower in CLIP model.
+"""
+
+import logging
+from collections import OrderedDict
+
+import torch
+import torch.nn as nn
+
+try:
+    import timm
+    from timm.models.layers import Mlp, to_2tuple
+
+    try:
+        # old timm imports < 0.8.1
+        from timm.models.layers.attention_pool2d import RotAttentionPool2d
+        from timm.models.layers.attention_pool2d import AttentionPool2d as AbsAttentionPool2d
+    except ImportError:
+        # new timm imports >= 0.8.1
+        from timm.layers import RotAttentionPool2d
+        from timm.layers import AttentionPool2d as AbsAttentionPool2d
+except ImportError:
+    timm = None
+
+from .utils import freeze_batch_norm_2d
+
+
+class TimmModel(nn.Module):
+    """timm model adapter
+    # FIXME this adapter is a work in progress, may change in ways that break weight compat
+    """
+
+    def __init__(self, model_name, embed_dim, image_size=224, pool="avg", proj="linear", proj_bias=False, drop=0.0, pretrained=False):
+        super().__init__()
+        if timm is None:
+            raise RuntimeError("Please `pip install timm` to use timm models.")
+
+        self.image_size = to_2tuple(image_size)
+        self.trunk = timm.create_model(model_name, pretrained=pretrained)
+        feat_size = self.trunk.default_cfg.get("pool_size", None)
+        feature_ndim = 1 if not feat_size else 2
+        if pool in ("abs_attn", "rot_attn"):
+            assert feature_ndim == 2
+            # if attn pooling used, remove both classifier and default pool
+            self.trunk.reset_classifier(0, global_pool="")
+        else:
+            # reset global pool if pool config set, otherwise leave as network default
+            reset_kwargs = dict(global_pool=pool) if pool else {}
+            self.trunk.reset_classifier(0, **reset_kwargs)
+        prev_chs = self.trunk.num_features
+
+        head_layers = OrderedDict()
+        if pool == "abs_attn":
+            head_layers["pool"] = AbsAttentionPool2d(prev_chs, feat_size=feat_size, out_features=embed_dim)
+            prev_chs = embed_dim
+        elif pool == "rot_attn":
+            head_layers["pool"] = RotAttentionPool2d(prev_chs, out_features=embed_dim)
+            prev_chs = embed_dim
+        else:
+            assert proj, "projection layer needed if non-attention pooling is used."
+
+        # NOTE attention pool ends with a projection layer, so proj should usually be set to '' if such pooling is used
+        if proj == "linear":
+            head_layers["drop"] = nn.Dropout(drop)
+            head_layers["proj"] = nn.Linear(prev_chs, embed_dim, bias=proj_bias)
+        elif proj == "mlp":
+            head_layers["mlp"] = Mlp(prev_chs, 2 * embed_dim, embed_dim, drop=drop, bias=(True, proj_bias))
+
+        self.head = nn.Sequential(head_layers)
+
+    def lock(self, unlocked_groups=0, freeze_bn_stats=False):
+        """lock modules
+        Args:
+            unlocked_groups (int): leave last n layer groups unlocked (default: 0)
+        """
+        if not unlocked_groups:
+            # lock full model
+            for param in self.trunk.parameters():
+                param.requires_grad = False
+            if freeze_bn_stats:
+                freeze_batch_norm_2d(self.trunk)
+        else:
+            # NOTE: partial freeze requires latest timm (master) branch and is subject to change
+            try:
+                # FIXME import here until API stable and in an official release
+                from timm.models.helpers import group_parameters, group_modules
+            except ImportError:
+                raise RuntimeError("Please install latest timm `pip install git+https://github.com/rwightman/pytorch-image-models`")
+            matcher = self.trunk.group_matcher()
+            gparams = group_parameters(self.trunk, matcher)
+            max_layer_id = max(gparams.keys())
+            max_layer_id = max_layer_id - unlocked_groups
+            for group_idx in range(max_layer_id + 1):
+                group = gparams[group_idx]
+                for param in group:
+                    self.trunk.get_parameter(param).requires_grad = False
+            if freeze_bn_stats:
+                gmodules = group_modules(self.trunk, matcher, reverse=True)
+                gmodules = {k for k, v in gmodules.items() if v <= max_layer_id}
+                freeze_batch_norm_2d(self.trunk, gmodules)
+
+    @torch.jit.ignore
+    def set_grad_checkpointing(self, enable=True):
+        try:
+            self.trunk.set_grad_checkpointing(enable)
+        except Exception as e:
+            logging.warning("grad checkpointing not supported for this timm image tower, continuing without...")
+
+    def forward(self, x):
+        x = self.trunk(x)
+        x = self.head(x)
+        return x

VLMEvalKit-sudoku/llava/model/multimodal_encoder/eva_clip/eva_clip_encoder.py

@@ -0,0 +1,74 @@
+import torch
+import torch.nn as nn
+
+from .eva_clip_processors import EvaClipImageTrainProcessor
+from .eva_vit import EVAEncoderWrapper
+from .factory import list_models, add_model_config, get_model_config
+
+from llava.utils import rank0_print
+
+
+class EvaClipVisionTower(nn.Module):
+    def __init__(self, vision_tower, args, delay_load=False):
+        super().__init__()
+
+        self.is_loaded = False
+        self.vision_tower_name = vision_tower
+        self.vision_tower_pretrained = args.vision_tower_pretrained
+        self.config = get_model_config(vision_tower)
+
+        if not delay_load:
+            rank0_print(f"Loading EVA ViT: {self.vision_tower_name}")
+            self.load_model()
+        elif getattr(args, "unfreeze_mm_vision_tower", False):
+            # TODO: better detector is needed.
+            rank0_print(f"The checkpoint seems to contain `vision_tower` weights: `unfreeze_mm_vision_tower`: True.")
+            self.load_model()
+        elif hasattr(args, "mm_tunable_parts") and "mm_vision_tower" in args.mm_tunable_parts:
+            rank0_print(f"The checkpoint seems to contain `vision_tower` weights: `mm_tunable_parts` contains `mm_vision_tower`.")
+            self.load_model()
+        else:
+            self.cfg_only = self.config
+
+    def load_model(self, device_map=None):
+        rank0_print(f"Pretrained: {self.vision_tower_pretrained}")
+        self.image_processor = EvaClipImageTrainProcessor(self.config["vision_cfg"]["image_size"])
+        self.vision_tower = EVAEncoderWrapper(self.vision_tower_pretrained, self.config)
+        rank0_print(f"Loaded image processor: {self.image_processor}")
+        self.vision_tower.requires_grad_(False)
+        self.is_loaded = True
+
+    def forward(self, images):
+        if type(images) is list:
+            image_features = []
+            for image in images:
+                image_feature = self.vision_tower(image.to(device=self.device, dtype=self.dtype).unsqueeze(0)).to(image.dtype)
+                image_features.append(image_feature)
+        else:
+            image_features = self.vision_tower(images.to(device=self.device, dtype=self.dtype)).to(images.dtype)
+
+        return image_features
+
+    @property
+    def dtype(self):
+        return self.vision_tower.dtype
+
+    @property
+    def device(self):
+        return self.vision_tower.device
+
+    @property
+    def hidden_size(self):
+        return self.config["vision_cfg"]["width"]
+
+    @property
+    def num_patches(self):
+        return (self.config["vision_cfg"]["image_size"] // self.config["vision_cfg"]["patch_size"]) ** 2
+
+    @property
+    def num_patches_per_side(self):
+        return self.config["vision_cfg"]["image_size"] // self.config["vision_cfg"]["patch_size"]
+
+    @property
+    def image_size(self):
+        return self.config["vision_cfg"]["image_size"]

