Initial DeepSeek-OCR vLLM Space

- Full vLLM 0.8.5 integration with AsyncEngine
- Custom model registration and vision encoders
- All resolution modes (tiny/small/base/large/gundam)
- Comprehensive documentation and examples
- HF Jobs ready

.dockerignore

@@ -0,0 +1,49 @@
+# Git files
+.git
+.gitignore
+.gitattributes
+
+# Python cache
+__pycache__
+*.pyc
+*.pyo
+*.pyd
+.Python
+*.so
+*.egg
+*.egg-info
+dist
+build
+
+# Jupyter
+*.ipynb
+.ipynb_checkpoints
+
+# IDE
+.vscode
+.idea
+*.swp
+*.swo
+*~
+
+# OS
+.DS_Store
+Thumbs.db
+
+# Documentation
+*.md
+!README.md
+
+# Tests
+tests
+test_*
+*_test.py
+
+# Template file
+config_template.py
+
+# Local data
+data/
+outputs/
+temp/
+tmp/

Dockerfile

@@ -0,0 +1,51 @@
+FROM nvidia/cuda:11.8.0-cudnn8-devel-ubuntu22.04
+
+# Set environment variables
+ENV PYTHONUNBUFFERED=1 \
+    DEBIAN_FRONTEND=noninteractive \
+    CUDA_HOME=/usr/local/cuda-11.8 \
+    TRITON_PTXAS_PATH=/usr/local/cuda-11.8/bin/ptxas \
+    VLLM_USE_V1=0
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y \
+    python3.11 \
+    python3.11-dev \
+    python3-pip \
+    git \
+    wget \
+    && rm -rf /var/lib/apt/lists/*
+
+# Set python3.11 as default
+RUN update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.11 1 && \
+    update-alternatives --install /usr/bin/python python /usr/bin/python3.11 1
+
+# Upgrade pip
+RUN python3 -m pip install --no-cache-dir --upgrade pip
+
+# Install PyTorch 2.6.0 with CUDA 11.8
+RUN pip install --no-cache-dir \
+    torch==2.6.0 \
+    torchvision==0.21.0 \
+    torchaudio==2.6.0 \
+    --index-url https://download.pytorch.org/whl/cu118
+
+# Install vLLM 0.8.5 from GitHub releases
+RUN pip install --no-cache-dir \
+    https://github.com/vllm-project/vllm/releases/download/v0.8.5/vllm-0.8.5%2Bcu118-cp38-abi3-manylinux1_x86_64.whl
+
+# Install Flash Attention 2.7.3
+RUN pip install --no-cache-dir flash-attn==2.7.3 --no-build-isolation
+
+# Set working directory
+WORKDIR /app
+
+# Copy requirements and install Python dependencies
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Copy application code
+COPY . .
+
+# Default command (can be overridden by HF Jobs)
+CMD ["python", "main.py", "--help"]

README.md

@@ -1,10 +1,290 @@
 ---
-title: Deepseek Ocr
-emoji: 📈
-colorFrom: pink
-colorTo: red
+title: DeepSeek-OCR
+emoji: 📄
+colorFrom: blue
+colorTo: purple
 sdk: docker
 pinned: false
+tags:
+  - ocr
+  - vision-language-model
+  - document-processing
+  - vllm
+  - deepseek
+license: mit
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# DeepSeek-OCR with vLLM
+
+High-performance document OCR using [DeepSeek-OCR](https://huggingface.co/deepseek-ai/DeepSeek-OCR) with vLLM for efficient batch processing.
+
+## 🚀 Quick Start with HuggingFace Jobs
+
+Process any image dataset without needing your own GPU:
+
+```bash
+# Basic usage (Gundam mode - adaptive resolution)
+hf jobs run --flavor l4x1 \
+    --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    input-dataset \
+    output-dataset
+
+# Quick test with 10 samples
+hf jobs run --flavor l4x1 \
+    --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    your-input-dataset \
+    your-output-dataset \
+    --max-samples 10 \
+    --resolution-mode tiny
+```
+
+That's it! The script will:
+- ✅ Process images from your dataset
+- ✅ Add OCR results as a new `markdown` column
+- ✅ Push results to a new dataset with automatic documentation
+- 📊 View results at: `https://huggingface.co/datasets/[your-output-dataset]`
+
+## 📋 Features
+
+### Model Capabilities
+
+- 📐 **LaTeX equations** - Mathematical formulas preserved in LaTeX format
+- 📊 **Tables** - Extracted and formatted as HTML/markdown
+- 📝 **Document structure** - Headers, lists, and formatting maintained
+- 🖼️ **Image grounding** - Spatial layout and bounding box information
+- 🔍 **Complex layouts** - Multi-column and hierarchical structures
+- 🌍 **Multilingual** - Supports multiple languages
+
+### Performance
+
+- ⚡ **vLLM AsyncEngine** - Optimized for throughput (~2500 tokens/s on A100)
+- 🎯 **Multiple resolution modes** - Choose speed vs quality
+- 🔥 **Large context** - Up to 8K tokens
+- 💪 **Batch optimized** - Efficient async processing
+
+## 🎛️ Resolution Modes
+
+| Mode | Resolution | Vision Tokens | Best For |
+|------|-----------|---------------|----------|
+| `tiny` | 512×512 | 64 | Fast testing, simple documents |
+| `small` | 640×640 | 100 | Balanced speed/quality |
+| `base` | 1024×1024 | 256 | High quality documents |
+| `large` | 1280×1280 | 400 | Maximum quality, detailed docs |
+| `gundam` | Dynamic | Adaptive | Large documents, best overall |
+
+## 💻 Usage Examples
+
+### Basic Processing
+
+```bash
+# Default (Gundam mode)
+hf jobs run --flavor l4x1 --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    my-images-dataset \
+    ocr-results
+```
+
+### High Quality Mode
+
+```bash
+hf jobs run --flavor l40sx1 --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    documents-dataset \
+    extracted-text \
+    --resolution-mode large
+```
+
+### Fast Processing for Testing
+
+```bash
+hf jobs run --flavor l4x1 --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    large-dataset \
+    test-output \
+    --resolution-mode tiny \
+    --max-samples 100
+```
+
+### Random Sampling
+
+```bash
+hf jobs run --flavor l4x1 --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    ordered-dataset \
+    random-sample \
+    --max-samples 50 \
+    --shuffle \
+    --seed 42
+```
+
+### Custom Image Column
+
+```bash
+hf jobs run --flavor a10g-large --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    davanstrien/ufo-ColPali \
+    ufo-ocr \
+    --image-column image \
+    --resolution-mode gundam
+```
+
+### Private Output Dataset
+
+```bash
+hf jobs run --flavor l4x1 --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    private-input \
+    private-output \
+    --private
+```
+
+## 📝 Command-Line Options
+
+### Required Arguments
+
+| Argument | Description |
+|----------|-------------|
+| `input_dataset` | Input dataset ID from Hugging Face Hub |
+| `output_dataset` | Output dataset ID for Hugging Face Hub |
+
+### Optional Arguments
+
+| Option | Default | Description |
+|--------|---------|-------------|
+| `--image-column` | `image` | Column containing images |
+| `--model` | `deepseek-ai/DeepSeek-OCR` | Model to use |
+| `--resolution-mode` | `gundam` | Resolution preset (tiny/small/base/large/gundam) |
+| `--max-model-len` | `8192` | Maximum model context length |
+| `--max-tokens` | `8192` | Maximum tokens to generate |
+| `--gpu-memory-utilization` | `0.75` | GPU memory usage (0.0-1.0) |
+| `--prompt` | `<image>\n<\|grounding\|>Convert...` | Custom prompt |
+| `--hf-token` | - | Hugging Face API token (or use env var) |
+| `--split` | `train` | Dataset split to process |
+| `--max-samples` | None | Limit samples (for testing) |
+| `--private` | False | Make output dataset private |
+| `--shuffle` | False | Shuffle dataset before processing |
+| `--seed` | `42` | Random seed for shuffling |
+
+## 📊 Output Format
+
+The script adds two new columns to your dataset:
+
+1. **`markdown`** - The OCR text in markdown format
+2. **`inference_info`** - JSON metadata about the processing
+
+### Inference Info Structure
+
+```json
+[
+  {
+    "column_name": "markdown",
+    "model_id": "deepseek-ai/DeepSeek-OCR",
+    "processing_date": "2025-10-21T12:00:00",
+    "resolution_mode": "gundam",
+    "base_size": 1024,
+    "image_size": 640,
+    "crop_mode": true,
+    "prompt": "<image>\n<|grounding|>Convert the document to markdown.",
+    "max_tokens": 8192,
+    "gpu_memory_utilization": 0.75,
+    "max_model_len": 8192,
+    "script": "main.py",
+    "script_version": "1.0.0",
+    "space_url": "https://huggingface.co/spaces/davanstrien/deepseek-ocr",
+    "implementation": "vllm-async (optimized)"
+  }
+]
+```
+
+## 🔧 Technical Details
+
+### Architecture
+
+- **Model**: DeepSeek-OCR (3B parameters, based on Qwen2.5-VL)
+- **Inference Engine**: vLLM 0.8.5 with AsyncEngine
+- **Image Preprocessing**: Custom dynamic tiling based on aspect ratio
+- **Vision Encoders**: Custom CLIP + SAM encoders
+- **Context Length**: Up to 8K tokens
+- **Optimization**: Flash Attention 2.7.3, async batch processing
+
+### Hardware Requirements
+
+- **Minimum**: L4 GPU (24GB VRAM) - `--flavor l4x1`
+- **Recommended**: L40S/A10G (48GB VRAM) - `--flavor l40sx1` or `--flavor a10g-large`
+- **Maximum Performance**: A100 (40GB+ VRAM) - `--flavor a100-large`
+
+### Speed Benchmarks
+
+| GPU | Resolution | Speed | Notes |
+|-----|-----------|-------|-------|
+| L4 | Tiny | ~5-8 img/s | Good for testing |
+| L4 | Gundam | ~2-3 img/s | Balanced |
+| A100 | Gundam | ~8-12 img/s | Production speed |
+| A100 | Large | ~5-7 img/s | Maximum quality |
+
+## 📚 Example Workflows
+
+### 1. Process Historical Documents
+
+```bash
+hf jobs run --flavor l40sx1 --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    historical-scans \
+    historical-text \
+    --resolution-mode large \
+    --shuffle
+```
+
+### 2. Extract Tables from Reports
+
+```bash
+hf jobs run --flavor a10g-large --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    financial-reports \
+    extracted-tables \
+    --resolution-mode gundam \
+    --prompt "<image>\n<|grounding|>Convert the document to markdown."
+```
+
+### 3. Multi-language Documents
+
+```bash
+hf jobs run --flavor l4x1 --secrets HF_TOKEN \
+    hf.co/spaces/davanstrien/deepseek-ocr \
+    python main.py \
+    multilingual-docs \
+    ocr-output \
+    --resolution-mode base
+```
+
+## 🔗 Related Resources
+
+- **Model**: [deepseek-ai/DeepSeek-OCR](https://huggingface.co/deepseek-ai/DeepSeek-OCR)
+- **vLLM**: [vllm-project/vllm](https://github.com/vllm-project/vllm)
+- **HF Jobs**: [Documentation](https://huggingface.co/docs/huggingface_hub/en/guides/jobs)
+
+## 📄 License
+
+MIT License - See model card for details
+
+## 🙏 Acknowledgments
+
+- DeepSeek AI for the OCR model
+- vLLM team for the inference engine
+- Hugging Face for Jobs infrastructure
+
+---
+
+Built with ❤️ using [vLLM](https://github.com/vllm-project/vllm) and [DeepSeek-OCR](https://huggingface.co/deepseek-ai/DeepSeek-OCR)