VLMEvalKit-sudoku/llava/model/multimodal_encoder/eva_clip/eva_vit.py

@@ -0,0 +1,856 @@
+"""
+# Adapted from https://github.com/baaivision/EVA/tree/master/EVA-CLIP
+"""
+
+from math import pi
+import torch
+from torch import nn
+from einops import rearrange, repeat
+import logging
+from llava.utils import rank0_print
+
+
+def broadcat(tensors, dim=-1):
+    num_tensors = len(tensors)
+    shape_lens = set(list(map(lambda t: len(t.shape), tensors)))
+    assert len(shape_lens) == 1, "tensors must all have the same number of dimensions"
+    shape_len = list(shape_lens)[0]
+    dim = (dim + shape_len) if dim < 0 else dim
+    dims = list(zip(*map(lambda t: list(t.shape), tensors)))
+    expandable_dims = [(i, val) for i, val in enumerate(dims) if i != dim]
+    assert all([*map(lambda t: len(set(t[1])) <= 2, expandable_dims)]), "invalid dimensions for broadcastable concatentation"
+    max_dims = list(map(lambda t: (t[0], max(t[1])), expandable_dims))
+    expanded_dims = list(map(lambda t: (t[0], (t[1],) * num_tensors), max_dims))
+    expanded_dims.insert(dim, (dim, dims[dim]))
+    expandable_shapes = list(zip(*map(lambda t: t[1], expanded_dims)))
+    tensors = list(map(lambda t: t[0].expand(*t[1]), zip(tensors, expandable_shapes)))
+    return torch.cat(tensors, dim=dim)
+
+
+def rotate_half(x):
+    x = rearrange(x, "... (d r) -> ... d r", r=2)
+    x1, x2 = x.unbind(dim=-1)
+    x = torch.stack((-x2, x1), dim=-1)
+    return rearrange(x, "... d r -> ... (d r)")
+
+
+class VisionRotaryEmbeddingFast(nn.Module):
+    def __init__(self, dim, pt_seq_len, ft_seq_len=None, custom_freqs=None, freqs_for="lang", theta=10000, max_freq=10, num_freqs=1, patch_dropout=0.0):
+        super().__init__()
+        if custom_freqs:
+            freqs = custom_freqs
+        elif freqs_for == "lang":
+            freqs = 1.0 / (theta ** (torch.arange(0, dim, 2)[: (dim // 2)].float() / dim))
+        elif freqs_for == "pixel":
+            freqs = torch.linspace(1.0, max_freq / 2, dim // 2) * pi
+        elif freqs_for == "constant":
+            freqs = torch.ones(num_freqs).float()
+        else:
+            raise ValueError(f"unknown modality {freqs_for}")
+
+        if ft_seq_len is None:
+            ft_seq_len = pt_seq_len
+        t = torch.arange(ft_seq_len) / ft_seq_len * pt_seq_len
+
+        freqs = torch.einsum("..., f -> ... f", t, freqs)
+        freqs = repeat(freqs, "... n -> ... (n r)", r=2)
+        freqs = broadcat((freqs[:, None, :], freqs[None, :, :]), dim=-1)
+
+        freqs_cos = freqs.cos().view(-1, freqs.shape[-1])
+        freqs_sin = freqs.sin().view(-1, freqs.shape[-1])
+
+        self.patch_dropout = patch_dropout
+
+        self.register_buffer("freqs_cos", freqs_cos)
+        self.register_buffer("freqs_sin", freqs_sin)
+
+        logging.info(f"Shape of rope freq: {self.freqs_cos.shape}")
+
+    def forward(self, t, patch_indices_keep=None):
+        if patch_indices_keep is not None:
+            batch = t.size()[0]
+            batch_indices = torch.arange(batch)
+            batch_indices = batch_indices[..., None]
+
+            freqs_cos = repeat(self.freqs_cos, "i j -> n i m j", n=t.shape[0], m=t.shape[1])
+            freqs_sin = repeat(self.freqs_sin, "i j -> n i m j", n=t.shape[0], m=t.shape[1])
+
+            freqs_cos = freqs_cos[batch_indices, patch_indices_keep]
+            freqs_cos = rearrange(freqs_cos, "n i m j -> n m i j")
+            freqs_sin = freqs_sin[batch_indices, patch_indices_keep]
+            freqs_sin = rearrange(freqs_sin, "n i m j -> n m i j")
+
+            return t * freqs_cos + rotate_half(t) * freqs_sin
+
+        return t * self.freqs_cos + rotate_half(t) * self.freqs_sin
+
+
+class LayerNorm(nn.LayerNorm):
+    """Subclass torch's LayerNorm (with cast back to input dtype)."""
+
+    def forward(self, x: torch.Tensor):
+        orig_type = x.dtype
+        x = F.layer_norm(x, self.normalized_shape, self.weight, self.bias, self.eps)
+        return x.to(orig_type)
+
+
+class PatchDropout(nn.Module):
+    """
+    https://arxiv.org/abs/2212.00794
+    """
+
+    def __init__(self, prob, exclude_first_token=True):
+        super().__init__()
+        assert 0 <= prob < 1.0
+        self.prob = prob
+        self.exclude_first_token = exclude_first_token  # exclude CLS token
+        logging.info(f"os.getenv('RoPE')={os.getenv('RoPE')}")
+
+    def forward(self, x):
+        if not self.training or self.prob == 0.0:
+            return x
+
+        if self.exclude_first_token:
+            cls_tokens, x = x[:, :1], x[:, 1:]
+        else:
+            cls_tokens = torch.jit.annotate(torch.Tensor, x[:, :1])
+
+        batch = x.size()[0]
+        num_tokens = x.size()[1]
+
+        batch_indices = torch.arange(batch)
+        batch_indices = batch_indices[..., None]
+
+        keep_prob = 1 - self.prob
+        num_patches_keep = max(1, int(num_tokens * keep_prob))
+
+        rand = torch.randn(batch, num_tokens)
+        patch_indices_keep = rand.topk(num_patches_keep, dim=-1).indices
+
+        x = x[batch_indices, patch_indices_keep]
+
+        if self.exclude_first_token:
+            x = torch.cat((cls_tokens, x), dim=1)
+
+        if self.training and os.getenv("RoPE") == "1":
+            return x, patch_indices_keep
+
+        return x
+
+
+# --------------------------------------------------------
+# Adapted from  https://github.com/microsoft/unilm/tree/master/beit
+# --------------------------------------------------------
+import math
+import os
+import torch.nn as nn
+import torch.nn.functional as F
+
+try:
+    from timm.models.layers import drop_path, to_2tuple, trunc_normal_
+except:
+    from timm.layers import drop_path, to_2tuple, trunc_normal_
+
+if os.getenv("ENV_TYPE") == "deepspeed":
+    try:
+        from deepspeed.runtime.activation_checkpointing.checkpointing import checkpoint
+    except:
+        from torch.utils.checkpoint import checkpoint
+else:
+    from torch.utils.checkpoint import checkpoint
+
+try:
+    import xformers.ops as xops
+except ImportError:
+    xops = None
+    # print("Please 'pip install xformers'")
+
+
+class DropPath(nn.Module):
+    """Drop paths (Stochastic Depth) per sample  (when applied in main path of residual blocks)."""
+
+    def __init__(self, drop_prob=None):
+        super(DropPath, self).__init__()
+        self.drop_prob = drop_prob
+
+    def forward(self, x):
+        return drop_path(x, self.drop_prob, self.training)
+
+    def extra_repr(self) -> str:
+        return "p={}".format(self.drop_prob)
+
+
+class Mlp(nn.Module):
+    def __init__(
+        self,
+        in_features,
+        hidden_features=None,
+        out_features=None,
+        act_layer=nn.GELU,
+        norm_layer=nn.LayerNorm,
+        drop=0.0,
+        subln=False,
+    ):
+        super().__init__()
+        out_features = out_features or in_features
+        hidden_features = hidden_features or in_features
+        self.fc1 = nn.Linear(in_features, hidden_features)
+        self.act = act_layer()
+
+        self.ffn_ln = norm_layer(hidden_features) if subln else nn.Identity()
+
+        self.fc2 = nn.Linear(hidden_features, out_features)
+        self.drop = nn.Dropout(drop)
+
+    def forward(self, x):
+        x = self.fc1(x)
+        x = self.act(x)
+        # x = self.drop(x)
+        # commit this for the orignal BERT implement
+        x = self.ffn_ln(x)
+
+        x = self.fc2(x)
+        x = self.drop(x)
+        return x
+
+
+class SwiGLU(nn.Module):
+    def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.SiLU, drop=0.0, norm_layer=nn.LayerNorm, subln=False):
+        super().__init__()
+        out_features = out_features or in_features
+        hidden_features = hidden_features or in_features
+
+        self.w1 = nn.Linear(in_features, hidden_features)
+        self.w2 = nn.Linear(in_features, hidden_features)
+
+        self.act = act_layer()
+        self.ffn_ln = norm_layer(hidden_features) if subln else nn.Identity()
+        self.w3 = nn.Linear(hidden_features, out_features)
+
+        self.drop = nn.Dropout(drop)
+
+    def forward(self, x):
+        x1 = self.w1(x)
+        x2 = self.w2(x)
+        hidden = self.act(x1) * x2
+        x = self.ffn_ln(hidden)
+        x = self.w3(x)
+        x = self.drop(x)
+        return x
+
+
+class Attention(nn.Module):
+    def __init__(self, dim, num_heads=8, qkv_bias=False, qk_scale=None, attn_drop=0.0, proj_drop=0.0, window_size=None, attn_head_dim=None, xattn=False, rope=None, subln=False, norm_layer=nn.LayerNorm):
+        super().__init__()
+        self.num_heads = num_heads
+        head_dim = dim // num_heads
+        if attn_head_dim is not None:
+            head_dim = attn_head_dim
+        all_head_dim = head_dim * self.num_heads
+        self.scale = qk_scale or head_dim**-0.5
+
+        self.subln = subln
+        if self.subln:
+            self.q_proj = nn.Linear(dim, all_head_dim, bias=False)
+            self.k_proj = nn.Linear(dim, all_head_dim, bias=False)
+            self.v_proj = nn.Linear(dim, all_head_dim, bias=False)
+        else:
+            self.qkv = nn.Linear(dim, all_head_dim * 3, bias=False)
+
+        if qkv_bias:
+            self.q_bias = nn.Parameter(torch.zeros(all_head_dim))
+            self.v_bias = nn.Parameter(torch.zeros(all_head_dim))
+        else:
+            self.q_bias = None
+            self.v_bias = None
+
+        if window_size:
+            self.window_size = window_size
+            self.num_relative_distance = (2 * window_size[0] - 1) * (2 * window_size[1] - 1) + 3
+            self.relative_position_bias_table = nn.Parameter(torch.zeros(self.num_relative_distance, num_heads))  # 2*Wh-1 * 2*Ww-1, nH
+            # cls to token & token 2 cls & cls to cls
+
+            # get pair-wise relative position index for each token inside the window
+            coords_h = torch.arange(window_size[0])
+            coords_w = torch.arange(window_size[1])
+            coords = torch.stack(torch.meshgrid([coords_h, coords_w]))  # 2, Wh, Ww
+            coords_flatten = torch.flatten(coords, 1)  # 2, Wh*Ww
+            relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]  # 2, Wh*Ww, Wh*Ww
+            relative_coords = relative_coords.permute(1, 2, 0).contiguous()  # Wh*Ww, Wh*Ww, 2
+            relative_coords[:, :, 0] += window_size[0] - 1  # shift to start from 0
+            relative_coords[:, :, 1] += window_size[1] - 1
+            relative_coords[:, :, 0] *= 2 * window_size[1] - 1
+            relative_position_index = torch.zeros(size=(window_size[0] * window_size[1] + 1,) * 2, dtype=relative_coords.dtype)
+            relative_position_index[1:, 1:] = relative_coords.sum(-1)  # Wh*Ww, Wh*Ww
+            relative_position_index[0, 0:] = self.num_relative_distance - 3
+            relative_position_index[0:, 0] = self.num_relative_distance - 2
+            relative_position_index[0, 0] = self.num_relative_distance - 1
+
+            self.register_buffer("relative_position_index", relative_position_index)
+        else:
+            self.window_size = None
+            self.relative_position_bias_table = None
+            self.relative_position_index = None
+
+        self.attn_drop = nn.Dropout(attn_drop)
+        self.inner_attn_ln = norm_layer(all_head_dim) if subln else nn.Identity()
+        # self.proj = nn.Linear(all_head_dim, all_head_dim)
+        self.proj = nn.Linear(all_head_dim, dim)
+        self.proj_drop = nn.Dropout(proj_drop)
+        self.xattn = xattn
+        self.xattn_drop = attn_drop
+
+        self.rope = rope
+
+    def forward(self, x, rel_pos_bias=None, attn_mask=None):
+        B, N, C = x.shape
+        if self.subln:
+            q = F.linear(input=x, weight=self.q_proj.weight, bias=self.q_bias)
+            k = F.linear(input=x, weight=self.k_proj.weight, bias=None)
+            v = F.linear(input=x, weight=self.v_proj.weight, bias=self.v_bias)
+
+            q = q.reshape(B, N, self.num_heads, -1).permute(0, 2, 1, 3)  # B, num_heads, N, C
+            k = k.reshape(B, N, self.num_heads, -1).permute(0, 2, 1, 3)
+            v = v.reshape(B, N, self.num_heads, -1).permute(0, 2, 1, 3)
+        else:
+
+            qkv_bias = None
+            if self.q_bias is not None:
+                qkv_bias = torch.cat((self.q_bias, torch.zeros_like(self.v_bias, requires_grad=False), self.v_bias))
+
+            qkv = F.linear(input=x, weight=self.qkv.weight, bias=qkv_bias)
+            qkv = qkv.reshape(B, N, 3, self.num_heads, -1).permute(2, 0, 3, 1, 4)  # 3, B, num_heads, N, C
+            q, k, v = qkv[0], qkv[1], qkv[2]
+
+        if self.rope:
+            # slightly fast impl
+            q_t = q[:, :, 1:, :]
+            ro_q_t = self.rope(q_t)
+            q = torch.cat((q[:, :, :1, :], ro_q_t), -2).type_as(v)
+
+            k_t = k[:, :, 1:, :]
+            ro_k_t = self.rope(k_t)
+            k = torch.cat((k[:, :, :1, :], ro_k_t), -2).type_as(v)
+
+        if self.xattn and xops is not None:
+            q = q.permute(0, 2, 1, 3)  # B, num_heads, N, C -> B, N, num_heads, C
+            k = k.permute(0, 2, 1, 3)
+            v = v.permute(0, 2, 1, 3)
+
+            x = xops.memory_efficient_attention(
+                q,
+                k,
+                v,
+                p=self.xattn_drop,
+                scale=self.scale,
+            )
+            x = x.reshape(B, N, -1)
+            x = self.inner_attn_ln(x)
+            x = self.proj(x)
+            x = self.proj_drop(x)
+        else:
+            q = q * self.scale
+            attn = q @ k.transpose(-2, -1)
+
+            if self.relative_position_bias_table is not None:
+                relative_position_bias = self.relative_position_bias_table[self.relative_position_index.view(-1)].view(self.window_size[0] * self.window_size[1] + 1, self.window_size[0] * self.window_size[1] + 1, -1)  # Wh*Ww,Wh*Ww,nH
+                relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()  # nH, Wh*Ww, Wh*Ww
+                attn = attn + relative_position_bias.unsqueeze(0).type_as(attn)
+
+            if rel_pos_bias is not None:
+                attn = attn + rel_pos_bias.type_as(attn)
+
+            if attn_mask is not None:
+                attn_mask = attn_mask.bool()
+                attn = attn.masked_fill(~attn_mask[:, None, None, :], float("-inf"))
+
+            attn = attn.softmax(dim=-1)
+            attn = self.attn_drop(attn)
+
+            x = (attn @ v).transpose(1, 2).reshape(B, N, -1)
+            x = self.inner_attn_ln(x)
+            x = self.proj(x)
+            x = self.proj_drop(x)
+        return x
+
+
+class Block(nn.Module):
+
+    def __init__(
+        self,
+        dim,
+        num_heads,
+        mlp_ratio=4.0,
+        qkv_bias=False,
+        qk_scale=None,
+        drop=0.0,
+        attn_drop=0.0,
+        drop_path=0.0,
+        init_values=None,
+        act_layer=nn.GELU,
+        norm_layer=nn.LayerNorm,
+        window_size=None,
+        attn_head_dim=None,
+        xattn=False,
+        rope=None,
+        postnorm=False,
+        subln=False,
+        naiveswiglu=False,
+    ):
+        super().__init__()
+        self.norm1 = norm_layer(dim)
+        self.attn = Attention(
+            dim, num_heads=num_heads, qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop, window_size=window_size, attn_head_dim=attn_head_dim, xattn=xattn, rope=rope, subln=subln, norm_layer=norm_layer
+        )
+        # NOTE: drop path for stochastic depth, we shall see if this is better than dropout here
+        self.drop_path = DropPath(drop_path) if drop_path > 0.0 else nn.Identity()
+        self.norm2 = norm_layer(dim)
+        mlp_hidden_dim = int(dim * mlp_ratio)
+
+        if naiveswiglu:
+            self.mlp = SwiGLU(
+                in_features=dim,
+                hidden_features=mlp_hidden_dim,
+                subln=subln,
+                norm_layer=norm_layer,
+            )
+        else:
+            self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, subln=subln, drop=drop)
+
+        if init_values is not None and init_values > 0:
+            self.gamma_1 = nn.Parameter(init_values * torch.ones((dim)), requires_grad=True)
+            self.gamma_2 = nn.Parameter(init_values * torch.ones((dim)), requires_grad=True)
+        else:
+            self.gamma_1, self.gamma_2 = None, None
+
+        self.postnorm = postnorm
+
+    def forward(self, x, rel_pos_bias=None, attn_mask=None):
+        if self.gamma_1 is None:
+            if self.postnorm:
+                x = x + self.drop_path(self.norm1(self.attn(x, rel_pos_bias=rel_pos_bias, attn_mask=attn_mask)))
+                x = x + self.drop_path(self.norm2(self.mlp(x)))
+            else:
+                x = x + self.drop_path(self.attn(self.norm1(x), rel_pos_bias=rel_pos_bias, attn_mask=attn_mask))
+                x = x + self.drop_path(self.mlp(self.norm2(x)))
+        else:
+            if self.postnorm:
+                x = x + self.drop_path(self.gamma_1 * self.norm1(self.attn(x, rel_pos_bias=rel_pos_bias, attn_mask=attn_mask)))
+                x = x + self.drop_path(self.gamma_2 * self.norm2(self.mlp(x)))
+            else:
+                x = x + self.drop_path(self.gamma_1 * self.attn(self.norm1(x), rel_pos_bias=rel_pos_bias, attn_mask=attn_mask))