config.py

@@ -0,0 +1,50 @@
+# Configuration for DeepSeek-OCR
+# These will be set programmatically by main.py based on command-line arguments
+
+# Resolution settings (set by resolution mode)
+BASE_SIZE = 1024
+IMAGE_SIZE = 640
+CROP_MODE = True
+
+# Processing settings
+MIN_CROPS = 2
+MAX_CROPS = 6  # max:9; If your GPU memory is small, it is recommended to set it to 6.
+MAX_CONCURRENCY = 100  # If you have limited GPU memory, lower the concurrency count.
+NUM_WORKERS = 64  # image pre-process (resize/padding) workers
+PRINT_NUM_VIS_TOKENS = False
+SKIP_REPEAT = True
+
+# Model settings
+MODEL_PATH = 'deepseek-ai/DeepSeek-OCR'
+
+# Paths (not used in Space version)
+INPUT_PATH = ''
+OUTPUT_PATH = ''
+
+# Default prompt
+PROMPT = '<image>\n<|grounding|>Convert the document to markdown.'
+
+# Tokenizer (will be initialized in main.py)
+TOKENIZER = None
+
+
+def set_resolution_mode(mode: str):
+    """Update global config based on resolution mode."""
+    global BASE_SIZE, IMAGE_SIZE, CROP_MODE
+
+    modes = {
+        "tiny": {"base_size": 512, "image_size": 512, "crop_mode": False},
+        "small": {"base_size": 640, "image_size": 640, "crop_mode": False},
+        "base": {"base_size": 1024, "image_size": 1024, "crop_mode": False},
+        "large": {"base_size": 1280, "image_size": 1280, "crop_mode": False},
+        "gundam": {"base_size": 1024, "image_size": 640, "crop_mode": True},
+    }
+
+    if mode not in modes:
+        raise ValueError(f"Unknown resolution mode: {mode}. Choose from {list(modes.keys())}")
+
+    BASE_SIZE = modes[mode]["base_size"]
+    IMAGE_SIZE = modes[mode]["image_size"]
+    CROP_MODE = modes[mode]["crop_mode"]
+
+    return BASE_SIZE, IMAGE_SIZE, CROP_MODE

config_template.py

@@ -0,0 +1,42 @@
+# TODO: change modes
+# Tiny: base_size = 512, image_size = 512, crop_mode = False
+# Small: base_size = 640, image_size = 640, crop_mode = False
+# Base: base_size = 1024, image_size = 1024, crop_mode = False
+# Large: base_size = 1280, image_size = 1280, crop_mode = False
+# Gundam: base_size = 1024, image_size = 640, crop_mode = True
+
+BASE_SIZE = 1024
+IMAGE_SIZE = 640
+CROP_MODE = True
+MIN_CROPS= 2
+MAX_CROPS= 6 # max:9; If your GPU memory is small, it is recommended to set it to 6.
+MAX_CONCURRENCY = 100 # If you have limited GPU memory, lower the concurrency count.
+NUM_WORKERS = 64 # image pre-process (resize/padding) workers 
+PRINT_NUM_VIS_TOKENS = False
+SKIP_REPEAT = True
+MODEL_PATH = 'deepseek-ai/DeepSeek-OCR' # change to your model path
+
+# TODO: change INPUT_PATH
+# .pdf: run_dpsk_ocr_pdf.py; 
+# .jpg, .png, .jpeg: run_dpsk_ocr_image.py; 
+# Omnidocbench images path: run_dpsk_ocr_eval_batch.py
+
+INPUT_PATH = '' 
+OUTPUT_PATH = ''
+
+PROMPT = '<image>\n<|grounding|>Convert the document to markdown.'
+# PROMPT = '<image>\nFree OCR.'
+# TODO commonly used prompts
+# document: <image>\n<|grounding|>Convert the document to markdown.
+# other image: <image>\n<|grounding|>OCR this image.
+# without layouts: <image>\nFree OCR.
+# figures in document: <image>\nParse the figure.
+# general: <image>\nDescribe this image in detail.
+# rec: <image>\nLocate <|ref|>xxxx<|/ref|> in the image.
+# '先天下之忧而忧'
+# .......
+
+
+from transformers import AutoTokenizer
+
+TOKENIZER = AutoTokenizer.from_pretrained(MODEL_PATH, trust_remote_code=True)