+                x = x + self.drop_path(self.gamma_2 * self.mlp(self.norm2(x)))
+        return x
+
+
+class PatchEmbed(nn.Module):
+    """Image to Patch Embedding"""
+
+    def __init__(self, img_size=224, patch_size=16, in_chans=3, embed_dim=768):
+        super().__init__()
+        img_size = to_2tuple(img_size)
+        patch_size = to_2tuple(patch_size)
+        num_patches = (img_size[1] // patch_size[1]) * (img_size[0] // patch_size[0])
+        self.patch_shape = (img_size[0] // patch_size[0], img_size[1] // patch_size[1])
+        self.img_size = img_size
+        self.patch_size = patch_size
+        self.num_patches = num_patches
+
+        self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)
+
+    def forward(self, x, **kwargs):
+        B, C, H, W = x.shape
+        # FIXME look at relaxing size constraints
+        assert H == self.img_size[0] and W == self.img_size[1], f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."
+        x = self.proj(x).flatten(2).transpose(1, 2)
+        return x
+
+
+class RelativePositionBias(nn.Module):
+
+    def __init__(self, window_size, num_heads):
+        super().__init__()
+        self.window_size = window_size
+        self.num_relative_distance = (2 * window_size[0] - 1) * (2 * window_size[1] - 1) + 3
+        self.relative_position_bias_table = nn.Parameter(torch.zeros(self.num_relative_distance, num_heads))  # 2*Wh-1 * 2*Ww-1, nH
+        # cls to token & token 2 cls & cls to cls
+
+        # get pair-wise relative position index for each token inside the window
+        coords_h = torch.arange(window_size[0])
+        coords_w = torch.arange(window_size[1])
+        coords = torch.stack(torch.meshgrid([coords_h, coords_w]))  # 2, Wh, Ww
+        coords_flatten = torch.flatten(coords, 1)  # 2, Wh*Ww
+        relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]  # 2, Wh*Ww, Wh*Ww
+        relative_coords = relative_coords.permute(1, 2, 0).contiguous()  # Wh*Ww, Wh*Ww, 2
+        relative_coords[:, :, 0] += window_size[0] - 1  # shift to start from 0
+        relative_coords[:, :, 1] += window_size[1] - 1
+        relative_coords[:, :, 0] *= 2 * window_size[1] - 1
+        relative_position_index = torch.zeros(size=(window_size[0] * window_size[1] + 1,) * 2, dtype=relative_coords.dtype)
+        relative_position_index[1:, 1:] = relative_coords.sum(-1)  # Wh*Ww, Wh*Ww
+        relative_position_index[0, 0:] = self.num_relative_distance - 3
+        relative_position_index[0:, 0] = self.num_relative_distance - 2
+        relative_position_index[0, 0] = self.num_relative_distance - 1
+
+        self.register_buffer("relative_position_index", relative_position_index)
+
+    def forward(self):
+        relative_position_bias = self.relative_position_bias_table[self.relative_position_index.view(-1)].view(self.window_size[0] * self.window_size[1] + 1, self.window_size[0] * self.window_size[1] + 1, -1)  # Wh*Ww,Wh*Ww,nH
+        return relative_position_bias.permute(2, 0, 1).contiguous()  # nH, Wh*Ww, Wh*Ww
+
+
+class EVAVisionTransformer(nn.Module):
+    """Vision Transformer with support for patch or hybrid CNN input stage"""
+
+    def __init__(
+        self,
+        img_size=224,
+        patch_size=16,
+        in_chans=3,
+        num_classes=1000,
+        embed_dim=768,
+        depth=12,
+        num_heads=12,
+        mlp_ratio=4.0,
+        qkv_bias=False,
+        qk_scale=None,
+        drop_rate=0.0,
+        attn_drop_rate=0.0,
+        drop_path_rate=0.0,
+        norm_layer=nn.LayerNorm,
+        init_values=None,
+        patch_dropout=0.0,
+        use_abs_pos_emb=True,
+        use_rel_pos_bias=False,
+        use_shared_rel_pos_bias=False,
+        rope=False,
+        use_mean_pooling=True,
+        init_scale=0.001,
+        grad_checkpointing=False,
+        xattn=False,
+        postnorm=False,
+        pt_hw_seq_len=16,
+        intp_freq=False,
+        naiveswiglu=False,
+        subln=False,
+    ):
+        super().__init__()
+        self.image_size = img_size
+        self.num_classes = num_classes
+        self.num_features = self.embed_dim = embed_dim  # num_features for consistency with other models
+
+        self.patch_embed = PatchEmbed(img_size=img_size, patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim)
+        num_patches = self.patch_embed.num_patches
+
+        self.cls_token = nn.Parameter(torch.zeros(1, 1, embed_dim))
+        # self.mask_token = nn.Parameter(torch.zeros(1, 1, embed_dim))
+        if use_abs_pos_emb:
+            self.pos_embed = nn.Parameter(torch.zeros(1, num_patches + 1, embed_dim))
+        else:
+            self.pos_embed = None
+        self.pos_drop = nn.Dropout(p=drop_rate)
+
+        if use_shared_rel_pos_bias:
+            self.rel_pos_bias = RelativePositionBias(window_size=self.patch_embed.patch_shape, num_heads=num_heads)
+        else:
+            self.rel_pos_bias = None
+
+        if rope:
+            half_head_dim = embed_dim // num_heads // 2
+            hw_seq_len = img_size // patch_size
+            self.rope = VisionRotaryEmbeddingFast(
+                dim=half_head_dim,
+                pt_seq_len=pt_hw_seq_len,
+                ft_seq_len=hw_seq_len if intp_freq else None,
+                # patch_dropout=patch_dropout
+            )
+        else:
+            self.rope = None
+
+        self.naiveswiglu = naiveswiglu
+
+        dpr = [x.item() for x in torch.linspace(0, drop_path_rate, depth)]  # stochastic depth decay rule
+        self.use_rel_pos_bias = use_rel_pos_bias
+        self.blocks = nn.ModuleList(
+            [
+                Block(
+                    dim=embed_dim,
+                    num_heads=num_heads,
+                    mlp_ratio=mlp_ratio,
+                    qkv_bias=qkv_bias,
+                    qk_scale=qk_scale,
+                    drop=drop_rate,
+                    attn_drop=attn_drop_rate,
+                    drop_path=dpr[i],
+                    norm_layer=norm_layer,
+                    init_values=init_values,
+                    window_size=self.patch_embed.patch_shape if use_rel_pos_bias else None,
+                    xattn=xattn,
+                    rope=self.rope,
+                    postnorm=postnorm,
+                    subln=subln,
+                    naiveswiglu=naiveswiglu,
+                )
+                for i in range(depth)
+            ]
+        )
+        self.norm = nn.Identity() if use_mean_pooling else norm_layer(embed_dim)
+        self.fc_norm = norm_layer(embed_dim) if use_mean_pooling else None
+        self.head = nn.Linear(embed_dim, num_classes) if num_classes > 0 else nn.Identity()
+
+        if self.pos_embed is not None:
+            trunc_normal_(self.pos_embed, std=0.02)
+
+        trunc_normal_(self.cls_token, std=0.02)
+        # trunc_normal_(self.mask_token, std=.02)
+
+        self.apply(self._init_weights)
+        self.fix_init_weight()
+
+        if isinstance(self.head, nn.Linear):
+            trunc_normal_(self.head.weight, std=0.02)
+            self.head.weight.data.mul_(init_scale)
+            self.head.bias.data.mul_(init_scale)
+
+        # setting a patch_dropout of 0. would mean it is disabled and this function would be the identity fn
+        self.patch_dropout = PatchDropout(patch_dropout) if patch_dropout > 0.0 else nn.Identity()
+
+        self.grad_checkpointing = grad_checkpointing
+
+    def fix_init_weight(self):
+        def rescale(param, layer_id):
+            param.div_(math.sqrt(2.0 * layer_id))
+
+        for layer_id, layer in enumerate(self.blocks):
+            rescale(layer.attn.proj.weight.data, layer_id + 1)
+            if self.naiveswiglu:
+                rescale(layer.mlp.w3.weight.data, layer_id + 1)
+            else:
+                rescale(layer.mlp.fc2.weight.data, layer_id + 1)
+
+    def get_cast_dtype(self) -> torch.dtype:
+        return self.blocks[0].mlp.fc2.weight.dtype
+
+    def _init_weights(self, m):
+        if isinstance(m, nn.Linear):
+            trunc_normal_(m.weight, std=0.02)
+            if m.bias is not None:
+                nn.init.constant_(m.bias, 0)
+        elif isinstance(m, nn.LayerNorm):
+            nn.init.constant_(m.bias, 0)
+            nn.init.constant_(m.weight, 1.0)
+
+    def get_num_layers(self):
+        return len(self.blocks)
+
+    def lock(self, unlocked_groups=0, freeze_bn_stats=False):
+        assert unlocked_groups == 0, "partial locking not currently supported for this model"
+        for param in self.parameters():
+            param.requires_grad = False
+
+    @torch.jit.ignore
+    def set_grad_checkpointing(self, enable=True):
+        self.grad_checkpointing = enable
+
+    @torch.jit.ignore
+    def no_weight_decay(self):
+        return {"pos_embed", "cls_token"}
+
+    def get_classifier(self):
+        return self.head
+
+    def reset_classifier(self, num_classes, global_pool=""):
+        self.num_classes = num_classes
+        self.head = nn.Linear(self.embed_dim, num_classes) if num_classes > 0 else nn.Identity()
+
+    def forward_features(self, x, return_all_features=False):
+
+        x = self.patch_embed(x)
+        batch_size, seq_len, _ = x.size()
+
+        cls_tokens = self.cls_token.expand(batch_size, -1, -1)  # stole cls_tokens impl from Phil Wang, thanks
+        x = torch.cat((cls_tokens, x), dim=1)
+        if self.pos_embed is not None:
+            x = x + self.pos_embed
+        x = self.pos_drop(x)
+
+        # a patch_dropout of 0. would mean it is disabled and this function would do nothing but return what was passed in
+        if os.getenv("RoPE") == "1":
+            if self.training and not isinstance(self.patch_dropout, nn.Identity):
+                x, patch_indices_keep = self.patch_dropout(x)
+                # Directly pass patch_indices_keep to self.rope.forward
+                x = self.rope.forward(x, patch_indices_keep=patch_indices_keep)
+            else:
+                # Pass None or omit the patch_indices_keep argument for default behavior
+                x = self.rope.forward(x, patch_indices_keep=None)
+                x = self.patch_dropout(x)
+        else:
+            x = self.patch_dropout(x)
+
+        rel_pos_bias = self.rel_pos_bias() if self.rel_pos_bias is not None else None
+        for i, blk in enumerate(self.blocks):
+            if i == len(self.blocks) - 1:
+                continue
+            if self.grad_checkpointing:
+                x = checkpoint(blk, x, (rel_pos_bias,))
+            else:
+                x = blk(x, rel_pos_bias=rel_pos_bias)
+
+        if not return_all_features:
+            x = self.norm(x)
+            if self.fc_norm is not None:
+                return self.fc_norm(x.mean(1))
+            else:
+                return x[:, 0]
+        return x
+
+    def forward(self, x, return_all_features=False):
+        if return_all_features:
+            return self.forward_features(x, return_all_features)
+        x = self.forward_features(x)
+        x = self.head(x)
+        return x
+
+
+def load_state_dict(checkpoint_path: str, map_location: str = "cpu", model_key: str = "model|module|state_dict", is_openai: bool = False, skip_list: list = []):
+    if is_openai:
+        model = torch.jit.load(checkpoint_path, map_location="cpu").eval()
+        state_dict = model.state_dict()
+        for key in ["input_resolution", "context_length", "vocab_size"]:
+            state_dict.pop(key, None)
+    else:
+        checkpoint = torch.load(checkpoint_path, map_location=map_location)
+        for mk in model_key.split("|"):
+            if isinstance(checkpoint, dict) and mk in checkpoint:
+                state_dict = checkpoint[mk]
+                break
+            else:
+                state_dict = checkpoint
+        if next(iter(state_dict.items()))[0].startswith("module"):
+            state_dict = {k[7:]: v for k, v in state_dict.items()}
+
+    for k in skip_list:
+        if k in list(state_dict.keys()):
+            logging.info(f"Removing key {k} from pretrained checkpoint")
+            del state_dict[k]
+
+    if os.getenv("RoPE") == "1":
+        for k in list(state_dict.keys()):
+            if "freqs_cos" in k or "freqs_sin" in k:
+                del state_dict[k]
+    return state_dict
+
+
+def load_clip_visual_state_dict(checkpoint_path: str, map_location: str = "cpu", is_openai: bool = False, skip_list: list = []):
+    state_dict = load_state_dict(checkpoint_path, map_location=map_location, is_openai=is_openai, skip_list=skip_list)
+    # for k in list(state_dict.keys()):
+    #     if not k.startswith("visual."):
+    #         del state_dict[k]
+    # for k in list(state_dict.keys()):
+    #     if k.startswith("visual."):
+    #         new_k = k[7:]
+    #         state_dict[new_k] = state_dict[k]
+    #         del state_dict[k]
+    return state_dict
+
+
+from dataclasses import dataclass
+from typing import Optional, Tuple, Union
+
+try:
+    from apex.normalization import FusedLayerNorm
+except:
+    FusedLayerNorm = LayerNorm
+    # print("Please build and install Nvidia apex package with option '--cuda_ext' according to https://github.com/NVIDIA/apex#from-source .")
+
+
+@dataclass
+class CLIPVisionCfg:
+    layers: Union[Tuple[int, int, int, int], int] = 12
+    width: int = 768
+    head_width: int = 64
+    mlp_ratio: float = 4.0
+    patch_size: int = 16
+    image_size: Union[Tuple[int, int], int] = 224
+    ls_init_value: Optional[float] = None  # layer scale initial value
+    patch_dropout: float = 0.0  # what fraction of patches to dropout during training (0 would mean disabled and no patches dropped) - 0.5 to 0.75 recommended in the paper for optimal results
+    global_average_pool: bool = False  # whether to global average pool the last embedding layer, instead of using CLS token (https://arxiv.org/abs/2205.01580)
+    drop_path_rate: Optional[float] = None  # drop path rate
+    timm_model_name: str = None  # a valid model name overrides layers, width, patch_size
+    timm_model_pretrained: bool = False  # use (imagenet) pretrained weights for named model
+    timm_pool: str = "avg"  # feature pooling for timm model ('abs_attn', 'rot_attn', 'avg', '')
+    timm_proj: str = "linear"  # linear projection for timm model output ('linear', 'mlp', '')
+    timm_proj_bias: bool = False  # enable bias final projection
+    eva_model_name: str = None  # a valid eva model name overrides layers, width, patch_size
+    qkv_bias: bool = True
+    fusedLN: bool = False
+    xattn: bool = False
+    postnorm: bool = False
+    rope: bool = False
+    pt_hw_seq_len: int = 16  # 224/14
+    intp_freq: bool = False
+    naiveswiglu: bool = False
+    subln: bool = False
+
+
+def create_norm_layer_factory(use_fused_ln, eps=1e-6):
+    # Otherwise, use the standard LayerNorm
+    return lambda num_features: nn.LayerNorm(num_features, eps=eps)
+
+
+def _build_vision_tower(vision_tower_path: str, embed_dim: int, vision_cfg: CLIPVisionCfg, **kwargs):
+    if isinstance(vision_cfg, dict):
+        vision_cfg = CLIPVisionCfg(**vision_cfg)
+
+    if vision_cfg.eva_model_name:
+        vision_heads = vision_cfg.width // vision_cfg.head_width
+        # Determine the appropriate norm layer factory based on the configuration
+        norm_layer_factory = create_norm_layer_factory(vision_cfg.fusedLN, eps=1e-6)
+
+        visual = EVAVisionTransformer(
+            img_size=vision_cfg.image_size,
+            patch_size=vision_cfg.patch_size,
+            num_classes=embed_dim,
+            use_mean_pooling=vision_cfg.global_average_pool,  # False
+            init_values=vision_cfg.ls_init_value,
+            patch_dropout=vision_cfg.patch_dropout,
+            embed_dim=vision_cfg.width,
+            depth=vision_cfg.layers,
+            num_heads=vision_heads,
+            mlp_ratio=vision_cfg.mlp_ratio,
+            qkv_bias=vision_cfg.qkv_bias,
+            drop_path_rate=vision_cfg.drop_path_rate,
+            norm_layer=norm_layer_factory,
+            xattn=vision_cfg.xattn,
+            rope=vision_cfg.rope,
+            postnorm=vision_cfg.postnorm,
+            pt_hw_seq_len=vision_cfg.pt_hw_seq_len,  # 224/14
+            intp_freq=vision_cfg.intp_freq,
+            naiveswiglu=vision_cfg.naiveswiglu,
+            subln=vision_cfg.subln,
+        )
+
+        state_dict = load_clip_visual_state_dict(vision_tower_path)
+        incompatible_keys = visual.load_state_dict(state_dict, strict=False)
+        rank0_print("EVA-CLIP incompatible_keys:", incompatible_keys)
+
+    return visual
+
+
+class EVAEncoderWrapper(nn.Module):
+    def __init__(self, vision_tower_pretrained, config):
+        super(EVAEncoderWrapper, self).__init__()
+        self.config = config
+        self.config["vision_tower_path"] = vision_tower_pretrained
+        self.model = _build_vision_tower(**self.config)
+
+    def forward(self, image, **kwargs):
+        encode = self.model(image, return_all_features=True)[:, 1:, :]  # remove the CLS token
+        return encode
+
+    @property
+    def dtype(self):
+        return list(self.parameters())[-1].dtype
+
+    @property
+    def device(self):
+        return list(self.parameters())[-1].device

VLMEvalKit-sudoku/llava/model/multimodal_encoder/eva_clip/model_configs/EVA-CLIP-18B.json

@@ -0,0 +1,27 @@
+{
+    "embed_dim": 1536,
+    "vision_cfg": {
+        "image_size": 224,
+        "layers": 48,
+        "width": 5120,
+        "head_width": 128,
+        "mlp_ratio": 5,
+        "patch_size": 14,
+        "eva_model_name": "eva-clip-18b-14-x",
+        "drop_path_rate": 0,
+        "qkv_bias": false,
+        "xattn": true,
+        "postnorm": true,
+        "fusedLN": false,
+        "use_rms_norm": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 1280,
+        "heads": 20,
+        "layers": 32,
+        "xattn": false,
+        "fusedLN": false
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/eva_clip/model_configs/EVA02-CLIP-B-16.json

@@ -0,0 +1,29 @@
+{
+    "embed_dim": 512,
+    "vision_cfg": {
+        "image_size": 224,
+        "layers": 12,
+        "width": 768,
+        "head_width": 64,
+        "patch_size": 16,
+        "mlp_ratio": 2.6667,
+        "eva_model_name": "eva-clip-b-16-X",
+        "drop_path_rate": 0.0,
+        "xattn": true,
+        "fusedLN": true,
+        "rope": true,
+        "pt_hw_seq_len": 16,
+        "intp_freq": true,
+        "naiveswiglu": true,
+        "subln": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 512,
+        "heads": 8,
+        "layers": 12,
+        "xattn": true,
+        "fusedLN": true
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/eva_clip/model_configs/EVA02-CLIP-bigE-14.json

@@ -0,0 +1,25 @@
+{
+    "embed_dim": 1024,
+    "vision_cfg": {
+        "image_size": 224,
+        "layers": 64,
+        "width": 1792,
+        "head_width": 112,
+        "mlp_ratio": 8.571428571428571,
+        "patch_size": 14,
+        "eva_model_name": "eva-clip-4b-14-x",
+        "drop_path_rate": 0,
+        "xattn": true,
+        "postnorm": true,
+        "fusedLN": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 1024,
+        "heads": 16,
+        "layers": 24,
+        "xattn": false,
+        "fusedLN": true
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/eva_clip/model_configs/Internal-EVA02-CLIP-10B-14.json

@@ -0,0 +1,25 @@
+{
+    "embed_dim": 1024,
+    "vision_cfg": {
+        "image_size": 224,
+        "layers": 77,
+        "width": 2304,
+        "head_width": 144,
+        "mlp_ratio": 10.9722,
+        "patch_size": 14,
+        "eva_model_name": "eva-clip-10b-14-x",
+        "drop_path_rate": 0,
+        "xattn": true,
+        "postnorm": false,
+        "fusedLN": true
+    },
+    "text_cfg": {
+        "context_length": 77,
+        "vocab_size": 49408,
+        "width": 1280,
+        "heads": 20,
+        "layers": 32,
+        "xattn": false,
+        "fusedLN": true
+    }
+}

VLMEvalKit-sudoku/llava/model/multimodal_encoder/hf_vision.py

@@ -0,0 +1,111 @@
+import torch
+import torch.nn as nn
+
+from transformers import AutoModel, AutoImageProcessor, AutoConfig, CLIPImageProcessor
+from llava.utils import rank0_print
+
+
+class HFVisionTower(nn.Module):
+    def __init__(self, vision_tower, args, delay_load=False):
+        super().__init__()
+
+        self.is_loaded = False
+
+        self.vision_tower_name = vision_tower.replace("hf:", "", 1)
+        self.select_layer = args.mm_vision_select_layer
+        self.select_feature = getattr(args, "mm_vision_select_feature", "patch")
+
+        if not delay_load:
+            self.load_model()
+        else:
+            self.cfg_only = AutoConfig.from_pretrained(self.vision_tower_name)
+
+    def load_model(self):
+        try:
+            self.image_processor = AutoImageProcessor.from_pretrained(self.vision_tower_name)
+        except Exception as e:
+            if "448" in self.vision_tower_name:
+                image_size = 448
+                # use image processor with conig
+                self.image_processor = CLIPImageProcessor(size={"shortest_edge": image_size}, do_center_crop=True, crop_size=image_size)
+            else:
+                self.image_processor = CLIPImageProcessor.from_pretrained("openai/clip-vit-large-patch14")
+        rank0_print(f"Loaded image processor: {self.image_processor}")
+        self.vision_tower = AutoModel.from_pretrained(self.vision_tower_name, torch_dtype=torch.bfloat16, trust_remote_code=True).to("cuda")
+        self.device = self.vision_tower.device
+        self.dtype = self.vision_tower.dtype
+        self.config = self.vision_tower.config
+
+        if hasattr(self.vision_tower, "vision_model"):
+            self.vision_tower = self.vision_tower.vision_model
+        self.vision_tower.requires_grad_(False)
+        # self.vision_tower.eval()
+        self.is_loaded = True
+
+    def feature_select(self, image_forward_outs):
+        select_feature_type = self.select_feature
+
+        if self.select_feature in ["slicefour_patch", "slicefour_cls_patch"]:
+            select_every_k_layer = len(image_forward_outs.hidden_states) // 4
+            image_features = torch.cat([image_forward_outs.hidden_states[i] for i in range(select_every_k_layer + self.select_layer, len(image_forward_outs.hidden_states), select_every_k_layer)], dim=-1)
+            select_feature_type = select_feature_type.replace("slicefour_", "")
+        else:
+            image_features = image_forward_outs.hidden_states[self.select_layer]
+
+        if select_feature_type == "patch":
+            image_features = image_features[:, 1:]
+        elif select_feature_type == "cls_patch":
+            image_features = image_features
+        else:
+            raise ValueError(f"Unexpected select feature: {select_feature_type}")
+        return image_features
+
+    def forward(self, images):
+        if type(images) is list:
+            image_features = []
+            for image in images:
+                image_forward_out = self.vision_tower(image.to(device=self.device, dtype=self.dtype).unsqueeze(0), output_hidden_states=True)
+                image_feature = self.feature_select(image_forward_out).to(image.dtype)
+                image_features.append(image_feature)
+        else:
+            image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype), output_hidden_states=True)
+            image_features = self.feature_select(image_forward_outs).to(images.dtype)
+
+        return image_features
+
+    @property
+    def dummy_feature(self):
+        return torch.zeros(1, self.hidden_size, device=self.device, dtype=self.dtype)
+
+    # @property
+    # def dtype(self):
+    #     return self.vision_tower.dtype
+
+    # @property
+    # def device(self):
+    #     return self.vision_tower.device
+
+    @property
+    def hidden_size(self):
+        try:
+            _hidden_size = self.config.hidden_size
+        except:
+            _hidden_size = self.config.vision_config.hidden_size
+        if "slicefour" in self.select_feature:
+            _hidden_size *= 4
+        return _hidden_size
+
+    @property
+    def num_patches(self):
+        _num_patches = (self.config.image_size // self.config.patch_size) ** 2
+        if "cls_patch" in self.select_feature:
+            _num_patches += 1
+        return _num_patches
+
+    @property
+    def num_patches_per_side(self):
+        return self.config.image_size // self.config.patch_size
+
+    @property
+    def image_size(self):
+        return self.config.image_size