deepencoder/build_linear.py

@@ -0,0 +1,174 @@
+import torch.nn as nn
+import torch
+import torch.nn.functional as F
+import copy
+
+
+class MlpProjector(nn.Module):
+
+    def __init__(self, cfg):
+
+        super().__init__()
+
+        self.cfg = cfg
+
+        if cfg.projector_type == "identity":
+            modules = nn.Identity()
+
+        elif cfg.projector_type == "linear":
+            modules = nn.Linear(cfg.input_dim, cfg.n_embed)
+
+        elif cfg.projector_type == "mlp_gelu":
+            mlp_depth = cfg.get("depth", 1)
+            modules = [nn.Linear(cfg.input_dim, cfg.n_embed)]
+            for _ in range(1, mlp_depth):
+                modules.append(nn.GELU())
+                modules.append(nn.Linear(cfg.n_embed, cfg.n_embed))
+            modules = nn.Sequential(*modules)
+        
+        elif cfg.projector_type == "normlayer_downsample_mlp_gelu":
+            mlp_depth = cfg.get("depth", 1)
+            mlp_ratio = cfg.get("mlp_ratio", 1)
+            modules = [
+                nn.LayerNorm(cfg.input_dim * cfg.downsample_ratio * cfg.downsample_ratio),
+                nn.Linear(cfg.input_dim * cfg.downsample_ratio * cfg.downsample_ratio, cfg.n_embed * mlp_ratio)
+            ]
+            for _ in range(1, mlp_depth - 1):
+                modules.append(nn.GELU())
+                modules.append(nn.Linear(cfg.n_embed * mlp_ratio, cfg.n_embed * mlp_ratio))
+            modules.append(nn.GELU())
+            modules.append(nn.Linear(cfg.n_embed * mlp_ratio, cfg.n_embed))
+            modules = nn.Sequential(*modules)
+        
+        elif cfg.projector_type == "downsample_mlp_gelu":
+            mlp_depth = cfg.get("depth", 1)
+            mlp_ratio = cfg.get("mlp_ratio", 1)
+            modules = [nn.Linear(cfg.input_dim * cfg.downsample_ratio * cfg.downsample_ratio, cfg.n_embed * mlp_ratio)]
+            for _ in range(1, mlp_depth - 1):
+                modules.append(nn.GELU())
+                modules.append(nn.Linear(cfg.n_embed * mlp_ratio, cfg.n_embed * mlp_ratio))
+            modules.append(nn.GELU())
+            modules.append(nn.Linear(cfg.n_embed * mlp_ratio, cfg.n_embed))
+            modules = nn.Sequential(*modules)
+
+        elif cfg.projector_type == "low_high_hybrid_split_mlp_gelu":
+            mlp_depth = cfg.get("depth", 1)
+            self.high_up_proj = nn.Linear(cfg.input_dim, cfg.n_embed // 2)
+            self.low_up_proj = nn.Linear(cfg.input_dim, cfg.n_embed // 2)
+
+            modules = []
+            for _ in range(1, mlp_depth):
+                modules.append(nn.GELU())
+                modules.append(nn.Linear(cfg.n_embed, cfg.n_embed))
+            modules = nn.Sequential(*modules)
+
+        elif cfg.projector_type == "hybrid_split_feature_mlp_gelu":
+            mlp_depth = cfg.get("depth", 1)
+            channel_div = cfg.get("channel_div", 0.5)
+            self.high_up_proj = nn.Linear(cfg.input_dim[0], int(cfg.n_embed * channel_div))
+            self.low_up_proj = nn.Linear(cfg.input_dim[1], cfg.n_embed - int(cfg.n_embed * channel_div))
+
+            modules = []
+            for _ in range(1, mlp_depth):
+                modules.append(nn.GELU())
+                modules.append(nn.Linear(cfg.n_embed, cfg.n_embed))
+            modules = nn.Sequential(*modules)
+
+        elif cfg.projector_type == "low_high_split_mlp_gelu":
+            mlp_depth = cfg.get("depth", 1)
+            modules = []
+            for _ in range(1, mlp_depth):
+                modules.append(nn.GELU())
+                modules.append(nn.Linear(cfg.n_embed // 2, cfg.n_embed // 2))
+            modules = nn.Sequential(*modules)
+            self.high_layers = nn.Sequential(*modules)
+            self.low_layers = copy.deepcopy(modules)
+
+        else:
+            raise ValueError(f"Unknown projector type: {cfg.projector_type}")
+
+        if cfg.get("token_pooling", False):
+            self.token_pooling_layer = nn.Linear(cfg.input_dim * 4, cfg.input_dim)
+
+        if cfg.get("conv_fusion_high_low_features", False):
+            self.fusion_layer = nn.Linear(cfg.input_dim, cfg.input_dim)
+        self.layers = modules
+
+    def forward(self, x):
+        if self.cfg.get("token_pooling", False):
+            batch_size, wxh, channels = x.shape
+            w = h = int(wxh**0.5)
+            x = x.view(batch_size, w, h, channels)
+            x = x.permute(0, 3, 1, 2)
+            # import ipdb; ipdb.set_trace()
+            patches = x.unfold(2, 2, 2).unfold(3, 2, 2)
+            batch_size, channels, h_patches, w_patches, _, _ = patches.size()
+            # 在通道维度上拼接
+            patches = patches.contiguous().view(batch_size, channels, h_patches * w_patches, -1)
+
+            # 通过线性层
+            patches = patches.permute(0, 2, 1, 3).contiguous()
+            patches = patches.view(batch_size, h_patches * w_patches, channels * 4)
+
+            x = self.token_pooling_layer(patches)
+        
+        if self.cfg.get("conv_fusion_high_low_features", False):
+            x = self.fusion_layer(x[:, 0]) + x[:, 1]
+
+        if self.cfg.projector_type == 'low_high_hybrid_split_mlp_gelu':
+            high_x, low_x = x[0], x[1]
+            high_x = self.high_up_proj(high_x)
+            low_x = self.low_up_proj(low_x)
+            x = torch.concat([high_x, low_x], dim=-1)
+        
+        if self.cfg.projector_type == 'hybrid_split_feature_mlp_gelu':
+            high_x = x[...,:self.cfg.input_dim[0]]
+            low_x = x[...,self.cfg.input_dim[0]:]
+            high_x = self.high_up_proj(high_x)
+            low_x = self.low_up_proj(low_x)
+            x = torch.concat([high_x, low_x], dim=-1)
+        
+        if self.cfg.projector_type == 'low_high_split_mlp_gelu':
+            high_x, low_x = x[0], x[1]
+            high_x = self.high_layers(high_x)
+            low_x = self.low_layers(low_x)
+            x = torch.concat([high_x, low_x], dim=-1)
+            return x
+        
+        if self.cfg.projector_type == 'downsample_mlp_gelu' or self.cfg.projector_type == 'normlayer_downsample_mlp_gelu':
+            bs, hw, input_dim = x.shape
+            h = w = int((hw) ** 0.5)
+
+            """compute padding"""
+            if h % self.cfg.downsample_ratio:
+                pad = self.cfg.downsample_ratio - h % self.cfg.downsample_ratio
+            else:
+                pad = 0
+            x = x.reshape(bs, h, w, input_dim)
+            if pad > 0:
+                x = F.pad(x, (0, 0, 0, pad, 0, pad), "constant", 0)
+
+            """4 to 1 concat"""
+            x = x.permute(0, 3, 1, 2)  # B, C, H, W
+            x = F.unfold(x, kernel_size=self.cfg.downsample_ratio, stride=self.cfg.downsample_ratio, padding=0) # B, C*4, HW // 4
+            x = x.permute(0, 2, 1)
+            
+        return self.layers(x)
+
+    @staticmethod
+    def get_flops_per_sample(cfg):
+        if cfg.projector_type == "linear":
+            fwd = 2 * cfg.input_dim * cfg.n_embed
+
+        elif "mlp_gelu" in cfg.projector_type :
+            mlp_depth = cfg.get("depth", 1)
+            downsample_ratio = cfg.get("downsample_ratio", 1)
+            input_dim = sum(cfg.input_dim) if isinstance(cfg.input_dim, list) else cfg.input_dim
+            input_dim = input_dim * downsample_ratio * downsample_ratio
+            fwd = 2 * input_dim * cfg.n_embed + (mlp_depth - 1) * 2 * cfg.n_embed * cfg.n_embed
+        else:
+            fwd = 0
+
+        return fwd * 3
+
+