VLMEvalKit-sudoku/llava/model/multimodal_encoder/hubconf.py

@@ -0,0 +1,213 @@
+# hubconf.py
+import torch
+from featup.featurizers.util import get_featurizer
+from featup.layers import ChannelNorm
+from featup.upsamplers import get_upsampler
+from torch.nn import Module
+import torch
+from torch.multiprocessing import Pool, set_start_method
+from functools import partial
+import torch.nn.functional as F
+
+dependencies = ['torch', 'torchvision', 'PIL', 'featup']  # List any dependencies here
+
+
+class UpsampledBackbone(Module):
+
+    def __init__(self, model_name, use_norm, scale):
+        super().__init__()
+        model, patch_size, self.dim = get_featurizer(model_name, "token", num_classes=1000)
+        if use_norm:
+            self.model = torch.nn.Sequential(model, ChannelNorm(self.dim))
+        else:
+            self.model = model
+        
+        if scale == '2x':
+            self.upsampler = get_upsampler("jbu_2x_stack", self.dim)
+        elif scale == '4x':
+            self.upsampler = get_upsampler("jbu_4x_stack", self.dim)
+        elif scale == '8x':
+            self.upsampler = get_upsampler("jbu_8x_stack", self.dim)
+        else:
+            self.upsampler = get_upsampler("jbu_stack", self.dim)
+
+    def forward(self, image):
+        return self.upsampler(self.model(image), image)
+    
+    def forward_with_internal_features(self, image, lowres):
+        if type(lowres) == list:
+            #return self.forward_with_features_list(image, lowres)
+            features_2x = []
+            features_4x = []
+            features_8x = []
+            for i in range(len(lowres)):
+                #lowres_norm = self.model[1](lowres[i])
+                res = self.upsampler.forward_with_internal_features(lowres[i], image[i].unsqueeze(0))
+                features_2x.append(res['feat2x'])
+                if 'feat4x' in res:
+                    features_4x.append(res['feat4x'])
+                if 'feat8x' in res:
+                    features_8x.append(res['feat8x'])
+                    
+            return features_2x, features_4x, features_8x
+        else:
+            feat2x = None
+            feat4x = None
+            feat8x = None
+            res = self.upsampler.forward_with_internal_features(lowres, image)
+            if 'feat2x' in res:
+                feat2x = res['feat2x']
+            if 'feat4x' in res:
+                feat4x = res['feat4x']
+            if 'feat8x' in res:
+                feat8x = res['feat8x']
+            return feat2x, feat4x, feat8x
+
+class Upsampled4xBackbone(Module):
+
+    def __init__(self, model_name, use_norm):
+        super().__init__()
+        model, patch_size, self.dim = get_featurizer(model_name, "token", num_classes=1000)
+        if use_norm:
+            self.model = torch.nn.Sequential(model, ChannelNorm(self.dim))
+        else:
+            self.model = model
+        self.upsampler = get_upsampler("jbu_4x_stack", self.dim)
+
+    def forward(self, image):
+        lowres = self.model(image).to(torch.bfloat16)
+        image = image.to(torch.bfloat16)
+        return self.upsampler(lowres, image)
+
+    def pad_to_square_tensor(self, feat, pad_res):#torch.Size([1, 1024, 25, 22]), 33
+        # Calculate padding for each dimension
+        pad_size = (0, pad_res-feat.size(3), 0, pad_res-feat.size(2))  # left, right, top, down
+        # Pad the image tensor
+        padded = F.pad(feat, pad_size, 'constant', 1)
+        return padded
+
+    def unpad_from_square_tensor(self, feat, h, w):
+        # Crop the image tensor from left top to the desired size
+        return feat[:, :h, :w]
+
+    def forward_with_features_list(self, images, lowres):
+        feature_scale = 14 # clip-large 336 -14
+        #list of torch.Size([3, 350, 308]), 
+        #list of torch.Size([1, 1024, 25, 22])
+        pad_res = 0
+        for i in range(len(lowres)):
+            pad_res = max(pad_res, max(lowres[i].size(2), lowres[i].size(3)))
+
+        lowres_tensor = [self.pad_to_square_tensor(lowres[i], pad_res) for i in range(len(lowres))]
+        lowres_tensor = torch.cat(lowres_tensor, dim=0)
+
+        images_tensor = [self.pad_to_square_tensor(images[i].unsqueeze(0), pad_res * feature_scale) for i in range(len(images))]
+        images_tensor = torch.cat(images_tensor, dim=0)
+
+        features_2x, features_4x = self.upsampler.forward_with_internal_features(lowres_tensor, images_tensor)
+
+        feat_2x_list = [self.unpad_from_square_tensor(features_2x[i], lowres[i].size(2) * 2, lowres[i].size(3) * 2).unsqueeze(0) for i in range(len(features_2x))]
+        feat_4x_list = [self.unpad_from_square_tensor(features_4x[i], lowres[i].size(2) * 4, lowres[i].size(3) * 4).unsqueeze(0) for i in range(len(features_4x))]
+
+        return feat_2x_list, feat_4x_list
+
+    def forward_with_internal_features(self, image, lowres):
+        if type(lowres) == list:
+            #return self.forward_with_features_list(image, lowres)
+            features_2x = []
+            features_4x = []
+            features_8x = []
+            for i in range(len(lowres)):
+                res = self.upsampler.forward_with_internal_features(lowres[i], image[i].unsqueeze(0))
+                features_2x.append(res['feat2x'])
+                if res.get('feat4x') is not None:
+                    features_4x.append(res['feat4x'])
+                if res.get('feat8x') is not None:
+                    features_8x.append(res['feat8x'])
+                    
+            return {'feat2x': features_2x, 'feat4x': features_4x, 'feat8x': features_8x}
+        else:
+            return self.upsampler.forward_with_internal_features(lowres, image)
+
+def _load_backbone(pretrained, use_norm, model_name):
+    """
+    The function that will be called by Torch Hub users to instantiate your model.
+    Args:
+        pretrained (bool): If True, returns a model pre-loaded with weights.
+    Returns:
+        An instance of your model.
+    """
+    model = UpsampledBackbone(model_name, use_norm)
+    if pretrained:
+        # Define how you load your pretrained weights here
+        # For example:
+        if use_norm:
+            exp_dir = ""
+        else:
+            exp_dir = "no_norm/"
+
+        checkpoint_url = f"https://marhamilresearch4.blob.core.windows.net/feature-upsampling-public/pretrained/{exp_dir}{model_name}_jbu_stack_cocostuff.ckpt"
+        state_dict = torch.hub.load_state_dict_from_url(checkpoint_url)["state_dict"]
+        state_dict = {k: v for k, v in state_dict.items() if "scale_net" not in k and "downsampler" not in k}
+        model.load_state_dict(state_dict, strict=False)
+    return model
+
+def _load_backbone_from_local(pretrained, use_norm, model_name, ckpt_path, scale = '16x'):
+    """
+    The function that will be called by Torch Hub users to instantiate your model.
+    Args:
+        pretrained (bool): If True, returns a model pre-loaded with weights.
+    Returns:
+        An instance of your model.
+    """
+    model = UpsampledBackbone(model_name, use_norm, scale)
+
+    if pretrained:
+        ckpt = torch.load(ckpt_path)
+        state_dict = ckpt['state_dict']
+        state_dict = {k: v for k, v in state_dict.items() if "scale_net" not in k and "downsampler" not in k}
+        model.load_state_dict(state_dict, strict=False)
+    return model
+
+def _get_state_dict(ckpt_path):
+    ckpt = torch.load(ckpt_path)
+    state_dict = ckpt['state_dict']
+    state_dict = {k: v for k, v in state_dict.items() if "scale_net" not in k and "downsampler" not in k}
+    return state_dict
+
+# def vit(pretrained=True, use_norm=True):
+#     return _load_backbone(pretrained, use_norm, "vit")
+
+def vit(pretrained=True, use_norm=True):
+    ckpt_path = '/home/jeeves/LowResCV/checkpoints/jbu/final_1e6_token_unnorm_pad30_jitter5_lr4_vit_jbu_stack_cocostuff_attention_crf_0.001_tv_0.0_ent_0.0/home/jeeves/LowResCV/checkpoints/jbu/final_1e6_token_unnorm_pad30_jitter5_lr4_vit_jbu_stack_cocostuff_attention_crf_0.001_tv_0.0_ent_0.0_685.ckpt'
+    return _load_backbone_from_local(pretrained, use_norm, "vit", ckpt_path)
+
+def dino16(pretrained=True, use_norm=True):
+    return _load_backbone(pretrained, use_norm, "dino16")
+
+
+def clip(pretrained=True, use_norm=True):
+    return _load_backbone(pretrained, use_norm, "clip")
+
+def clipLarge(pretrained=True, use_norm=True, scale = '4x'):
+    ckpt_path = '/mnt/data/user/tc_agi/zyp/featup/upsampler/0919/checkpoints/clip-large_jbu_4x_stack_cocostuff_attention_crf_0.001_tv_0.0_ent_0.0-0.001-True-30-2-5_2000.ckpt'
+    return _load_backbone_from_local(pretrained, use_norm, "clip-large", ckpt_path, scale)
+
+
+def featup(model_name='clip-large', pretrained=False, use_norm=False, scale = '4x'):
+    ckpt_path = '/mnt/data/user/tc_agi/zyp/featup/upsampler/0919/checkpoints/clip-large_jbu_4x_stack_cocostuff_attention_crf_0.001_tv_0.0_ent_0.0-0.001-True-30-2-5_2000.ckpt'
+    return _load_backbone_from_local(pretrained, use_norm, model_name, ckpt_path, scale)
+
+def get_featup_state_dict(ckpt_path='/mnt/data/user/tc_agi/zyp/featup/upsampler/0919/checkpoints/clip-large_jbu_4x_stack_cocostuff_attention_crf_0.001_tv_0.0_ent_0.0-0.001-True-30-2-5_2000.ckpt'):
+    return _get_state_dict(ckpt_path)
+
+def dinov2(pretrained=True, use_norm=True):
+    return _load_backbone(pretrained, use_norm, "dinov2")
+
+
+def resnet50(pretrained=True, use_norm=True):
+    return _load_backbone(pretrained, use_norm, "resnet50")
+
+def maskclip(pretrained=True, use_norm=True):
+    assert not use_norm, "MaskCLIP only supports unnormed model"
+    return _load_backbone(pretrained, use_norm, "maskclip")