deepencoder/clip_sdpa.py

@@ -0,0 +1,504 @@
+from contextlib import nullcontext
+import math
+from typing import Optional, Tuple
+# from megatron.model import LayerNorm
+from easydict import EasyDict as adict
+import torch
+from torch.nn import functional as F
+from torch import nn
+from flash_attn import flash_attn_qkvpacked_func, flash_attn_func
+# from optimus import flash_attn_func
+# from megatron.core import tensor_parallel
+# from megatron.core import parallel_state as mpu
+# from megatron.core.utils import make_viewless_tensor, divide
+# from megatron.model.fused_rms_norm import RMSNorm
+# from megatron.model.transformer import (
+#     FlashSelfAttention,
+#     NoopTransformerLayer,
+#     _cfg_to_kwargs,
+# )
+# from megatron.model.enums import AttnMaskType, AttnType
+# from megatron.model.fused_softmax import FusedScaleMaskSoftmax
+# from megatron.model.utils import attention_mask_func
+
+# from megatron.model.module import MegatronModule
+
+# try:
+#     from einops import rearrange
+# except ImportError:
+#     rearrange = None
+
+# from flash_attn import flash_attn_varlen_func as flash_attn_unpadded_func
+
+# try:
+#     # flash attention 2.x
+#     from flash_attn import flash_attn_varlen_func as flash_attn_unpadded_func
+# except ImportError:
+#     try:
+#         # flash attention 1.x
+#         from flash_attn.flash_attn_interface import flash_attn_unpadded_func
+#     except ImportError:
+#         flash_attn_unpadded_func = None
+
+# try:
+#     from flash_attn.flash_attn_interface import flash_attn_unpadded_relative_attention_bias_func
+# except ImportError:
+#     flash_attn_unpadded_relative_attention_bias_func = None
+
+# try:
+#     from flash_attn.flash_attn_interface import mask_flash_attn_unpadded_func
+# except ImportError:
+#     mask_flash_attn_unpadded_func = None
+
+
+class LayerNormfp32(torch.nn.LayerNorm):
+    """Subclass torch's LayerNorm to handle fp16."""
+
+    def forward(self, x: torch.Tensor):
+        orig_type = x.dtype
+        ret = super().forward(x.type(torch.float32))
+        return ret.type(orig_type)
+
+
+def get_abs_pos(abs_pos, tgt_size):
+    # abs_pos: L, C
+    # tgt_size: M
+    # return: M, C
+
+    # print(tgt_size)
+    # print(abs_pos.shape)
+    # exit()
+    dim = abs_pos.size(-1)
+    # print(dim)
+    abs_pos_new = abs_pos.squeeze(0)
+    cls_token, old_pos_embed = abs_pos_new[:1], abs_pos_new[1:]
+
+
+
+    src_size = int(math.sqrt(abs_pos_new.shape[0] - 1))
+    tgt_size = int(math.sqrt(tgt_size))
+    dtype = abs_pos.dtype
+
+    if src_size != tgt_size:
+        old_pos_embed = old_pos_embed.view(1, src_size, src_size, dim).permute(0, 3, 1,
+                                                                                    2).contiguous()
+        old_pos_embed = old_pos_embed.to(torch.float32)
+        new_pos_embed = F.interpolate(
+            old_pos_embed,
+            size=(tgt_size, tgt_size),
+            mode='bicubic',
+            antialias=True,
+            align_corners=False,
+        ).to(dtype)
+        new_pos_embed = new_pos_embed.permute(0, 2, 3, 1)
+        new_pos_embed = new_pos_embed.view(tgt_size * tgt_size, dim)
+        vision_pos_embed = torch.cat([cls_token, new_pos_embed], dim=0)
+        vision_pos_embed = vision_pos_embed.view(1, tgt_size * tgt_size + 1, dim)
+        return vision_pos_embed
+    else:
+        return abs_pos
+
+@torch.jit.script
+def quick_gelu(x):
+    return x * torch.sigmoid(1.702 * x)
+
+
+
+class CLIPVisionEmbeddings(nn.Module):
+    def __init__(self, hidden_size=1024, image_size=224, patch_size=14, num_channels=3):
+        super().__init__()
+        self.embed_dim = hidden_size
+        self.image_size = image_size
+        self.patch_size = patch_size
+
+        self.class_embedding = torch.nn.Parameter(torch.randn(self.embed_dim))
+
+        self.patch_embedding = torch.nn.Conv2d(
+            in_channels=num_channels,
+            out_channels=self.embed_dim,
+            kernel_size=self.patch_size,
+            stride=self.patch_size,
+            bias=False,
+        )
+
+        self.num_patches = (self.image_size // self.patch_size) ** 2
+        self.num_positions = self.num_patches + 1
+        self.position_embedding = torch.nn.Embedding(self.num_positions, self.embed_dim)
+        self.register_buffer(
+            "position_ids", torch.arange(self.num_positions).expand((1, -1))
+        )
+
+    def forward(self, pixel_values, patch_embeds):
+        batch_size = pixel_values.shape[0]
+        # patch_embeds = self.patch_embedding(
+        #     pixel_values
+        # )  # shape = [*, width, grid, grid]
+
+
+        if patch_embeds is not None:
+            patch_embeds = patch_embeds
+            # print(patch_embeds.shape)
+        else:
+            patch_embeds = self.patch_embedding(pixel_values)  
+            # print(111111)
+        # shape = [*, width, grid, grid]
+        # patch_embeds = patch_embeds.flatten(2).transpose(1, 2)
+
+        patch_embeds = patch_embeds.flatten(2).transpose(1, 2)
+
+
+        class_embeds = self.class_embedding.expand(batch_size, 1, -1)
+        embeddings = torch.cat([class_embeds, patch_embeds], dim=1)
+
+        # x = torch.cat([cls_token, x], dim=1)
+        embeddings = embeddings + get_abs_pos(self.position_embedding(self.position_ids), embeddings.size(1))
+        # embeddings = embeddings + self.position_embedding(self.position_ids)
+        return embeddings
+
+
+class NoTPFeedForward(nn.Module):
+    def __init__(
+            self,
+            cfg,
+            dim: int,
+            hidden_dim: int,
+    ):
+        super().__init__()
+
+        self.fc1 = torch.nn.Linear(dim, hidden_dim, bias=True)
+        self.fc2 = torch.nn.Linear(hidden_dim, dim, bias=True)
+
+    def forward(self, x):
+        output = self.fc2(quick_gelu(self.fc1(x)))
+        return output
+
+
+# from optimus.flash_attn_interface import flash_attn_qkvpacked_func
+
+
+# class NoTPAttention(nn.Module):
+#     def __init__(self, cfg):
+#         super().__init__()
+#         self.num_heads = cfg.num_attention_heads
+#         self.n_local_heads = cfg.num_attention_heads
+#         self.head_dim = cfg.hidden_size // cfg.num_attention_heads
+#         self.max_seq_len = cfg.seq_length
+#         self.use_flash_attention = cfg.use_flash_attn
+
+#         self.qkv_proj = torch.nn.Linear(cfg.hidden_size, cfg.hidden_size * 3, bias=True)
+#         self.out_proj = torch.nn.Linear(cfg.hidden_size, cfg.hidden_size, bias=True)
+
+#         # self.core_attention = CoreAttention(cfg, AttnType.self_attn)
+
+#         self.attn_drop = cfg.attention_dropout
+
+#     def forward(
+#             self,
+#             x: torch.Tensor,
+#     ):
+#         bsz, seqlen, _ = x.shape
+#         xqkv = self.qkv_proj(x)
+#         xqkv = xqkv.view(bsz, seqlen, 3, self.num_heads, self.head_dim)
+
+#         if self.use_flash_attention:
+#             output = flash_attn_qkvpacked_func(xqkv)
+#             output = output.view(bsz, seqlen, -1)
+#         else:
+#             xq, xk, xv = torch.split(xqkv, 1, dim=2)
+#             xq = xq.squeeze(2)
+#             xk = xk.squeeze(2)
+#             xv = xv.squeeze(2)
+#             # xq, xk, xv = xqkv[:, :, 0, ...], xqkv[:, :, 1, ...], xqkv[:, :, 2, ...]
+
+#             # （B, num_head, S, head_size)
+#             xq = xq.permute(0, 2, 1, 3)
+#             xk = xk.permute(0, 2, 1, 3)
+#             xv = xv.permute(0, 2, 1, 3)
+
+#             output = torch.nn.functional.scaled_dot_product_attention(xq, xk, xv, attn_mask=None)
+#             utput = output.permute(0, 2, 1, 3).view(bsz, seqlen, -1)
+#         output = self.out_proj(output)
+#         return output
+
+
+# from optimus.flash_attn_interface import flash_attn_qkvpacked_func
+
+
+class NoTPAttention(torch.nn.Module):
+    def __init__(self, cfg):
+        super().__init__()
+        self.num_heads = cfg.num_attention_heads
+        self.n_local_heads = cfg.num_attention_heads
+        self.head_dim = cfg.hidden_size // cfg.num_attention_heads
+        self.max_seq_len = cfg.seq_length
+        self.use_flash_attention = cfg.use_flash_attn
+
+        self.qkv_proj = torch.nn.Linear(cfg.hidden_size, cfg.hidden_size * 3, bias=True)
+        self.out_proj = torch.nn.Linear(cfg.hidden_size, cfg.hidden_size, bias=True)
+
+        # self.core_attention = CoreAttention(cfg, AttnType.self_attn)
+
+        self.attn_drop = cfg.attention_dropout
+
+    def forward(
+            self,
+            x: torch.Tensor,
+    ):
+        bsz, seqlen, _ = x.shape
+        xqkv = self.qkv_proj(x)
+        xqkv = xqkv.view(bsz, seqlen, 3, self.num_heads, self.head_dim)
+
+        if self.use_flash_attention:
+            output = flash_attn_qkvpacked_func(xqkv)
+            output = output.view(bsz, seqlen, -1)
+            # xq, xk, xv = torch.split(xqkv, 1, dim=2)
+            # xq = xq.squeeze(2)
+            # xk = xk.squeeze(2)
+            # xv = xv.squeeze(2)
+            # # xq, xk, xv = xqkv[:, :, 0, ...], xqkv[:, :, 1, ...], xqkv[:, :, 2, ...]
+
+            # # （B, num_head, S, head_size)
+            # xq = xq.permute(0, 2, 1, 3)
+            # xk = xk.permute(0, 2, 1, 3)
+            # xv = xv.permute(0, 2, 1, 3)
+            # # with torch.backends.cuda.sdp_kernel(enable_flash=True, enable_math=False, enable_mem_efficient=False):
+            # output = torch.nn.functional.scaled_dot_product_attention(xq, xk, xv, attn_mask=None)
+            # output = output.permute(0, 2, 1, 3).reshape(bsz, seqlen, -1)
+                # output = output.permute(0, 2, 1, 3).contiguous().view(bsz, seqlen, -1)
+        else:
+            # output = flash_attn_qkvpacked_func(xqkv)
+            xq, xk, xv = torch.split(xqkv, 1, dim=2)
+            xq = xq.squeeze(2)
+            xk = xk.squeeze(2)
+            xv = xv.squeeze(2)
+            # xq, xk, xv = xqkv[:, :, 0, ...], xqkv[:, :, 1, ...], xqkv[:, :, 2, ...]
+
+            # （B, num_head, S, head_size)
+            xq = xq.permute(0, 2, 1, 3)
+            xk = xk.permute(0, 2, 1, 3)
+            xv = xv.permute(0, 2, 1, 3)
+            # with torch.backends.cuda.sdp_kernel(enable_flash=True, enable_math=False, enable_mem_efficient=False):
+            output = torch.nn.functional.scaled_dot_product_attention(xq, xk, xv, attn_mask=None)
+            output = output.permute(0, 2, 1, 3).reshape(bsz, seqlen, -1)
+        output = self.out_proj(output)
+        return output
+
+class NoTPTransformerBlock(nn.Module):
+    def __init__(self, cfg, layer_id: int, multiple_of=256):
+        super().__init__()
+
+        self.n_heads = cfg.num_attention_heads
+        self.dim = cfg.hidden_size
+        self.head_dim = cfg.hidden_size // cfg.num_attention_heads
+        self.self_attn = NoTPAttention(cfg)
+        self.mlp = NoTPFeedForward(
+            cfg, dim=cfg.hidden_size, hidden_dim=cfg.ffn_hidden_size
+        )
+        self.layer_id = layer_id
+        self.layer_norm1 = torch.nn.LayerNorm(
+            cfg.hidden_size, eps=cfg.layernorm_epsilon
+        )
+        self.layer_norm2 = torch.nn.LayerNorm(
+            cfg.hidden_size, eps=cfg.layernorm_epsilon
+        )
+
+    def forward(self, x: torch.Tensor):
+        residual = self.self_attn.forward(self.layer_norm1(x))
+        h = x + residual
+        out = h + self.mlp.forward(self.layer_norm2(h))
+        return out
+
+
+class NoTPTransformer(nn.Module):
+    def __init__(self, cfg):
+        super().__init__()
+
+        self.cfg = cfg
+        # self.recompute_list = self.cfg.get("recompute_list", [])
+        self.num_layers = cfg.num_layers  # _get_num_layers(cfg)
+
+        self.layers = torch.nn.ModuleList()
+        for layer_id in range(self.num_layers):
+            self.layers.append(
+                NoTPTransformerBlock(
+                    cfg,
+                    layer_id + 1,
+                )
+            )
+
+    def forward(
+            self,
+            hidden_states,
+    ):
+
+        for lid, layer in enumerate(self.layers):
+            # if lid in self.recompute_list:
+            #     def custom(layer_id):
+            #         def custom_forward(*args, **kwargs):
+            #             x_ = self.layers[layer_id](*args, **kwargs)
+            #             return x_
+
+            #         return custom_forward
+
+            #     assert hidden_states.requires_grad == True, logger.warning(
+            #         "When using recalculation, the input must have grad fn"
+            #     )
+            #     hidden_states = tensor_parallel.checkpoint(
+            #         custom(lid),
+            #         False,
+            #         hidden_states.contiguous()
+            #     )
+            # else:
+            hidden_states = layer(hidden_states)
+
+        return hidden_states
+
+
+# from megatron.core.tensor_parallel.layers import non_tensor_paralleled, local_dp_reduce, local_dp_scatter
+
+class VitModel(nn.Module):
+    def __init__(
+            self,
+            cfg,
+            freeze_embed=False,
+            freeze_pre_norm=False
+    ) -> None:
+        super().__init__()
+
+        self.embeddings = CLIPVisionEmbeddings(hidden_size=cfg.hidden_size, image_size=cfg.image_size, patch_size=cfg.patch_size)
+
+        if freeze_embed:
+            for name, param in self.embeddings.named_parameters():
+                param.requires_grad = False
+
+        self.transformer = NoTPTransformer(cfg=cfg)
+
+        if cfg.get("fp32norm", False):
+            logger.info("Load fp32 layernorm for ViT.")
+            self.pre_layrnorm = LayerNormfp32(
+                cfg.hidden_size,
+                eps=cfg.get("pre_layernorm_epsilon", 1e-5),
+            )
+        else:
+            self.pre_layrnorm = torch.nn.LayerNorm(
+                cfg.hidden_size,
+                eps=cfg.get("pre_layernorm_epsilon", 1e-5),
+            )
+
+        # self.pre_layrnorm = RMSNorm(
+        #     cfg.hidden_size,
+        #     eps=cfg.get("pre_layernorm_epsilon", 1e-5),
+        #     sequence_parallel=False,
+        #     use_fp32=True,
+        #     use_optimus=True,
+        # )
+
+        if freeze_pre_norm:
+            for name, param in self.pre_layrnorm.named_parameters():
+                param.requires_grad = False
+
+        for p in self.parameters():
+            p.micro_dp = True
+
+    def set_input_tensor(self, input_tensor):
+        if not isinstance(input_tensor, list):
+            input_tensor = [input_tensor]
+        self.transformer.set_input_tensor(input_tensor[0])
+
+    def __str__(self) -> str:
+        return "open_clip"
+
+    def forward(
+            self,
+            x,
+            patch_embeds
+    ):
+        x = self.embeddings(x, patch_embeds)
+        hidden_states = self.pre_layrnorm(x)
+
+        # hidden_states, dis = local_dp_scatter(hidden_states)
+        output = self.transformer(hidden_states)
+
+        # output = local_dp_reduce(output, dis)
+
+        return output
+
+
+vit_model_cfg = adict(
+    num_layers=24,
+    hidden_size=1024,
+    num_heads = 16,
+    num_attention_heads=16,
+    ffn_hidden_size=4096,
+    seq_length=256,
+    max_position_embeddings=256,
+    use_flash_attn=False,
+    understand_projector_stride=2,
+    hidden_dropout = 0.0,
+    attention_dropout = 0.0,
+    no_persist_layer_norm = False,
+    layernorm_epsilon = 1e-5,
+    pre_layernorm_epsilon = 1e-5,
+    image_size = 224,
+    patch_size = 14,
+    recompute_list = []
+)
+
+def build_clip_l():
+    return VitModel(
+        cfg=vit_model_cfg,
+        freeze_embed=False,
+        freeze_pre_norm=False,
+    )
+
+
+if __name__ == '__main__':
+
+    
+    from mmgpt.model.vision_encoder.sam_b import build_sam_vit_b
+
+
+
+    vit_model_cfg = adict(
+        num_layers=24,
+        hidden_size=1024,
+        num_attention_heads=16,
+        ffn_hidden_size=4096,
+        seq_length=256,
+        max_position_embeddings=256,
+        use_flash_attn=False,
+        understand_projector_stride=2,
+        hidden_dropout = 0.0,
+        attention_dropout = 0.0,
+        no_persist_layer_norm = False,
+        layernorm_epsilon = 1e-5,
+        pre_layernorm_epsilon = 1e-5,
+        image_size = 224,
+        patch_size = 14,
+        recompute_list = []
+    )
+
+    sam_model = build_sam_vit_b()
+
+
+    vision_model = VitModel(
+        cfg=vit_model_cfg,
+        freeze_embed=False,
+        freeze_pre_norm=False,
+    )
+
+    # model = VitModel(1344)
+    # x = torch.zeros(2, 3, 224, 224)
+    x = torch.zeros(2, 3, 1024, 1024)
+
+    
+    with torch.no_grad():
+        # y = vision_model(x)
+        patch_embed = sam_model(x)
+        print(patch_embed.shape)
+        y = vision_model(x, patch_embed)
+        print(y.shape)
+
+        image_feature = torch.add(y[:, 1:], patch_embed.flatten(2).permute(0, 2, 1))
+
+        print(image_feature.shape)