VLMEvalKit-sudoku/llava/model/multimodal_encoder/modeling_siglip2.py

@@ -0,0 +1,1386 @@
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+#           This file was automatically generated from src/transformers/models/siglip2/modular_siglip2.py.
+#               Do NOT edit this file manually as any edits will be overwritten by the generation of
+#             the file from the modular. If any change should be done, please apply the change to the
+#                          modular_siglip2.py file directly. One of our CI enforces this.
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+# coding=utf-8
+# Copyright 2025 The HuggingFace Inc. team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import math
+import warnings
+from dataclasses import dataclass
+from functools import partial, reduce
+import torch.utils.checkpoint
+from PIL import Image
+from typing import Any, Optional, Tuple, Union, Dict
+import os
+
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+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,
+)
+from transformers.configuration_utils import PretrainedConfig
+from transformers.image_processing_utils import BatchFeature, get_size_dict
+from transformers.image_transforms import (
+    convert_to_rgb,
+    normalize,
+    rescale,
+    resize,
+    to_channel_dimension_format,
+)
+from transformers.image_utils import (
+    ChannelDimension,
+    PILImageResampling,
+    to_numpy_array,
+)
+from transformers.activations import ACT2FN
+from transformers.modeling_outputs import BaseModelOutput, BaseModelOutputWithPooling
+from transformers.modeling_utils import PreTrainedModel
+from transformers import PretrainedConfig
+from transformers.utils import ModelOutput
+from llava.utils import rank0_print
+from einops import rearrange
+
+# if is_flash_attn_2_available():
+#     from transformers.modeling_flash_attention_utils import _flash_attention_forward
+
+
+class SigLipImageProcessor:
+    def __init__(self, image_mean=(0.5, 0.5, 0.5), image_std=(0.5, 0.5, 0.5), size=(384, 384), crop_size: Dict[str, int] = None, resample=PILImageResampling.BICUBIC, rescale_factor=1 / 255, data_format=ChannelDimension.FIRST):
+        crop_size = crop_size if crop_size is not None else {"height": 384, "width": 384}
+        crop_size = get_size_dict(crop_size, default_to_square=True, param_name="crop_size")
+
+        self.image_mean = image_mean
+        self.image_std = image_std
+        self.size = size
+        self.resample = resample
+        self.rescale_factor = rescale_factor
+        self.data_format = data_format
+        self.crop_size = crop_size
+
+    def preprocess(self, images, do_resize = True, do_center_crop = True, do_rescale = True, do_normalize = True, return_tensors = 'pt'):
+        if isinstance(images, Image.Image):
+            images = [images]
+        else:
+            # to adapt video data
+            images = [to_numpy_array(image) for image in images]
+            assert isinstance(images, list)
+
+        # do_resize=False, do_center_crop=False, do_rescale=True, do_normalize=True, 
+
+        transforms = [
+            convert_to_rgb,
+            to_numpy_array
+        ]
+
+        if do_resize:
+            transforms.append(partial(resize, size=self.size, resample=self.resample, data_format=self.data_format))
+        if do_rescale:
+            transforms.append(partial(rescale, scale=self.rescale_factor, data_format=self.data_format))
+        if do_normalize:
+            transforms.append(partial(normalize, mean=self.image_mean, std=self.image_std, data_format=self.data_format))
+        
+        transforms.append(partial(to_channel_dimension_format, channel_dim=self.data_format, input_channel_dim=self.data_format))
+
+        images = reduce(lambda x, f: [*map(f, x)], transforms, images)
+        data = {"pixel_values": images}
+
+        return BatchFeature(data=data, tensor_type=return_tensors)
+
+
+class Siglip2TextConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`Siglip2TextModel`]. It is used to instantiate a
+    Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-224](https://huggingface.co/google/siglip2-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 Siglip2 text model. Defines the number of different tokens that can be represented by
+            the `inputs_ids` passed when calling [`Siglip2Model`].
+        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.
+        projection_size (`int`, *optional*, defaults to `hidden_size`):
+            The size of the projection head.
+
+    Example:
+
+    ```python
+    >>> from transformers import Siglip2TextConfig, Siglip2TextModel
+
+    >>> # Initializing a Siglip2TextConfig with google/siglip2-base-patch16-224 style configuration
+    >>> configuration = Siglip2TextConfig()
+
+    >>> # Initializing a Siglip2TextModel (with random weights) from the google/siglip2-base-patch16-224 style configuration
+    >>> model = Siglip2TextModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip2_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/siglip2
+        # See https://github.com/huggingface/transformers/pull/24773#issuecomment-1632287538
+        pad_token_id=1,
+        bos_token_id=49406,
+        eos_token_id=49407,
+        projection_size=None,
+        **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
+        self.projection_size = projection_size if projection_size is not None else hidden_size
+
+
+class Siglip2VisionConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`Siglip2VisionModel`]. It is used to instantiate a
+    Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-naflex](https://huggingface.co/google/siglip2-base-patch16-naflex) 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.
+        num_patches (`int`, *optional*, defaults to 256):
+            The number of patches in the image with the size of (`patch_size`, `patch_size`).
+            The image is resized to fill maximum of this number of patches, and to preserve
+            the aspect ratio. In case the resulted number of patches is lower, the image is
+            padded in "patch" dimension.
+        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 Siglip2VisionConfig, Siglip2VisionModel
+
+    >>> # Initializing a Siglip2VisionConfig with google/siglip2-base-patch16-naflex style configuration
+    >>> configuration = Siglip2VisionConfig()
+
+    >>> # Initializing a Siglip2VisionModel (with random weights) from the google/siglip2-base-patch16-naflex style configuration
+    >>> model = Siglip2VisionModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip2_vision_model"
+    base_config_key = "vision_config"
+
+    def __init__(
+        self,
+        hidden_size=1152,
+        intermediate_size=4304,
+        num_hidden_layers=27,
+        num_attention_heads=16,
+        num_channels=3,
+        num_patches=256,
+        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 = 384 #fixme
+        self.attention_dropout = attention_dropout
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+        self.num_patches = num_patches
+
+
+class Siglip2Config(PretrainedConfig):
+    r"""
+    [`Siglip2Config`] is the configuration class to store the configuration of a [`Siglip2Model`]. It is used to
+    instantiate a Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-224](https://huggingface.co/google/siglip2-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 [`Siglip2TextConfig`].
+        vision_config (`dict`, *optional*):
+            Dictionary of configuration options used to initialize [`Siglip2VisionConfig`].
+        kwargs (*optional*):
+            Dictionary of keyword arguments.
+
+    Example:
+
+    ```python
+    >>> from transformers import Siglip2Config, Siglip2Model
+
+    >>> # Initializing a Siglip2Config with google/siglip2-base-patch16-224 style configuration
+    >>> configuration = Siglip2Config()
+
+    >>> # Initializing a Siglip2Model (with random weights) from the google/siglip2-base-patch16-224 style configuration
+    >>> model = Siglip2Model(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+
+    >>> # We can also initialize a Siglip2Config from a Siglip2TextConfig and a Siglip2VisionConfig
+    >>> from transformers import Siglip2TextConfig, Siglip2VisionConfig
+
+    >>> # Initializing a Siglip2Text and Siglip2Vision configuration
+    >>> config_text = Siglip2TextConfig()
+    >>> config_vision = Siglip2VisionConfig()
+
+    >>> config = Siglip2Config.from_text_vision_configs(config_text, config_vision)
+    ```"""
+
+    model_type = "siglip2"
+    sub_configs = {"text_config": Siglip2TextConfig, "vision_config": Siglip2VisionConfig}
+
+    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 `Siglip2TextConfig` with default values.")
+
+        if vision_config is None:
+            vision_config = {}
+            logger.info("`vision_config` is `None`. initializing the `Siglip2VisionConfig` with default values.")
+
+        self.text_config = Siglip2TextConfig(**text_config)
+        self.vision_config = Siglip2VisionConfig(**vision_config)
+
+        self.initializer_factor = 1.0
+
+    @classmethod
+    def from_text_vision_configs(cls, text_config: Siglip2TextConfig, vision_config: Siglip2VisionConfig, **kwargs):
+        r"""
+        Instantiate a [`Siglip2Config`] (or a derived class) from siglip2 text model configuration and siglip2 vision
+        model configuration.
+
+        Returns:
+            [`Siglip2Config`]: An instance of a configuration object
+        """
+
+        return cls(text_config=text_config.to_dict(), vision_config=vision_config.to_dict(), **kwargs)
+    r"""
+    This is the configuration class to store the configuration of a [`Siglip2VisionModel`]. It is used to instantiate a
+    Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-naflex](https://huggingface.co/google/siglip2-base-patch16-naflex) 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.
+        num_patches (`int`, *optional*, defaults to 256):
+            The number of patches in the image with the size of (`patch_size`, `patch_size`).
+            The image is resized to fill maximum of this number of patches, and to preserve
+            the aspect ratio. In case the resulted number of patches is lower, the image is
+            padded in "patch" dimension.
+        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 Siglip2VisionConfig, Siglip2VisionModel
+
+    >>> # Initializing a Siglip2VisionConfig with google/siglip2-base-patch16-naflex style configuration
+    >>> configuration = Siglip2VisionConfig()
+
+    >>> # Initializing a Siglip2VisionModel (with random weights) from the google/siglip2-base-patch16-naflex style configuration
+    >>> model = Siglip2VisionModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip2_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,
+        num_patches=256,
+        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.attention_dropout = attention_dropout
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+        self.num_patches = num_patches
+
+logger = logging.get_logger(__name__)
+
+# General docstring
+_CONFIG_FOR_DOC = "Siglip2VisionConfig"
+
+
+@dataclass
+class Siglip2VisionOutput(ModelOutput):
+    """
+    Base class for vision model's outputs that also contains image embeddings of the pooling of the last hidden states.
+
+    Args:
+        image_embeds (`torch.FloatTensor` of shape `(batch_size, output_dim)` *optional* returned when model is initialized with `with_projection=True`):
+            The image embeddings obtained by applying the projection layer to the pooler_output.
+        last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
+            Sequence of hidden-states at the output of the last layer of the model.
+        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
+            Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
+            one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.
+
+            Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
+        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
+            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
+            sequence_length)`.
+
+            Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
+            heads.
+    """
+
+    image_embeds: Optional[torch.FloatTensor] = None
+    last_hidden_state: torch.FloatTensor = None
+    hidden_states: Optional[Tuple[torch.FloatTensor, ...]] = None
+    attentions: Optional[Tuple[torch.FloatTensor, ...]] = None
+
+
+class Siglip2VisionEmbeddings(nn.Module):
+    def __init__(self, config: Siglip2VisionConfig):
+        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.Linear(
+            in_features=config.num_channels * self.patch_size * self.patch_size,
+            out_features=self.embed_dim,
+        )
+
+        self.num_patches = config.num_patches
+        self.position_embedding_size = int(self.num_patches**0.5)
+        self.position_embedding = nn.Embedding(self.num_patches, self.embed_dim)
+    
+    @staticmethod
+    def resize_positional_embeddings(
+        positional_embeddings: torch.Tensor,
+        spatial_shapes: torch.LongTensor,
+        max_length: int,
+    ) -> torch.Tensor:
+        """
+        Resize positional embeddings to image-specific size and pad to a fixed size.
+
+        Args:
+            positional_embeddings (`torch.Tensor`):
+                Position embeddings of shape (height, width, embed_dim)
+            spatial_shapes (`torch.LongTensor`):
+                Spatial shapes of shape (batch_size, 2) to resize the positional embeddings to
+            max_length (`int`):
+                Maximum length of the positional embeddings to pad resized positional embeddings to
+
+        Returns:
+            `torch.Tensor`: Embeddings of shape (batch_size, max_length, embed_dim)
+        """
+        # import pdb; pdb.set_trace()
+        batch_size = spatial_shapes.shape[0]
+        embed_dim = positional_embeddings.shape[-1]
+        source_dtype = positional_embeddings.dtype
+
+        resulted_positional_embeddings = torch.empty(
+            (batch_size, max_length, embed_dim),
+            device=positional_embeddings.device,
+            dtype=source_dtype,
+        )
+
+        # (height, width, embed_dim) -> (1, embed_dim, height, width) for interpolation
+        positional_embeddings = positional_embeddings.permute(2, 0, 1).unsqueeze(0)
+
+        # Upcast to float32 on CPU because antialias is not supported for bfloat16/float16 on CPU
+        if positional_embeddings.device.type == "cpu":
+            positional_embeddings = positional_embeddings.to(torch.float32)
+
+        for i in range(batch_size):
+            # (1, dim, height, width) -> (1, dim, target_height, target_width)
+            height, width = spatial_shapes[i]
+            resized_embeddings = F.interpolate(
+                positional_embeddings,
+                size=(height, width),
+                mode="bilinear",
+                align_corners=False,
+                antialias=True,
+            )
+
+            # (1, dim, target_height, target_width) -> (target_height * target_width, dim)
+            resized_embeddings = resized_embeddings.reshape(embed_dim, height * width).transpose(0, 1)
+
+            # Cast to original dtype
+            resized_embeddings = resized_embeddings.to(source_dtype)
+
+            resulted_positional_embeddings[i, : height * width] = resized_embeddings
+            resulted_positional_embeddings[i, height * width :] = resized_embeddings[0]
+
+        return resulted_positional_embeddings
+
+    def forward(self, pixel_values: torch.FloatTensor, spatial_shapes: torch.LongTensor) -> torch.Tensor:
+        """
+        Args:
+            pixel_values (`torch.FloatTensor`):
+                Pixel values of shape (batch_size, max_num_patches, num_channels * patch_size * patch_size)
+            spatial_shapes (`List[Tuple[int, int]]`):
+                Spatial shapes of shape (batch_size, 2) to resize the positional embeddings to
+        """
+        # import pdb; pdb.set_trace()
+        # Apply patch embeddings to already patchified pixel values
+        target_dtype = self.patch_embedding.weight.dtype
+        patch_embeds = self.patch_embedding(pixel_values.to(dtype=target_dtype))
+
+        # Get positional resized and padded positional embeddings
+        positional_embeddings = self.position_embedding.weight.reshape(
+            self.position_embedding_size, self.position_embedding_size, -1
+        )
+        resized_positional_embeddings = self.resize_positional_embeddings(
+            positional_embeddings, spatial_shapes, max_length=pixel_values.shape[1]
+        )
+
+        # Add positional embeddings to patch embeddings
+        embeddings = patch_embeds + resized_positional_embeddings
+        return embeddings
+
+
+class Siglip2Attention(nn.Module):
+    """Multi-headed attention from 'Attention Is All You Need' paper"""
+
+    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 Siglip2SdpaAttention(Siglip2Attention):
+    """
+    Siglip2 attention module using torch.nn.functional.scaled_dot_product_attention. This module inherits from
+    `Siglip2Attention` 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 Siglip2Attention.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(
+                "Siglip2Model is using Siglip2SdpaAttention, 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
+
+class Siglip2FlashAttention2(Siglip2Attention):
+    """
+    Siglip2Attention flash attention module. This module inherits from `Siglip2Attention` 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)
+        key_states = key_states.view(batch_size, q_len, self.num_heads, self.head_dim)
+        value_states = value_states.view(batch_size, q_len, self.num_heads, self.head_dim)
+
+        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 Siglip2MLP(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
+
+
+SIGLIP2_ATTENTION_CLASSES = {
+    "eager": Siglip2Attention,
+    "sdpa": Siglip2SdpaAttention,
+    "flash_attention_2": Siglip2FlashAttention2,
+}
+
+
+class Siglip2EncoderLayer(nn.Module):
+    def __init__(self, config: Siglip2Config):
+        super().__init__()
+        self.embed_dim = config.hidden_size
+        self.self_attn = SIGLIP2_ATTENTION_CLASSES[config._attn_implementation](config=config)
+        self.layer_norm1 = nn.LayerNorm(self.embed_dim, eps=config.layer_norm_eps)
+        self.mlp = Siglip2MLP(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 Siglip2Encoder(nn.Module):
+    """
+    Transformer encoder consisting of `config.num_hidden_layers` self attention layers. Each layer is a
+    [`Siglip2EncoderLayer`].
+
+    Args:
+        config: Siglip2Config
+    """
+
+    def __init__(self, config: Siglip2Config):
+        super().__init__()
+        self.config = config
+        self.layers = nn.ModuleList([Siglip2EncoderLayer(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
+        )
+
+
+SIGLIP2_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.
+"""
+
+
+class Siglip2VisionTransformer(nn.Module):
+    def __init__(self, config: Siglip2VisionConfig):
+        super().__init__()
+        self.config = config
+        embed_dim = config.hidden_size
+
+        self.embeddings = Siglip2VisionEmbeddings(config)
+        self.encoder = Siglip2Encoder(config)
+        self.post_layernorm = nn.LayerNorm(embed_dim, eps=config.layer_norm_eps)
+        self.use_head = False if not hasattr(config, "vision_use_head") else config.vision_use_head
+        self._use_flash_attention_2 = config._attn_implementation == "flash_attention_2"        
+
+    @add_start_docstrings_to_model_forward(SIGLIP2_VISION_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=Siglip2VisionConfig)
+    def forward(
+        self,
+        pixel_values,
+        attention_mask: torch.Tensor,
+        spatial_shapes: torch.LongTensor,
+        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
+
+        hidden_states = self.embeddings(pixel_values, spatial_shapes)
+        # import pdb; pdb.set_trace()
+        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]
+            encoder_attention_mask = _prepare_4d_attention_mask(attention_mask, hidden_states.dtype)
+        else:
+            encoder_attention_mask = attention_mask
+
+        encoder_outputs = self.encoder(
+            inputs_embeds=hidden_states,
+            attention_mask=encoder_attention_mask,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        last_hidden_state = encoder_outputs[0]
+        last_hidden_state = self.post_layernorm(last_hidden_state)
+
+        pooled_output = self.head(last_hidden_state)
+
+        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,
+        )
+
+
+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")
+
+
+class Siglip2PreTrainedModel(PreTrainedModel):
+    """
+    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
+    models.
+    """
+
+    config_class = Siglip2Config
+    base_model_prefix = "siglip2"
+    supports_gradient_checkpointing = True
+
+    _no_split_modules = [
+        "Siglip2TextEmbeddings",
+        "Siglip2EncoderLayer",
+        "Siglip2VisionEmbeddings",
+        "Siglip2EncoderLayer",
+        "Siglip2MultiheadAttentionPoolingHead",
+    ]
+    _supports_flash_attn_2 = True
+    _supports_sdpa = True
+
+    def _init_weights(self, module):
+        """Initialize the weights"""
+        if isinstance(module, Siglip2VisionEmbeddings):
+            width = self.config.hidden_size
+            nn.init.normal_(module.position_embedding.weight, std=1 / np.sqrt(width))
+        elif isinstance(module, nn.Embedding):
+            default_flax_embed_init(module.weight)
+        elif isinstance(module, Siglip2Attention):
+            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, Siglip2MLP):
+            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, (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)
+
+
+class Siglip2VisionModel(Siglip2PreTrainedModel):
+    config_class = Siglip2VisionConfig
+    main_input_name = "pixel_values"
+
+    def __init__(self, config: Siglip2VisionConfig):
+        super().__init__(config)
+
+        self.vision_model = Siglip2VisionTransformer(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(SIGLIP2_VISION_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=Siglip2VisionConfig)
+    def forward(
+        self,
+        pixel_values: torch.FloatTensor,
+        pixel_attention_mask: torch.Tensor,
+        spatial_shapes: torch.LongTensor,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        Returns:
+
+        Examples:
+
+        ```python
+        >>> from PIL import Image
+        >>> import requests
+        >>> from transformers import AutoProcessor, Siglip2VisionModel
+
+        >>> model = Siglip2VisionModel.from_pretrained("google/siglip2-base-patch16-224")
+        >>> processor = AutoProcessor.from_pretrained("google/siglip2-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,
+            attention_mask=pixel_attention_mask,
+            spatial_shapes=spatial_shapes,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+
+class SigLip2VisionTower(nn.Module):
+    def __init__(self, vision_tower, vision_tower_cfg, delay_load=False):
+        super().__init__()
+
+        self.is_loaded = False
+
+        self.config = Siglip2VisionConfig()
+
+        self.vision_tower_name = vision_tower
+
+        self.image_processor = SigLipImageProcessor()
+
+        if not delay_load:
+            rank0_print(f"Loading vision tower: {vision_tower}")
+            self.load_model()
+        elif getattr(vision_tower_cfg, "unfreeze_mm_vision_tower", False):
+            # TODO: better detector is needed.
+            rank0_print(f"The checkpoint seems to contain `vision_tower` weights: `unfreeze_mm_vision_tower`: True.")
+            self.load_model()
+        elif hasattr(vision_tower_cfg, "mm_tunable_parts") and "mm_vision_tower" in vision_tower_cfg.mm_tunable_parts:
+            rank0_print(f"The checkpoint seems to contain `vision_tower` weights: `mm_tunable_parts` contains `mm_vision_tower`.")
+            self.load_model()
+        else:
+            self.cfg_only = self.config
+
+    def load_model(self, device_map=None):
+        if self.is_loaded:
+            rank0_print("{} is already loaded, `load_model` called again, skipping.".format(self.vision_tower_name))
+            return
+        # self.vision_tower = Siglip2VisionModel.from_pretrained(self.vision_tower_name, device_map=device_map, ignore_mismatched_sizes=True)
+        self.vision_tower = Siglip2VisionModel.from_pretrained(self.vision_tower_name, device_map=device_map)
+        # import pdb; pdb.set_trace()
+        print('siglip2_naflex')
+        self.vision_tower.vision_model.head = nn.Identity()
+        self.vision_tower.requires_grad_(False)
+        self.is_loaded = True
+
+    def forward(self, images, patch_sizes):
+        if type(images) is list: 
+            # image_features = []
+            # for image, spatial_shape in zip(images, patch_sizes):
+            #     spatial_shape = torch.as_tensor(spatial_shape)[None]
+            #     image_forward_out = self.vision_tower(image.to(device=self.device, dtype=self.dtype).unsqueeze(0), spatial_shapes=spatial_shape, output_hidden_states=True)
+            #     # 根据崇屹经验，siglip2和siglip都是最后一层
+            #     image_feature = image_forward_out.last_hidden_state.to(image.dtype)
+            #     # hidden_states的长度为 ViT层数+1，因为patching后的embedding也有
+            #     # image_feature = image_forward_out.hidden_states[-2].to(image.dtype) 
+            #     image_features.append(image_feature)
+            pixel_values = []
+            pixel_attention_masks = []
+            spatial_shapes = []
+            max_length = max([patch_size[0] * patch_size[1] for patch_size in patch_sizes])
+            encoder_patch_size = self.vision_tower.vision_model.embeddings.patch_size
+            for image, spatial_shape in zip(images, patch_sizes):
+                valid_pixel_num = spatial_shape[0] * spatial_shape[1]
+                spatial_shape = torch.as_tensor(spatial_shape)[None]
+                image = image.to(device=self.device, dtype=self.dtype).unsqueeze(0)
+                pixel_value = rearrange(image, 'b c (h p1) (w p2) -> b (h w) (p1 p2 c)', p1=encoder_patch_size, p2=encoder_patch_size)
+                # b, n, c
+                padding_pixel = torch.zeros_like(pixel_value)[:, :1]
+                pixel_value = torch.cat([pixel_value, padding_pixel.repeat(1, max_length - valid_pixel_num, 1)], dim=1)
+                pixel_attention_mask = torch.zeros_like(pixel_value[:, :, 0])
+                pixel_attention_mask[:, :valid_pixel_num] = 1
+                
+                pixel_values.append(pixel_value)
+                pixel_attention_masks.append(pixel_attention_mask)
+                spatial_shapes.append(spatial_shape)
+            
+            pixel_values = torch.cat(pixel_values)
+            pixel_attention_masks = torch.cat(pixel_attention_masks)
+            spatial_shapes = torch.cat(spatial_shapes)
+
+            image_forward_outs = self.vision_tower(pixel_values, 
+                                                   pixel_attention_mask=pixel_attention_masks, 
+                                                   spatial_shapes=spatial_shapes, output_hidden_states=True)
+
+            # import pdb; pdb.set_trace()
+            image_features = image_forward_outs.last_hidden_state.to(pixel_values.dtype)
+            image_features = image_features.split(1)
+            # 应该为list
+        
+        else: #torch.Size([1, 3, 384, 384])
+            print('no support for paralla')
+            exit()
+            image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype),spatial_shapes=patch_sizes, output_hidden_states=True)
+            image_features = image_forward_outs.last_hidden_state.to(images.dtype)
+            # image_features = image_forward_outs.hidden_states[-2].to(images.dtype)
+
+        return image_features
+    
+    # def forward_uhd_v2(self, images, patch_sizes):
+    #     if type(images) is list:
+    #         image_features = []
+    #         for image in images:
+    #             image_forward_out = self.vision_tower(image.to(device=self.device, dtype=self.dtype).unsqueeze(0), spatial_shapes=patch_sizes,output_hidden_states=True)
+    #             image_feature = image_forward_out.hidden_states[-1].to(image.dtype)
+    #             image_features.append(image_feature)
+    #     else:
+    #         image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype), spatial_shapes=patch_sizes,output_hidden_states=True)
+    #         image_features = image_forward_outs.hidden_states[-1].to(images.dtype)
+
+    #     return image_features
+
+    @property
+    def dummy_feature(self):
+        return torch.zeros(1, self.hidden_size, device=self.device, dtype=self.dtype)
+
+    @property
+    def dtype(self):
+        for p in self.vision_tower.parameters():
+            return p.dtype
+
+    @property
+    def device(self):
+        for p in self.vision_tower.parameters():
+            return p.device
+
+    @property
+    def hidden_size(self):
+        return self.config.hidden_size
+
+    @property
+    def num_patches(self):
+        return (self.config.image_size // self.config.patch_size) ** 2
+
+    @property
+    def num_patches_per_side(self):
+        return self.config.image_size // self.config.patch_size
+        # return self.model_config["vision_cfg"]["image_size"] // self.model_config["vision_cfg"]["patch_size"]
+
+    @property
+    def image_size(self):
+        return self.config.image_size