deepencoder/sam_vary_sdpa.py

@@ -0,0 +1,528 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+from typing import Optional, Tuple, Type
+from functools import partial
+from flash_attn import flash_attn_qkvpacked_func
+# from .common import LayerNorm2d, MLPBlock
+
+# from mmgpt.model.vision_encoder.flash_4 import _attention_rel_h_rel_w
+
+
+def get_abs_pos(abs_pos, tgt_size):
+
+    dtype = abs_pos.dtype
+
+    src_size = abs_pos.size(1)
+
+    if src_size != tgt_size:
+        old_pos_embed = abs_pos.permute(0, 3, 1, 2)
+        old_pos_embed = old_pos_embed.to(torch.float32)
+        new_pos_embed = F.interpolate(
+            old_pos_embed,
+            size=(tgt_size, tgt_size),
+            mode='bicubic',
+            antialias=True,
+            align_corners=False,
+        ).to(dtype)
+        new_pos_embed = new_pos_embed.permute(0, 2, 3, 1)
+        return new_pos_embed
+    else:
+        return abs_pos
+
+
+
+
+class MLPBlock(nn.Module):
+    def __init__(
+        self,
+        embedding_dim: int,
+        mlp_dim: int,
+        act: Type[nn.Module] = nn.GELU,
+    ) -> None:
+        super().__init__()
+        self.lin1 = nn.Linear(embedding_dim, mlp_dim)
+        self.lin2 = nn.Linear(mlp_dim, embedding_dim)
+        self.act = act()
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        return self.lin2(self.act(self.lin1(x)))
+
+
+# From https://github.com/facebookresearch/detectron2/blob/main/detectron2/layers/batch_norm.py # noqa
+# Itself from https://github.com/facebookresearch/ConvNeXt/blob/d1fa8f6fef0a165b27399986cc2bdacc92777e40/models/convnext.py#L119  # noqa
+class LayerNorm2d(nn.Module):
+    def __init__(self, num_channels: int, eps: float = 1e-6) -> None:
+        super().__init__()
+        self.weight = nn.Parameter(torch.ones(num_channels))
+        self.bias = nn.Parameter(torch.zeros(num_channels))
+        self.eps = eps
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        u = x.mean(1, keepdim=True)
+        s = (x - u).pow(2).mean(1, keepdim=True)
+        x = (x - u) / torch.sqrt(s + self.eps)
+        x = self.weight[:, None, None] * x + self.bias[:, None, None]
+        return x
+
+
+# This class and its supporting functions below lightly adapted from the ViTDet backbone available at: https://github.com/facebookresearch/detectron2/blob/main/detectron2/modeling/backbone/vit.py # noqa
+class ImageEncoderViT(nn.Module):
+    def __init__(
+        self,
+        img_size: int = 1024,
+        patch_size: int = 16,
+        in_chans: int = 3,
+        embed_dim: int = 768,
+        depth: int = 12,
+        num_heads: int = 12,
+        mlp_ratio: float = 4.0,
+        out_chans: int = 256,
+        qkv_bias: bool = True,
+        norm_layer: Type[nn.Module] = nn.LayerNorm,
+        act_layer: Type[nn.Module] = nn.GELU,
+        use_abs_pos: bool = True,
+        use_rel_pos: bool = False,
+        rel_pos_zero_init: bool = True,
+        window_size: int = 0,
+        global_attn_indexes: Tuple[int, ...] = (),
+    ) -> None:
+        """
+        Args:
+            img_size (int): Input image size.
+            patch_size (int): Patch size.
+            in_chans (int): Number of input image channels.
+            embed_dim (int): Patch embedding dimension.
+            depth (int): Depth of ViT.
+            num_heads (int): Number of attention heads in each ViT block.
+            mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+            qkv_bias (bool): If True, add a learnable bias to query, key, value.
+            norm_layer (nn.Module): Normalization layer.
+            act_layer (nn.Module): Activation layer.
+            use_abs_pos (bool): If True, use absolute positional embeddings.
+            use_rel_pos (bool): If True, add relative positional embeddings to the attention map.
+            rel_pos_zero_init (bool): If True, zero initialize relative positional parameters.
+            window_size (int): Window size for window attention blocks.
+            global_attn_indexes (list): Indexes for blocks using global attention.
+        """
+        super().__init__()
+        self.img_size = img_size
+
+        self.patch_embed = PatchEmbed(
+            kernel_size=(patch_size, patch_size),
+            stride=(patch_size, patch_size),
+            in_chans=in_chans,
+            embed_dim=embed_dim,
+        )
+
+        self.pos_embed: Optional[nn.Parameter] = None
+        if use_abs_pos:
+            # Initialize absolute positional embedding with pretrain image size.
+            self.pos_embed = nn.Parameter(
+                torch.zeros(1, img_size // patch_size, img_size // patch_size, embed_dim)
+            )
+
+        self.blocks = nn.ModuleList()
+        for i in range(depth):
+            block = Block(
+                dim=embed_dim,
+                num_heads=num_heads,
+                mlp_ratio=mlp_ratio,
+                qkv_bias=qkv_bias,
+                norm_layer=norm_layer,
+                act_layer=act_layer,
+                use_rel_pos=use_rel_pos,
+                rel_pos_zero_init=rel_pos_zero_init,
+                window_size=window_size if i not in global_attn_indexes else 0,
+                input_size=(img_size // patch_size, img_size // patch_size),
+            )
+            self.blocks.append(block)
+
+        self.neck = nn.Sequential(
+            nn.Conv2d(
+                embed_dim,
+                out_chans,
+                kernel_size=1,
+                bias=False,
+            ),
+            LayerNorm2d(out_chans),
+            nn.Conv2d(
+                out_chans,
+                out_chans,
+                kernel_size=3,
+                padding=1,
+                bias=False,
+            ),
+            LayerNorm2d(out_chans),
+        )
+
+        self.net_2 = nn.Conv2d(256, 512, kernel_size=3, stride=2, padding=1, bias=False)
+        self.net_3 = nn.Conv2d(512, 1024, kernel_size=3, stride=2, padding=1, bias=False)
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        x = self.patch_embed(x)
+        if self.pos_embed is not None:
+            # x = x + self.pos_embed
+            x = x + get_abs_pos(self.pos_embed, x.size(1))
+
+        for blk in self.blocks:
+            x = blk(x)
+
+        neck_output  = self.neck(x.permute(0, 3, 1, 2))
+        conv2_output  = self.net_2(neck_output)
+        # print(f"conv2_output shape: {conv2_output.shape}")
+        conv3_output  = self.net_3(conv2_output)
+
+        return conv3_output 
+
+
+class Block(nn.Module):
+    """Transformer blocks with support of window attention and residual propagation blocks"""
+
+    def __init__(
+        self,
+        dim: int,
+        num_heads: int,
+        mlp_ratio: float = 4.0,
+        qkv_bias: bool = True,
+        norm_layer: Type[nn.Module] = nn.LayerNorm,
+        act_layer: Type[nn.Module] = nn.GELU,
+        use_rel_pos: bool = False,
+        rel_pos_zero_init: bool = True,
+        window_size: int = 0,
+        input_size: Optional[Tuple[int, int]] = None,
+    ) -> None:
+        """
+        Args:
+            dim (int): Number of input channels.
+            num_heads (int): Number of attention heads in each ViT block.
+            mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+            qkv_bias (bool): If True, add a learnable bias to query, key, value.
+            norm_layer (nn.Module): Normalization layer.
+            act_layer (nn.Module): Activation layer.
+            use_rel_pos (bool): If True, add relative positional embeddings to the attention map.
+            rel_pos_zero_init (bool): If True, zero initialize relative positional parameters.
+            window_size (int): Window size for window attention blocks. If it equals 0, then
+                use global attention.
+            input_size (tuple(int, int) or None): Input resolution for calculating the relative
+                positional parameter size.
+        """
+        super().__init__()
+        self.norm1 = norm_layer(dim)
+        self.attn = Attention(
+            dim,
+            num_heads=num_heads,
+            qkv_bias=qkv_bias,
+            use_rel_pos=use_rel_pos,
+            rel_pos_zero_init=rel_pos_zero_init,
+            input_size=input_size if window_size == 0 else (window_size, window_size),
+        )
+
+        self.norm2 = norm_layer(dim)
+        self.mlp = MLPBlock(embedding_dim=dim, mlp_dim=int(dim * mlp_ratio), act=act_layer)
+
+        self.window_size = window_size
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        shortcut = x
+        x = self.norm1(x)
+        # Window partition
+        if self.window_size > 0:
+            H, W = x.shape[1], x.shape[2]
+            x, pad_hw = window_partition(x, self.window_size)
+
+        x = self.attn(x)
+        # Reverse window partition
+        if self.window_size > 0:
+            x = window_unpartition(x, self.window_size, pad_hw, (H, W))
+
+        x = shortcut + x
+        x = x + self.mlp(self.norm2(x))
+
+        return x
+
+
+class Attention(nn.Module):
+    """Multi-head Attention block with relative position embeddings."""
+
+    def __init__(
+        self,
+        dim: int,
+        num_heads: int = 8,
+        qkv_bias: bool = True,
+        use_rel_pos: bool = False,
+        rel_pos_zero_init: bool = True,
+        input_size: Optional[Tuple[int, int]] = None,
+    ) -> None:
+        """
+        Args:
+            dim (int): Number of input channels.
+            num_heads (int): Number of attention heads.
+            qkv_bias (bool):  If True, add a learnable bias to query, key, value.
+            rel_pos (bool): If True, add relative positional embeddings to the attention map.
+            rel_pos_zero_init (bool): If True, zero initialize relative positional parameters.
+            input_size (tuple(int, int) or None): Input resolution for calculating the relative
+                positional parameter size.
+        """
+        super().__init__()
+        self.num_heads = num_heads
+        head_dim = dim // num_heads
+        self.scale = head_dim**-0.5
+
+        self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
+        self.proj = nn.Linear(dim, dim)
+
+        self.use_rel_pos = use_rel_pos
+        if self.use_rel_pos:
+            assert (
+                input_size is not None
+            ), "Input size must be provided if using relative positional encoding."
+            # initialize relative positional embeddings
+            self.rel_pos_h = nn.Parameter(torch.zeros(2 * input_size[0] - 1, head_dim))
+            self.rel_pos_w = nn.Parameter(torch.zeros(2 * input_size[1] - 1, head_dim))
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        B, H, W, _ = x.shape
+        # qkv with shape (3, B, nHead, H * W, C)
+        qkv = self.qkv(x).reshape(B, H * W, 3, self.num_heads, -1).permute(2, 0, 3, 1, 4)
+        # q, k, v with shape (B * nHead, H * W, C)
+        q, k, v = qkv.reshape(3, B * self.num_heads, H * W, -1).unbind(0)
+
+        rel_h, rel_w = None, None
+        if self.use_rel_pos:
+            rel_h, rel_w = add_decomposed_rel_pos(q, self.rel_pos_h, self.rel_pos_w, (H, W), (H, W))
+
+        q = q.view(B, self.num_heads, H * W, -1)
+        k = k.view(B, self.num_heads, H * W, -1)
+        v = v.view(B, self.num_heads, H * W, -1)
+
+        if self.use_rel_pos:
+            rel_h = rel_h.view(B, self.num_heads, rel_h.size(1), rel_h.size(2), rel_h.size(3))
+            rel_w = rel_w.view(B, self.num_heads, rel_w.size(1), rel_w.size(2), rel_w.size(3))
+            attn_bias = (rel_h + rel_w).view(B, self.num_heads, rel_h.size(2), rel_h.size(3) * rel_w.size(4))
+            x = torch.nn.functional.scaled_dot_product_attention(q, k, v, attn_mask=attn_bias)
+            # x = _attention_rel_h_rel_w(q, k, v, rel_h, rel_w)
+        else:
+            x = torch.nn.functional.scaled_dot_product_attention(q, k, v)
+            # qkv = torch.stack([q, k, v], dim=1).transpose(1, 3).reshape(B, H * W, 3, self.num_heads, -1)
+            # x = flash_attn_qkvpacked_func(qkv, dropout_p=0.0, causal=False).transpose(1, 2)
+
+        
+
+        x = x.view(B, self.num_heads, H, W, -1).permute(0, 2, 3, 1, 4).reshape(B, H, W, -1)
+
+        x = self.proj(x)
+
+        return x
+
+
+def window_partition(x: torch.Tensor, window_size: int) -> Tuple[torch.Tensor, Tuple[int, int]]:
+    """
+    Partition into non-overlapping windows with padding if needed.
+    Args:
+        x (tensor): input tokens with [B, H, W, C].
+        window_size (int): window size.
+
+    Returns:
+        windows: windows after partition with [B * num_windows, window_size, window_size, C].
+        (Hp, Wp): padded height and width before partition
+    """
+    B, H, W, C = x.shape
+
+    pad_h = (window_size - H % window_size) % window_size
+    pad_w = (window_size - W % window_size) % window_size
+    if pad_h > 0 or pad_w > 0:
+        x = F.pad(x, (0, 0, 0, pad_w, 0, pad_h))
+    Hp, Wp = H + pad_h, W + pad_w
+
+    x = x.view(B, Hp // window_size, window_size, Wp // window_size, window_size, C)
+    windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
+    return windows, (Hp, Wp)
+
+
+def window_unpartition(
+    windows: torch.Tensor, window_size: int, pad_hw: Tuple[int, int], hw: Tuple[int, int]
+) -> torch.Tensor:
+    """
+    Window unpartition into original sequences and removing padding.
+    Args:
+        windows (tensor): input tokens with [B * num_windows, window_size, window_size, C].
+        window_size (int): window size.
+        pad_hw (Tuple): padded height and width (Hp, Wp).
+        hw (Tuple): original height and width (H, W) before padding.
+
+    Returns:
+        x: unpartitioned sequences with [B, H, W, C].
+    """
+    Hp, Wp = pad_hw
+    H, W = hw
+    B = windows.shape[0] // (Hp * Wp // window_size // window_size)
+    x = windows.view(B, Hp // window_size, Wp // window_size, window_size, window_size, -1)
+    x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, Hp, Wp, -1)
+
+    if Hp > H or Wp > W:
+        x = x[:, :H, :W, :].contiguous()
+    return x
+
+
+def get_rel_pos(q_size: int, k_size: int, rel_pos: torch.Tensor) -> torch.Tensor:
+    """
+    Get relative positional embeddings according to the relative positions of
+        query and key sizes.
+    Args:
+        q_size (int): size of query q.
+        k_size (int): size of key k.
+        rel_pos (Tensor): relative position embeddings (L, C).
+
+    Returns:
+        Extracted positional embeddings according to relative positions.
+    """
+    max_rel_dist = int(2 * max(q_size, k_size) - 1)
+    # Interpolate rel pos if needed.
+    if rel_pos.shape[0] != max_rel_dist:
+        # Interpolate rel pos.
+        dtype = rel_pos.dtype
+        rel_pos = rel_pos.to(torch.float32)
+        rel_pos_resized = F.interpolate(
+            rel_pos.reshape(1, rel_pos.shape[0], -1).permute(0, 2, 1),
+            size=max_rel_dist,
+            mode="linear",
+        ).to(dtype)
+        rel_pos_resized = rel_pos_resized.reshape(-1, max_rel_dist).permute(1, 0)
+    else:
+        rel_pos_resized = rel_pos
+
+    # Scale the coords with short length if shapes for q and k are different.
+    q_coords = torch.arange(q_size, device=rel_pos.device)[:, None] * max(k_size / q_size, 1.0)
+    k_coords = torch.arange(k_size, device=rel_pos.device)[None, :] * max(q_size / k_size, 1.0)
+    relative_coords = (q_coords - k_coords) + (k_size - 1) * max(q_size / k_size, 1.0)
+
+    return rel_pos_resized[relative_coords.long()]
+
+
+def add_decomposed_rel_pos(
+    q: torch.Tensor,
+    rel_pos_h: torch.Tensor,
+    rel_pos_w: torch.Tensor,
+    q_size: Tuple[int, int],
+    k_size: Tuple[int, int],
+) -> torch.Tensor:
+    """
+    Calculate decomposed Relative Positional Embeddings from :paper:`mvitv2`.
+    https://github.com/facebookresearch/mvit/blob/19786631e330df9f3622e5402b4a419a263a2c80/mvit/models/attention.py   # noqa B950
+    Args:
+        q (Tensor): query q in the attention layer with shape (B, q_h * q_w, C).
+        rel_pos_h (Tensor): relative position embeddings (Lh, C) for height axis.
+        rel_pos_w (Tensor): relative position embeddings (Lw, C) for width axis.
+        q_size (Tuple): spatial sequence size of query q with (q_h, q_w).
+        k_size (Tuple): spatial sequence size of key k with (k_h, k_w).
+
+    Returns:
+        attn (Tensor): attention map with added relative positional embeddings.
+    """
+    q_h, q_w = q_size
+    k_h, k_w = k_size
+    Rh = get_rel_pos(q_h, k_h, rel_pos_h)
+    Rw = get_rel_pos(q_w, k_w, rel_pos_w)
+
+    B, _, dim = q.shape
+    r_q = q.reshape(B, q_h, q_w, dim)
+    rel_h = torch.einsum("bhwc,hkc->bhwk", r_q, Rh)
+    rel_w = torch.einsum("bhwc,wkc->bhwk", r_q, Rw)
+    rel_h = rel_h.unsqueeze(-1)
+    rel_w = rel_w.unsqueeze(-2)
+    rel_h = rel_h.reshape(B, q_h * q_w, k_h, 1)
+    rel_w = rel_w.reshape(B, q_h * q_w, 1, k_w)
+
+    return rel_h, rel_w
+
+
+class PatchEmbed(nn.Module):
+    """
+    Image to Patch Embedding.
+    """
+
+    def __init__(
+        self,
+        kernel_size: Tuple[int, int] = (16, 16),
+        stride: Tuple[int, int] = (16, 16),
+        padding: Tuple[int, int] = (0, 0),
+        in_chans: int = 3,
+        embed_dim: int = 768,
+    ) -> None:
+        """
+        Args:
+            kernel_size (Tuple): kernel size of the projection layer.
+            stride (Tuple): stride of the projection layer.
+            padding (Tuple): padding size of the projection layer.
+            in_chans (int): Number of input image channels.
+            embed_dim (int): Patch embedding dimension.
+        """
+        super().__init__()
+
+        self.proj = nn.Conv2d(
+            in_chans, embed_dim, kernel_size=kernel_size, stride=stride, padding=padding
+        )
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        x = self.proj(x)
+        # B C H W -> B H W C
+        x = x.permute(0, 2, 3, 1)
+        return x
+
+
+def build_sam_vit_b(checkpoint=None):
+    return _build_sam(
+        encoder_embed_dim=768,
+        encoder_depth=12,
+        encoder_num_heads=12,
+        encoder_global_attn_indexes=[2, 5, 8, 11],
+        checkpoint=checkpoint,
+    )
+
+
+def _build_sam(
+    encoder_embed_dim,
+    encoder_depth,
+    encoder_num_heads,
+    encoder_global_attn_indexes,
+    checkpoint=None,
+):
+    prompt_embed_dim = 256
+    image_size = 1024
+    vit_patch_size = 16
+    image_embedding_size = image_size // vit_patch_size
+    image_encoder=ImageEncoderViT(
+            depth=encoder_depth,
+            embed_dim=encoder_embed_dim,
+            img_size=image_size,
+            mlp_ratio=4,
+            norm_layer=partial(torch.nn.LayerNorm, eps=1e-6),
+            num_heads=encoder_num_heads,
+            patch_size=vit_patch_size,
+            qkv_bias=True,
+            use_rel_pos=True,
+            global_attn_indexes=encoder_global_attn_indexes,
+            window_size=14,
+            out_chans=prompt_embed_dim,
+        )
+    
+    if checkpoint is not None:
+        # with open(checkpoint, "rb") as f:
+        state_dict = torch.load(checkpoint)
+        # print(state_dict.keys())
+        # for key in state_dict:
+        # image_encoder.load_state_dict({k[14:]: v for k, v in state_dict.items() if 'image_encoder' in k}, strict=False)
+        # ocr-anyting
+        # image_encoder.load_state_dict(state_dict, strict=True)
+        # tob
+        image_encoder.load_state_dict({k[30:]: v for k, v in state_dict.items() if 'vision_tower_high' in k}, strict=True)
+        print(checkpoint)
+    return image_encoder