VLMEvalKit-sudoku/llava/model/multimodal_encoder/modeling_siglip2_cjm.py

@@ -0,0 +1,1187 @@
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+#           This file was automatically generated from src/transformers/models/siglip2/modular_siglip2.py.
+#               Do NOT edit this file manually as any edits will be overwritten by the generation of
+#             the file from the modular. If any change should be done, please apply the change to the
+#                          modular_siglip2.py file directly. One of our CI enforces this.
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+# coding=utf-8
+# Copyright 2025 The HuggingFace Inc. team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import math
+import warnings
+from dataclasses import dataclass
+from typing import Any, Optional, Tuple, Union
+
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+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,
+)
+from transformers.configuration_utils import PretrainedConfig
+
+
+class Siglip2TextConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`Siglip2TextModel`]. It is used to instantiate a
+    Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-224](https://huggingface.co/google/siglip2-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 Siglip2 text model. Defines the number of different tokens that can be represented by
+            the `inputs_ids` passed when calling [`Siglip2Model`].
+        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.
+        projection_size (`int`, *optional*, defaults to `hidden_size`):
+            The size of the projection head.
+
+    Example:
+
+    ```python
+    >>> from transformers import Siglip2TextConfig, Siglip2TextModel
+
+    >>> # Initializing a Siglip2TextConfig with google/siglip2-base-patch16-224 style configuration
+    >>> configuration = Siglip2TextConfig()
+
+    >>> # Initializing a Siglip2TextModel (with random weights) from the google/siglip2-base-patch16-224 style configuration
+    >>> model = Siglip2TextModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip2_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/siglip2
+        # See https://github.com/huggingface/transformers/pull/24773#issuecomment-1632287538
+        pad_token_id=1,
+        bos_token_id=49406,
+        eos_token_id=49407,
+        projection_size=None,
+        **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
+        self.projection_size = projection_size if projection_size is not None else hidden_size
+
+
+class Siglip2VisionConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`Siglip2VisionModel`]. It is used to instantiate a
+    Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-naflex](https://huggingface.co/google/siglip2-base-patch16-naflex) 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.
+        num_patches (`int`, *optional*, defaults to 256):
+            The number of patches in the image with the size of (`patch_size`, `patch_size`).
+            The image is resized to fill maximum of this number of patches, and to preserve
+            the aspect ratio. In case the resulted number of patches is lower, the image is
+            padded in "patch" dimension.
+        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 Siglip2VisionConfig, Siglip2VisionModel
+
+    >>> # Initializing a Siglip2VisionConfig with google/siglip2-base-patch16-naflex style configuration
+    >>> configuration = Siglip2VisionConfig()
+
+    >>> # Initializing a Siglip2VisionModel (with random weights) from the google/siglip2-base-patch16-naflex style configuration
+    >>> model = Siglip2VisionModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip2_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,
+        num_patches=256,
+        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.attention_dropout = attention_dropout
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+        self.num_patches = num_patches
+
+
+class Siglip2Config(PretrainedConfig):
+    r"""
+    [`Siglip2Config`] is the configuration class to store the configuration of a [`Siglip2Model`]. It is used to
+    instantiate a Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-224](https://huggingface.co/google/siglip2-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 [`Siglip2TextConfig`].
+        vision_config (`dict`, *optional*):
+            Dictionary of configuration options used to initialize [`Siglip2VisionConfig`].
+        kwargs (*optional*):
+            Dictionary of keyword arguments.
+
+    Example:
+
+    ```python
+    >>> from transformers import Siglip2Config, Siglip2Model
+
+    >>> # Initializing a Siglip2Config with google/siglip2-base-patch16-224 style configuration
+    >>> configuration = Siglip2Config()
+
+    >>> # Initializing a Siglip2Model (with random weights) from the google/siglip2-base-patch16-224 style configuration
+    >>> model = Siglip2Model(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+
+    >>> # We can also initialize a Siglip2Config from a Siglip2TextConfig and a Siglip2VisionConfig
+    >>> from transformers import Siglip2TextConfig, Siglip2VisionConfig
+
+    >>> # Initializing a Siglip2Text and Siglip2Vision configuration
+    >>> config_text = Siglip2TextConfig()
+    >>> config_vision = Siglip2VisionConfig()
+
+    >>> config = Siglip2Config.from_text_vision_configs(config_text, config_vision)
+    ```"""
+
+    model_type = "siglip2"
+    sub_configs = {"text_config": Siglip2TextConfig, "vision_config": Siglip2VisionConfig}
+
+    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 `Siglip2TextConfig` with default values.")
+
+        if vision_config is None:
+            vision_config = {}
+            logger.info("`vision_config` is `None`. initializing the `Siglip2VisionConfig` with default values.")
+
+        self.text_config = Siglip2TextConfig(**text_config)
+        self.vision_config = Siglip2VisionConfig(**vision_config)
+
+        self.initializer_factor = 1.0
+
+    @classmethod
+    def from_text_vision_configs(cls, text_config: Siglip2TextConfig, vision_config: Siglip2VisionConfig, **kwargs):
+        r"""
+        Instantiate a [`Siglip2Config`] (or a derived class) from siglip2 text model configuration and siglip2 vision
+        model configuration.
+
+        Returns:
+            [`Siglip2Config`]: An instance of a configuration object
+        """
+
+        return cls(text_config=text_config.to_dict(), vision_config=vision_config.to_dict(), **kwargs)
+    r"""
+    This is the configuration class to store the configuration of a [`Siglip2VisionModel`]. It is used to instantiate a
+    Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-naflex](https://huggingface.co/google/siglip2-base-patch16-naflex) 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.
+        num_patches (`int`, *optional*, defaults to 256):
+            The number of patches in the image with the size of (`patch_size`, `patch_size`).
+            The image is resized to fill maximum of this number of patches, and to preserve
+            the aspect ratio. In case the resulted number of patches is lower, the image is
+            padded in "patch" dimension.
+        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 Siglip2VisionConfig, Siglip2VisionModel
+
+    >>> # Initializing a Siglip2VisionConfig with google/siglip2-base-patch16-naflex style configuration
+    >>> configuration = Siglip2VisionConfig()
+
+    >>> # Initializing a Siglip2VisionModel (with random weights) from the google/siglip2-base-patch16-naflex style configuration
+    >>> model = Siglip2VisionModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip2_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,
+        num_patches=256,
+        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.attention_dropout = attention_dropout
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+        self.num_patches = num_patches
+
+
+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 = "Siglip2VisionConfig"
+
+
+@dataclass
+class Siglip2VisionOutput(ModelOutput):
+    """
+    Base class for vision model's outputs that also contains image embeddings of the pooling of the last hidden states.
+
+    Args:
+        image_embeds (`torch.FloatTensor` of shape `(batch_size, output_dim)` *optional* returned when model is initialized with `with_projection=True`):
+            The image embeddings obtained by applying the projection layer to the pooler_output.
+        last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
+            Sequence of hidden-states at the output of the last layer of the model.
+        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
+            Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
+            one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.
+
+            Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
+        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
+            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
+            sequence_length)`.
+
+            Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
+            heads.
+    """
+
+    image_embeds: Optional[torch.FloatTensor] = None
+    last_hidden_state: torch.FloatTensor = None
+    hidden_states: Optional[Tuple[torch.FloatTensor, ...]] = None
+    attentions: Optional[Tuple[torch.FloatTensor, ...]] = None
+
+
+class Siglip2VisionEmbeddings(nn.Module):
+    def __init__(self, config: Siglip2VisionConfig):
+        super().__init__()
+        self.config = config
+        self.embed_dim = config.hidden_size
+        self.patch_size = config.patch_size
+
+        self.patch_embedding = nn.Linear(
+            in_features=config.num_channels * self.patch_size * self.patch_size,
+            out_features=self.embed_dim,
+        )
+
+        self.num_patches = config.num_patches
+        self.position_embedding_size = int(self.num_patches**0.5)
+        self.position_embedding = nn.Embedding(self.num_patches, self.embed_dim)
+
+    @staticmethod
+    def resize_positional_embeddings(
+        positional_embeddings: torch.Tensor,
+        spatial_shapes: torch.LongTensor,
+        max_length: int,
+    ) -> torch.Tensor:
+        """
+        Resize positional embeddings to image-specific size and pad to a fixed size.
+
+        Args:
+            positional_embeddings (`torch.Tensor`):
+                Position embeddings of shape (height, width, embed_dim)
+            spatial_shapes (`torch.LongTensor`):
+                Spatial shapes of shape (batch_size, 2) to resize the positional embeddings to
+            max_length (`int`):
+                Maximum length of the positional embeddings to pad resized positional embeddings to
+
+        Returns:
+            `torch.Tensor`: Embeddings of shape (batch_size, max_length, embed_dim)
+        """
+        batch_size = spatial_shapes.shape[0]
+        embed_dim = positional_embeddings.shape[-1]
+        source_dtype = positional_embeddings.dtype
+
+        resulted_positional_embeddings = torch.empty(
+            (batch_size, max_length, embed_dim),
+            device=positional_embeddings.device,
+            dtype=source_dtype,
+        )
+
+        # (height, width, embed_dim) -> (1, embed_dim, height, width) for interpolation
+        positional_embeddings = positional_embeddings.permute(2, 0, 1).unsqueeze(0)
+
+        # Upcast to float32 on CPU because antialias is not supported for bfloat16/float16 on CPU
+        if positional_embeddings.device.type == "cpu":
+            positional_embeddings = positional_embeddings.to(torch.float32)
+
+        for i in range(batch_size):
+            # (1, dim, height, width) -> (1, dim, target_height, target_width)
+            height, width = spatial_shapes[i]
+            resized_embeddings = F.interpolate(
+                positional_embeddings,
+                size=(height, width),
+                mode="bilinear",
+                align_corners=False,
+                antialias=True,
+            )
+
+            # (1, dim, target_height, target_width) -> (target_height * target_width, dim)
+            resized_embeddings = resized_embeddings.reshape(embed_dim, height * width).transpose(0, 1)
+
+            # Cast to original dtype
+            resized_embeddings = resized_embeddings.to(source_dtype)
+
+            resulted_positional_embeddings[i, : height * width] = resized_embeddings
+            resulted_positional_embeddings[i, height * width :] = resized_embeddings[0]
+
+        return resulted_positional_embeddings
+
+    def forward(self, pixel_values: torch.FloatTensor, spatial_shapes: torch.LongTensor) -> torch.Tensor:
+        """
+        Args:
+            pixel_values (`torch.FloatTensor`):
+                Pixel values of shape (batch_size, max_num_patches, num_channels * patch_size * patch_size)
+            spatial_shapes (`List[Tuple[int, int]]`):
+                Spatial shapes of shape (batch_size, 2) to resize the positional embeddings to
+        """
+
+        # Apply patch embeddings to already patchified pixel values
+        target_dtype = self.patch_embedding.weight.dtype
+        patch_embeds = self.patch_embedding(pixel_values.to(dtype=target_dtype))
+
+        # Get positional resized and padded positional embeddings
+        positional_embeddings = self.position_embedding.weight.reshape(
+            self.position_embedding_size, self.position_embedding_size, -1
+        )
+        resized_positional_embeddings = self.resize_positional_embeddings(
+            positional_embeddings, spatial_shapes, max_length=pixel_values.shape[1]
+        )
+
+        # Add positional embeddings to patch embeddings
+        embeddings = patch_embeds + resized_positional_embeddings
+        return embeddings
+
+
+class Siglip2Attention(nn.Module):
+    """Multi-headed attention from 'Attention Is All You Need' paper"""
+
+    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 Siglip2FlashAttention2(Siglip2Attention):
+    """
+    Siglip2Attention flash attention module. This module inherits from `Siglip2Attention` 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)
+        key_states = key_states.view(batch_size, q_len, self.num_heads, self.head_dim)
+        value_states = value_states.view(batch_size, q_len, self.num_heads, self.head_dim)
+
+        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 Siglip2SdpaAttention(Siglip2Attention):
+    """
+    Siglip2 attention module using torch.nn.functional.scaled_dot_product_attention. This module inherits from
+    `Siglip2Attention` 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 Siglip2Attention.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(
+                "Siglip2Model is using Siglip2SdpaAttention, 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
+
+
+class Siglip2MLP(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
+
+
+SIGLIP2_ATTENTION_CLASSES = {
+    "eager": Siglip2Attention,
+    "flash_attention_2": Siglip2FlashAttention2,
+    "sdpa": Siglip2SdpaAttention,
+}
+
+
+class Siglip2EncoderLayer(nn.Module):
+    def __init__(self, config: Siglip2Config):
+        super().__init__()
+        self.embed_dim = config.hidden_size
+        self.self_attn = SIGLIP2_ATTENTION_CLASSES[config._attn_implementation](config=config)
+        self.layer_norm1 = nn.LayerNorm(self.embed_dim, eps=config.layer_norm_eps)
+        self.mlp = Siglip2MLP(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 Siglip2Encoder(nn.Module):
+    """
+    Transformer encoder consisting of `config.num_hidden_layers` self attention layers. Each layer is a
+    [`Siglip2EncoderLayer`].
+
+    Args:
+        config: Siglip2Config
+    """
+
+    def __init__(self, config: Siglip2Config):
+        super().__init__()
+        self.config = config
+        self.layers = nn.ModuleList([Siglip2EncoderLayer(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
+        )
+
+
+SIGLIP2_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.
+"""
+
+
+class Siglip2VisionTransformer(nn.Module):
+    def __init__(self, config: Siglip2VisionConfig):
+        super().__init__()
+        self.config = config
+        embed_dim = config.hidden_size
+
+        self.embeddings = Siglip2VisionEmbeddings(config)
+        self.encoder = Siglip2Encoder(config)
+        self.post_layernorm = nn.LayerNorm(embed_dim, eps=config.layer_norm_eps)
+        self.use_head = False if not hasattr(config, "vision_use_head") else config.vision_use_head
+        self._use_flash_attention_2 = config._attn_implementation == "flash_attention_2"
+
+    @add_start_docstrings_to_model_forward(SIGLIP2_VISION_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=Siglip2VisionConfig)
+    def forward(
+        self,
+        pixel_values: torch.FloatTensor,
+        attention_mask: torch.Tensor,
+        spatial_shapes: torch.LongTensor,
+        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
+
+        hidden_states = self.embeddings(pixel_values, spatial_shapes)
+
+        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]
+            encoder_attention_mask = _prepare_4d_attention_mask(attention_mask, hidden_states.dtype)
+        else:
+            encoder_attention_mask = attention_mask
+
+        encoder_outputs = self.encoder(
+            inputs_embeds=hidden_states,
+            attention_mask=encoder_attention_mask,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        last_hidden_state = encoder_outputs[0]
+        last_hidden_state = self.post_layernorm(last_hidden_state)
+
+        pooler_output = self.head(last_hidden_state, attention_mask) 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,
+        )
+
+
+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")
+
+
+class Siglip2PreTrainedModel(PreTrainedModel):
+    """
+    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
+    models.
+    """
+
+    config_class = Siglip2Config
+    base_model_prefix = "siglip2"
+    supports_gradient_checkpointing = True
+
+    _no_split_modules = [
+        "Siglip2TextEmbeddings",
+        "Siglip2EncoderLayer",
+        "Siglip2VisionEmbeddings",
+        "Siglip2EncoderLayer",
+        "Siglip2MultiheadAttentionPoolingHead",
+    ]
+    _supports_flash_attn_2 = True
+    _supports_sdpa = True
+
+    def _init_weights(self, module):
+        """Initialize the weights"""
+        if isinstance(module, Siglip2VisionEmbeddings):
+            width = self.config.hidden_size
+            nn.init.normal_(module.position_embedding.weight, std=1 / np.sqrt(width))
+        elif isinstance(module, nn.Embedding):
+            default_flax_embed_init(module.weight)
+        elif isinstance(module, Siglip2Attention):
+            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, Siglip2MLP):
+            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, (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)
+
+
+class Siglip2VisionModel(Siglip2PreTrainedModel):
+    config_class = Siglip2VisionConfig
+    main_input_name = "pixel_values"
+
+    def __init__(self, config: Siglip2VisionConfig):
+        super().__init__(config)
+
+        self.vision_model = Siglip2VisionTransformer(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(SIGLIP2_VISION_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=Siglip2VisionConfig)
+    def forward(
+        self,
+        pixel_values: torch.FloatTensor,
+        pixel_attention_mask: torch.Tensor,
+        spatial_shapes: torch.LongTensor,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        Returns:
+
+        Examples:
+
+        ```python
+        >>> from PIL import Image
+        >>> import requests
+        >>> from transformers import AutoProcessor, Siglip2VisionModel
+
+        >>> model = Siglip2VisionModel.from_pretrained("google/siglip2-base-patch16-224")
+        >>> processor = AutoProcessor.from_pretrained("google/siglip2-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,
+            attention_mask=pixel_attention_mask,
+            spatial_shapes=spatial_shapes,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )

VLMEvalKit-sudoku/llava/model/multimodal_encoder/modeling_swin_siglip2.py

@@ -0,0 +1,1319 @@
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+#           This file was automatically generated from src/transformers/models/siglip2/modular_siglip2.py.
+#               Do NOT edit this file manually as any edits will be overwritten by the generation of
+#             the file from the modular. If any change should be done, please apply the change to the
+#                          modular_siglip2.py file directly. One of our CI enforces this.
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+# coding=utf-8
+# Copyright 2025 The HuggingFace Inc. team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import math
+import warnings
+from dataclasses import dataclass
+from functools import partial, reduce
+import torch.utils.checkpoint
+from PIL import Image
+from typing import Any, Optional, Tuple, Union, Dict
+import os
+
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+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,
+)
+from transformers.configuration_utils import PretrainedConfig
+from transformers.image_processing_utils import BatchFeature, get_size_dict
+from transformers.image_transforms import (
+    convert_to_rgb,
+    normalize,
+    rescale,
+    resize,
+    to_channel_dimension_format,
+)
+from transformers.image_utils import (
+    ChannelDimension,
+    PILImageResampling,
+    to_numpy_array,
+)
+from transformers.activations import ACT2FN
+from transformers.modeling_outputs import BaseModelOutput, BaseModelOutputWithPooling
+from transformers.modeling_utils import PreTrainedModel
+from transformers import PretrainedConfig
+from transformers.utils import ModelOutput
+from llava.utils import rank0_print
+from einops import rearrange
+
+class SigLipImageProcessor:
+    def __init__(self, image_mean=(0.5, 0.5, 0.5), image_std=(0.5, 0.5, 0.5), size=(384, 384), crop_size: Dict[str, int] = None, resample=PILImageResampling.BICUBIC, rescale_factor=1 / 255, data_format=ChannelDimension.FIRST):
+        crop_size = crop_size if crop_size is not None else {"height": 384, "width": 384}
+        crop_size = get_size_dict(crop_size, default_to_square=True, param_name="crop_size")
+
+        self.image_mean = image_mean
+        self.image_std = image_std
+        self.size = size
+        self.resample = resample
+        self.rescale_factor = rescale_factor
+        self.data_format = data_format
+        self.crop_size = crop_size
+
+    def preprocess(self, images, do_resize = True, do_center_crop = True, do_rescale = True, do_normalize = True, return_tensors = 'pt'):
+        if isinstance(images, Image.Image):
+            images = [images]
+        else:
+            # to adapt video data
+            images = [to_numpy_array(image) for image in images]
+            assert isinstance(images, list)
+
+        # do_resize=False, do_center_crop=False, do_rescale=True, do_normalize=True, 
+
+        transforms = [
+            convert_to_rgb,
+            to_numpy_array
+        ]
+
+        if do_resize:
+            transforms.append(partial(resize, size=self.size, resample=self.resample, data_format=self.data_format))
+        if do_rescale:
+            transforms.append(partial(rescale, scale=self.rescale_factor, data_format=self.data_format))
+        if do_normalize:
+            transforms.append(partial(normalize, mean=self.image_mean, std=self.image_std, data_format=self.data_format))
+        
+        transforms.append(partial(to_channel_dimension_format, channel_dim=self.data_format, input_channel_dim=self.data_format))
+
+        images = reduce(lambda x, f: [*map(f, x)], transforms, images)
+        data = {"pixel_values": images}
+
+        return BatchFeature(data=data, tensor_type=return_tensors)
+
+
+class Siglip2TextConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`Siglip2TextModel`]. It is used to instantiate a
+    Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-224](https://huggingface.co/google/siglip2-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 Siglip2 text model. Defines the number of different tokens that can be represented by
+            the `inputs_ids` passed when calling [`Siglip2Model`].
+        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.
+        projection_size (`int`, *optional*, defaults to `hidden_size`):
+            The size of the projection head.
+
+    Example:
+
+    ```python
+    >>> from transformers import Siglip2TextConfig, Siglip2TextModel
+
+    >>> # Initializing a Siglip2TextConfig with google/siglip2-base-patch16-224 style configuration
+    >>> configuration = Siglip2TextConfig()
+
+    >>> # Initializing a Siglip2TextModel (with random weights) from the google/siglip2-base-patch16-224 style configuration
+    >>> model = Siglip2TextModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip2_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/siglip2
+        # See https://github.com/huggingface/transformers/pull/24773#issuecomment-1632287538
+        pad_token_id=1,
+        bos_token_id=49406,
+        eos_token_id=49407,
+        projection_size=None,
+        **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
+        self.projection_size = projection_size if projection_size is not None else hidden_size
+
+
+class Siglip2VisionConfig(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`Siglip2VisionModel`]. It is used to instantiate a
+    Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-naflex](https://huggingface.co/google/siglip2-base-patch16-naflex) 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.
+        num_patches (`int`, *optional*, defaults to 256):
+            The number of patches in the image with the size of (`patch_size`, `patch_size`).
+            The image is resized to fill maximum of this number of patches, and to preserve
+            the aspect ratio. In case the resulted number of patches is lower, the image is
+            padded in "patch" dimension.
+        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 Siglip2VisionConfig, Siglip2VisionModel
+
+    >>> # Initializing a Siglip2VisionConfig with google/siglip2-base-patch16-naflex style configuration
+    >>> configuration = Siglip2VisionConfig()
+
+    >>> # Initializing a Siglip2VisionModel (with random weights) from the google/siglip2-base-patch16-naflex style configuration
+    >>> model = Siglip2VisionModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip2_vision_model"
+    base_config_key = "vision_config"
+
+    def __init__(
+        self,
+        hidden_size=1152,
+        intermediate_size=4304,
+        num_hidden_layers=27,
+        num_attention_heads=16,
+        num_channels=3,
+        num_patches=256,
+        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 = 384 #fixme
+        self.attention_dropout = attention_dropout
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+        self.num_patches = num_patches
+
+
+class Siglip2Config(PretrainedConfig):
+    r"""
+    [`Siglip2Config`] is the configuration class to store the configuration of a [`Siglip2Model`]. It is used to
+    instantiate a Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-224](https://huggingface.co/google/siglip2-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 [`Siglip2TextConfig`].
+        vision_config (`dict`, *optional*):
+            Dictionary of configuration options used to initialize [`Siglip2VisionConfig`].
+        kwargs (*optional*):
+            Dictionary of keyword arguments.
+
+    Example:
+
+    ```python
+    >>> from transformers import Siglip2Config, Siglip2Model
+
+    >>> # Initializing a Siglip2Config with google/siglip2-base-patch16-224 style configuration
+    >>> configuration = Siglip2Config()
+
+    >>> # Initializing a Siglip2Model (with random weights) from the google/siglip2-base-patch16-224 style configuration
+    >>> model = Siglip2Model(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+
+    >>> # We can also initialize a Siglip2Config from a Siglip2TextConfig and a Siglip2VisionConfig
+    >>> from transformers import Siglip2TextConfig, Siglip2VisionConfig
+
+    >>> # Initializing a Siglip2Text and Siglip2Vision configuration
+    >>> config_text = Siglip2TextConfig()
+    >>> config_vision = Siglip2VisionConfig()
+
+    >>> config = Siglip2Config.from_text_vision_configs(config_text, config_vision)
+    ```"""
+
+    model_type = "siglip2"
+    sub_configs = {"text_config": Siglip2TextConfig, "vision_config": Siglip2VisionConfig}
+
+    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 `Siglip2TextConfig` with default values.")
+
+        if vision_config is None:
+            vision_config = {}
+            logger.info("`vision_config` is `None`. initializing the `Siglip2VisionConfig` with default values.")
+
+        self.text_config = Siglip2TextConfig(**text_config)
+        self.vision_config = Siglip2VisionConfig(**vision_config)
+
+        self.initializer_factor = 1.0
+
+    @classmethod
+    def from_text_vision_configs(cls, text_config: Siglip2TextConfig, vision_config: Siglip2VisionConfig, **kwargs):
+        r"""
+        Instantiate a [`Siglip2Config`] (or a derived class) from siglip2 text model configuration and siglip2 vision
+        model configuration.
+
+        Returns:
+            [`Siglip2Config`]: An instance of a configuration object
+        """
+
+        return cls(text_config=text_config.to_dict(), vision_config=vision_config.to_dict(), **kwargs)
+    r"""
+    This is the configuration class to store the configuration of a [`Siglip2VisionModel`]. It is used to instantiate a
+    Siglip2 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 Siglip2
+    [google/siglip2-base-patch16-naflex](https://huggingface.co/google/siglip2-base-patch16-naflex) 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.
+        num_patches (`int`, *optional*, defaults to 256):
+            The number of patches in the image with the size of (`patch_size`, `patch_size`).
+            The image is resized to fill maximum of this number of patches, and to preserve
+            the aspect ratio. In case the resulted number of patches is lower, the image is
+            padded in "patch" dimension.
+        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 Siglip2VisionConfig, Siglip2VisionModel
+
+    >>> # Initializing a Siglip2VisionConfig with google/siglip2-base-patch16-naflex style configuration
+    >>> configuration = Siglip2VisionConfig()
+
+    >>> # Initializing a Siglip2VisionModel (with random weights) from the google/siglip2-base-patch16-naflex style configuration
+    >>> model = Siglip2VisionModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "siglip2_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,
+        num_patches=256,
+        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.attention_dropout = attention_dropout
+        self.layer_norm_eps = layer_norm_eps
+        self.hidden_act = hidden_act
+        self.num_patches = num_patches
+
+logger = logging.get_logger(__name__)
+
+# General docstring
+_CONFIG_FOR_DOC = "Siglip2VisionConfig"
+
+
+@dataclass
+class Siglip2VisionOutput(ModelOutput):
+    """
+    Base class for vision model's outputs that also contains image embeddings of the pooling of the last hidden states.
+
+    Args:
+        image_embeds (`torch.FloatTensor` of shape `(batch_size, output_dim)` *optional* returned when model is initialized with `with_projection=True`):
+            The image embeddings obtained by applying the projection layer to the pooler_output.
+        last_hidden_state (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`):
+            Sequence of hidden-states at the output of the last layer of the model.
+        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
+            Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
+            one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.
+
+            Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
+        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
+            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
+            sequence_length)`.
+
+            Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
+            heads.
+    """
+
+    image_embeds: Optional[torch.FloatTensor] = None
+    last_hidden_state: torch.FloatTensor = None
+    hidden_states: Optional[Tuple[torch.FloatTensor, ...]] = None
+    attentions: Optional[Tuple[torch.FloatTensor, ...]] = None
+
+
+class Siglip2VisionEmbeddings(nn.Module):
+    def __init__(self, config: Siglip2VisionConfig):
+        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.Linear(
+            in_features=config.num_channels * self.patch_size * self.patch_size,
+            out_features=self.embed_dim,
+        )
+
+        self.num_patches = config.num_patches
+        self.position_embedding_size = int(self.num_patches**0.5)
+        self.position_embedding = nn.Embedding(self.num_patches, self.embed_dim)
+    
+    @staticmethod
+    def resize_positional_embeddings(
+        positional_embeddings: torch.Tensor,
+        spatial_shapes: torch.LongTensor,
+        max_length: int,
+    ) -> torch.Tensor:
+        """
+        Resize positional embeddings to image-specific size and pad to a fixed size.
+
+        Args:
+            positional_embeddings (`torch.Tensor`):
+                Position embeddings of shape (height, width, embed_dim)
+            spatial_shapes (`torch.LongTensor`):
+                Spatial shapes of shape (batch_size, 2) to resize the positional embeddings to
+            max_length (`int`):
+                Maximum length of the positional embeddings to pad resized positional embeddings to
+
+        Returns:
+            `torch.Tensor`: Embeddings of shape (batch_size, max_length, embed_dim)
+        """
+        batch_size = spatial_shapes.shape[0]
+        embed_dim = positional_embeddings.shape[-1]
+        source_dtype = positional_embeddings.dtype
+
+        resulted_positional_embeddings = torch.empty(
+            (batch_size, max_length, embed_dim),
+            device=positional_embeddings.device,
+            dtype=source_dtype,
+        )
+
+        # (height, width, embed_dim) -> (1, embed_dim, height, width) for interpolation
+        positional_embeddings = positional_embeddings.permute(2, 0, 1).unsqueeze(0)
+
+        # Upcast to float32 on CPU because antialias is not supported for bfloat16/float16 on CPU
+        if positional_embeddings.device.type == "cpu":
+            positional_embeddings = positional_embeddings.to(torch.float32)
+
+        for i in range(batch_size):
+            # (1, dim, height, width) -> (1, dim, target_height, target_width)
+            height, width = spatial_shapes[i]
+            resized_embeddings = F.interpolate(
+                positional_embeddings,
+                size=(height, width),
+                mode="bilinear",
+                align_corners=False,
+                antialias=True,
+            )
+
+            # (1, dim, target_height, target_width) -> (target_height * target_width, dim)
+            resized_embeddings = resized_embeddings.reshape(embed_dim, height * width).transpose(0, 1)
+
+            # Cast to original dtype
+            resized_embeddings = resized_embeddings.to(source_dtype)
+
+            resulted_positional_embeddings[i, : height * width] = resized_embeddings
+            resulted_positional_embeddings[i, height * width :] = resized_embeddings[0]
+
+        return resulted_positional_embeddings
+
+    def forward(self, pixel_values: torch.FloatTensor, spatial_shapes: torch.LongTensor) -> torch.Tensor:
+        """
+        Args:
+            pixel_values (`torch.FloatTensor`):
+                Pixel values of shape (batch_size, max_num_patches, num_channels * patch_size * patch_size)
+            spatial_shapes (`List[Tuple[int, int]]`):
+                Spatial shapes of shape (batch_size, 2) to resize the positional embeddings to
+        """
+
+        # Apply patch embeddings to already patchified pixel values
+        #(batch, num_patches, channels, patch_size, patch_size)
+
+        pixel_values = rearrange(pixel_values, 'b c (h p1) (w p2) -> b (h w) (p1 p2 c)', p1=self.patch_size, p2=self.patch_size)
+        target_dtype = self.patch_embedding.weight.dtype
+        patch_embeds = self.patch_embedding(pixel_values.to(dtype=target_dtype))
+        
+        # Get positional resized and padded positional embeddings
+        positional_embeddings = self.position_embedding.weight.reshape(
+            self.position_embedding_size, self.position_embedding_size, -1
+        )
+        resized_positional_embeddings = self.resize_positional_embeddings(
+            positional_embeddings, spatial_shapes, max_length=pixel_values.shape[1]
+        )
+        # Add positional embeddings to patch embeddings
+        embeddings = patch_embeds + resized_positional_embeddings
+        return embeddings
+
+
+class Siglip2Attention(nn.Module):
+    """Multi-headed attention from 'Attention Is All You Need' paper"""
+
+    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 Siglip2SdpaAttention(Siglip2Attention):
+    """
+    Siglip2 attention module using torch.nn.functional.scaled_dot_product_attention. This module inherits from
+    `Siglip2Attention` 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 Siglip2Attention.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(
+                "Siglip2Model is using Siglip2SdpaAttention, 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
+
+
+class Siglip2MLP(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
+
+
+SIGLIP2_ATTENTION_CLASSES = {
+    "eager": Siglip2Attention,
+    "sdpa": Siglip2SdpaAttention,
+}
+
+
+
+class PatchMerging(nn.Module):
+    r""" Patch Merging Layer.
+
+    Args:
+        input_resolution (tuple[int]): Resolution of input feature.
+        dim (int): Number of input channels.
+        norm_layer (nn.Module, optional): Normalization layer.  Default: nn.LayerNorm
+    """
+
+    # def __init__(self, input_resolution, dim, norm_layer=nn.LayerNorm):
+    def __init__(self, dim, norm_layer=nn.LayerNorm):
+        super().__init__()
+        # self.input_resolution = input_resolution
+        self.dim = dim
+        self.norm = norm_layer(4 * dim)
+        self.reduction = nn.Sequential(
+            nn.Linear(4 * dim, 4 * dim, bias=False),
+            nn.GELU(),
+            nn.Linear(4 * dim, dim, bias=False),
+        )
+        # self.dim = dim
+        # self.reduction = nn.Linear(4 * dim, 2 * dim, bias=False)
+        # self.norm = norm_layer(4 * dim)
+        
+
+    def forward(self, x, spatial_shapes):
+        B, L, C = x.shape
+        H, W = spatial_shapes[0]
+        assert H * W == L
+        # TODO: 查看spatial_shapes是hw还是wh 确定了：是hw
+        x = x.reshape(B, H, W, C)
+        x = rearrange(x, 'b (h p1) (w p2) c -> b (h w) (p1 p2 c)', p1=2, p2=2) 
+        x = self.norm(x)
+        x = self.reduction(x)
+        return x, spatial_shapes // 2
+    
+class Siglip2EncoderLayer(nn.Module):
+    def __init__(self, config: Siglip2Config, layer_index):
+        super().__init__()
+        self.embed_dim = config.hidden_size
+        self.self_attn = SIGLIP2_ATTENTION_CLASSES[config._attn_implementation](config=config)
+        self.layer_norm1 = nn.LayerNorm(self.embed_dim, eps=config.layer_norm_eps)
+        self.mlp = Siglip2MLP(config)
+        self.layer_norm2 = nn.LayerNorm(self.embed_dim, eps=config.layer_norm_eps)
+        # add layer_index 来指示哪里存在merger层
+        self.layer_index = layer_index
+        if layer_index in config.vision_config['merger_layer_index']:
+            self.merger = PatchMerging(config.hidden_size)
+        else:
+            self.merger = None
+
+    # Ignore copy
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        spatial_shapes,
+        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
+
+        # 如果有merger就执行
+        if self.merger is not None:
+            hidden_states, spatial_shapes = self.merger(hidden_states, spatial_shapes)
+            outputs = (hidden_states, spatial_shapes)
+        else:
+            outputs = (hidden_states,)
+
+        if output_attentions:
+            outputs += (attn_weights,)
+
+        return outputs
+
+
+class Siglip2Encoder(nn.Module):
+    """
+    Transformer encoder consisting of `config.num_hidden_layers` self attention layers. Each layer is a
+    [`Siglip2EncoderLayer`].
+
+    Args:
+        config: Siglip2Config
+    """
+
+    def __init__(self, config: Siglip2Config):
+        super().__init__()
+        self.config = config
+        # self.layers = nn.ModuleList([Siglip2EncoderLayer(config) for _ in range(config.num_hidden_layers)])
+        self.layers = nn.ModuleList([Siglip2EncoderLayer(config, layer_index=i) for i in range(config.num_hidden_layers)])
+        self.gradient_checkpointing = False
+
+    # Ignore copy
+    def forward(
+        self,
+        inputs_embeds,
+        spatial_shapes,
+        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,
+                    spatial_shapes,
+                    attention_mask,
+                    output_attentions,
+                )
+            else:
+                layer_outputs = encoder_layer(
+                    hidden_states,
+                    spatial_shapes,
+                    attention_mask,
+                    output_attentions=output_attentions,
+                )
+
+            hidden_states = layer_outputs[0]
+
+            # 说明是swin的更新 spatial shapes
+            if len(layer_outputs) == 2 and not output_attentions:
+                spatial_shapes = layer_outputs[1]
+
+            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
+        )
+
+
+SIGLIP2_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.
+"""
+
+
+class Siglip2VisionTransformer(nn.Module):
+    def __init__(self, config: Siglip2VisionConfig):
+        super().__init__()
+        self.config = config
+        embed_dim = config.hidden_size
+
+        self.embeddings = Siglip2VisionEmbeddings(config)
+        self.encoder = Siglip2Encoder(config)
+        self.post_layernorm = nn.LayerNorm(embed_dim, eps=config.layer_norm_eps)
+        self.use_head = False if not hasattr(config, "vision_use_head") else config.vision_use_head
+        self._use_flash_attention_2 = config._attn_implementation == "flash_attention_2"
+
+    @add_start_docstrings_to_model_forward(SIGLIP2_VISION_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=Siglip2VisionConfig)
+    def forward(
+        self,
+        pixel_values,
+        spatial_shapes: torch.LongTensor,
+        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
+
+        hidden_states = self.embeddings(pixel_values, spatial_shapes)
+
+        encoder_outputs = self.encoder(
+            inputs_embeds=hidden_states,
+            spatial_shapes=spatial_shapes,
+            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)
+
+        pooled_output = self.head(last_hidden_state)
+
+        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,
+        )
+
+
+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")
+
+
+class Siglip2PreTrainedModel(PreTrainedModel):
+    """
+    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
+    models.
+    """
+
+    config_class = Siglip2Config
+    base_model_prefix = "siglip2"
+    supports_gradient_checkpointing = True
+
+    _no_split_modules = [
+        "Siglip2TextEmbeddings",
+        "Siglip2EncoderLayer",
+        "Siglip2VisionEmbeddings",
+        "Siglip2EncoderLayer",
+        "Siglip2MultiheadAttentionPoolingHead",
+    ]
+    _supports_flash_attn_2 = True
+    _supports_sdpa = True
+
+    def _init_weights(self, module):
+        """Initialize the weights"""
+        if isinstance(module, Siglip2VisionEmbeddings):
+            width = self.config.hidden_size
+            nn.init.normal_(module.position_embedding.weight, std=1 / np.sqrt(width))
+        elif isinstance(module, nn.Embedding):
+            default_flax_embed_init(module.weight)
+        elif isinstance(module, Siglip2Attention):
+            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, Siglip2MLP):
+            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, (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)
+
+
+class Siglip2VisionModel(Siglip2PreTrainedModel):
+    config_class = Siglip2VisionConfig
+    main_input_name = "pixel_values"
+
+    def __init__(self, config: Siglip2VisionConfig):
+        super().__init__(config)
+
+        self.vision_model = Siglip2VisionTransformer(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(SIGLIP2_VISION_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=BaseModelOutputWithPooling, config_class=Siglip2VisionConfig)
+    def forward(
+        self,
+        pixel_values: torch.FloatTensor,
+        spatial_shapes: torch.LongTensor,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        r"""
+        Returns:
+
+        Examples:
+
+        ```python
+        >>> from PIL import Image
+        >>> import requests
+        >>> from transformers import AutoProcessor, Siglip2VisionModel
+
+        >>> model = Siglip2VisionModel.from_pretrained("google/siglip2-base-patch16-224")
+        >>> processor = AutoProcessor.from_pretrained("google/siglip2-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,
+            spatial_shapes=spatial_shapes,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+
+
+class NaFlexSigLip2SwinVisionTower(nn.Module):
+    def __init__(self, vision_tower, vision_tower_cfg, delay_load=False):
+        super().__init__()
+
+        self.is_loaded = False
+
+        self.config = Siglip2VisionConfig()
+
+        self.vision_tower_name = vision_tower
+
+        self.image_processor = SigLipImageProcessor()
+
+        if not delay_load:
+            rank0_print(f"Loading vision tower: {vision_tower}")
+            self.load_model()
+        elif getattr(vision_tower_cfg, "unfreeze_mm_vision_tower", False):
+            # TODO: better detector is needed.
+            rank0_print(f"The checkpoint seems to contain `vision_tower` weights: `unfreeze_mm_vision_tower`: True.")
+            self.load_model()
+        elif hasattr(vision_tower_cfg, "mm_tunable_parts") and "mm_vision_tower" in vision_tower_cfg.mm_tunable_parts:
+            rank0_print(f"The checkpoint seems to contain `vision_tower` weights: `mm_tunable_parts` contains `mm_vision_tower`.")
+            self.load_model()
+        else:
+            self.cfg_only = self.config
+
+    def load_model(self, device_map=None):
+        if self.is_loaded:
+            rank0_print("{} is already loaded, `load_model` called again, skipping.".format(self.vision_tower_name))
+            return
+
+        self.vision_tower = Siglip2VisionModel.from_pretrained(self.vision_tower_name, device_map=device_map)
+        self.vision_tower.vision_model.head = nn.Identity()
+        self.vision_tower.requires_grad_(False)
+        self.is_loaded = True
+
+    def forward(self, images, patch_sizes):
+        if type(images) is list: 
+            image_features = []
+            for image, spatial_shape in zip(images, patch_sizes):
+                # 服务后面的维度：1, h, w
+                spatial_shape = torch.as_tensor(spatial_shape)[None]
+                image_forward_out = self.vision_tower(image.to(device=self.device, dtype=self.dtype).unsqueeze(0), spatial_shapes=spatial_shape, output_hidden_states=True)
+                # 根据崇屹经验，siglip2和siglip都是最后一层
+                # hidden_states的长度为 ViT层数+1，因为patching后的embedding也有
+                # image_feature = image_forward_out.hidden_states[-1].to(image.dtype)
+                image_feature = image_forward_out.last_hidden_state.to(image.dtype)
+                image_features.append(image_feature)
+        else: #torch.Size([1, 3, 384, 384])
+            print('no support for paralla')
+            exit()
+            image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype),spatial_shapes=patch_sizes, output_hidden_states=True)
+            # image_features = image_forward_outs.hidden_states[-1].to(images.dtype)
+            image_features = image_forward_outs.last_hidden_state.to(images.dtype)
+
+        return image_features
+    
+    # def forward_uhd_v2(self, images, patch_sizes):
+    #     if type(images) is list:
+    #         image_features = []
+    #         for image in images:
+    #             image_forward_out = self.vision_tower(image.to(device=self.device, dtype=self.dtype).unsqueeze(0), spatial_shapes=patch_sizes,output_hidden_states=True)
+    #             # 根据崇屹经验，siglip2和siglip都是最后一层
+    #             # hidden_states的长度为 ViT层数+1，因为patching后的embedding也有
+    #             image_feature = image_forward_out.last_hidden_state.to(image.dtype)
+    #             # image_feature = image_forward_out.hidden_states[-1].to(image.dtype)
+    #             image_features.append(image_feature)
+    #     else:
+    #         print('no support for paralla')
+    #         exit()
+    #         image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype), spatial_shapes=patch_sizes,output_hidden_states=True)
+    #         image_features = image_forward_outs.last_hidden_state.to(images.dtype)
+    #         # image_features = image_forward_outs.hidden_states[-1].to(images.dtype)
+
+    #     return image_features
+
+    @property
+    def dummy_feature(self):
+        return torch.zeros(1, self.hidden_size, device=self.device, dtype=self.dtype)
+
+    @property
+    def dtype(self):
+        for p in self.vision_tower.parameters():
+            return p.dtype
+
+    @property
+    def device(self):
+        for p in self.vision_tower.parameters():
+            return p.device
+
+    @property
+    def hidden_size(self):
+        return self.config.hidden_size
+
+    @property
+    def num_patches(self):
+        return (self.config.image_size // self.config.patch_size) ** 2
+
+    @property
+    def num_patches_per_side(self):
+        return self.config.image_size // self.config.patch_size
+        # return self.model_config["vision_cfg"]["image_size"] // self.model_config["vision_cfg"]["patch_size"]
+
+    @property
+    def image_size(self):
+        return self.config.image_size