deepseek_ocr.py

@@ -0,0 +1,582 @@
+
+"""Inference-only Deepseek-OCR model compatible with HuggingFace weights."""
+import math
+from collections.abc import Iterable, Mapping, Sequence
+from typing import List, Literal, Optional, Set, Tuple, TypedDict, Union
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from einops import rearrange, repeat
+from transformers import BatchFeature
+
+from vllm.config import VllmConfig
+from vllm.model_executor import SamplingMetadata
+from vllm.model_executor.layers.quantization import QuantizationConfig
+from vllm.model_executor.model_loader.utils import set_default_torch_dtype
+from vllm.multimodal import MULTIMODAL_REGISTRY
+from vllm.multimodal.inputs import (MultiModalDataDict, MultiModalFieldConfig,
+                                    MultiModalKwargs, NestedTensors)
+from vllm.multimodal.parse import (ImageEmbeddingItems, ImageProcessorItems,
+                                   ImageSize, MultiModalDataItems)
+from vllm.multimodal.processing import (BaseMultiModalProcessor,
+                                        BaseProcessingInfo, PromptReplacement,
+                                        PromptUpdate)
+from vllm.multimodal.profiling import BaseDummyInputsBuilder
+from vllm.sequence import IntermediateTensors
+from vllm.transformers_utils.configs.deepseek_vl2 import (DeepseekVLV2Config,
+                                                          MlpProjectorConfig,
+                                                          VisionEncoderConfig)
+from process.image_process import (
+    DeepseekOCRProcessor, count_tiles)
+from vllm.transformers_utils.tokenizer import cached_tokenizer_from_config
+# from vllm.utils import is_list_of
+
+from vllm.model_executor.models.interfaces import MultiModalEmbeddings, SupportsMultiModal, SupportsPP
+from vllm.model_executor.models.utils import (AutoWeightsLoader, WeightsMapper, flatten_bn,
+                    init_vllm_registered_model, maybe_prefix,
+                    merge_multimodal_embeddings)
+
+from deepencoder.sam_vary_sdpa import build_sam_vit_b
+from deepencoder.clip_sdpa import build_clip_l
+from deepencoder.build_linear import MlpProjector
+from addict import Dict
+# import time
+from config import IMAGE_SIZE, BASE_SIZE, CROP_MODE, PRINT_NUM_VIS_TOKENS, PROMPT
+# The image token id may be various
+_IMAGE_TOKEN = "<image>"
+
+
+class DeepseekOCRProcessingInfo(BaseProcessingInfo):
+
+    def get_hf_config(self):
+        return self.ctx.get_hf_config(DeepseekVLV2Config)
+
+    def get_hf_processor(self, **kwargs: object):
+        return self.ctx.get_hf_processor(DeepseekOCRProcessor, **kwargs)
+
+    def get_supported_mm_limits(self) -> Mapping[str, Optional[int]]:
+        return {"image": None}
+
+    def get_num_image_tokens(self,
+                             *,
+                             image_width: int,
+                             image_height: int,
+                             cropping: bool = True) -> int:
+        hf_processor = self.get_hf_processor()
+
+
+        # image_size = hf_processor.image_size
+        # patch_size = hf_processor.patch_size
+        # downsample_ratio = hf_processor.downsample_ratio
+
+        image_size = IMAGE_SIZE
+        base_size = BASE_SIZE
+        patch_size = 16
+        downsample_ratio = 4
+
+        if CROP_MODE:
+            if image_width <= 640 and image_height <= 640:
+                crop_ratio = [1, 1]
+            else:
+                # images_crop_raw, crop_ratio = hf_processor.dynamic_preprocess(image)
+
+                # find the closest aspect ratio to the target
+                crop_ratio = count_tiles(image_width, image_height, image_size=IMAGE_SIZE)
+
+                # print('===========')
+                # print('crop_ratio ', crop_ratio)
+                # print('============')
+                
+            num_width_tiles, num_height_tiles = crop_ratio
+        else:
+            num_width_tiles = num_height_tiles = 1
+
+        h = w = math.ceil((base_size // patch_size) / downsample_ratio)
+
+        h2 = w2 = math.ceil((image_size // patch_size) / downsample_ratio)
+
+        global_views_tokens = h * (w + 1)
+        if num_width_tiles >1 or num_height_tiles>1:
+            local_views_tokens = (num_height_tiles * h2) * (num_width_tiles * w2 + 1)
+        else:
+            local_views_tokens = 0
+
+
+        return global_views_tokens + local_views_tokens + 1
+
+    def get_image_size_with_most_features(self) -> ImageSize:
+
+        if IMAGE_SIZE == 1024 and BASE_SIZE == 1280:
+            return ImageSize(width=1024*2, height=1024*2)
+        return ImageSize(width=640*2, height=640*2)
+
+
+class DeepseekOCRDummyInputsBuilder(
+        BaseDummyInputsBuilder[DeepseekOCRProcessingInfo]):
+
+    def get_dummy_text(self, mm_counts: Mapping[str, int]) -> str:
+        num_images = mm_counts.get("image", 0)
+
+        processor = self.info.get_hf_processor()
+        image_token = processor.image_token
+
+        return image_token * num_images
+
+    def get_dummy_mm_data(
+        self,
+        seq_len: int,
+        mm_counts: Mapping[str, int],
+    ) -> MultiModalDataDict:
+        num_images = mm_counts.get("image", 0)
+
+        max_image_size = self.info.get_image_size_with_most_features()
+
+        if '<image>' in PROMPT:
+            return {
+                "image":
+                DeepseekOCRProcessor().tokenize_with_images(images = self._get_dummy_images(width=max_image_size.width,
+                                    height=max_image_size.height,
+                                    num_images=num_images), bos=True, eos=True, cropping=CROP_MODE)
+            }
+        else:
+            return {
+                "image": []
+            }
+
+
+
+
+class DeepseekOCRMultiModalProcessor(
+        BaseMultiModalProcessor[DeepseekOCRProcessingInfo]):
+    
+
+    def _call_hf_processor(
+        self,
+        prompt: str,
+        mm_data: Mapping[str, object],
+        mm_kwargs: Mapping[str, object],
+    ) -> BatchFeature:
+        
+        
+        # print(mm_data)
+        if mm_data:
+            processed_outputs = self.info.ctx.call_hf_processor(
+                self.info.get_hf_processor(**mm_kwargs),
+                dict(prompt=prompt, **mm_data),
+                mm_kwargs,
+            )
+
+        else:
+            tokenizer = self.info.get_tokenizer()
+            processed_outputs = tokenizer(prompt,
+                                          add_special_tokens=True,
+                                          return_tensors="pt")
+
+        return processed_outputs
+
+    def _get_mm_fields_config(
+        self,
+        hf_inputs: BatchFeature,
+        hf_processor_mm_kwargs: Mapping[str, object],
+    ) -> Mapping[str, MultiModalFieldConfig]:
+        return dict(
+            pixel_values=MultiModalFieldConfig.batched("image"),
+            images_spatial_crop=MultiModalFieldConfig.batched("image"),
+            # image_embeds=MultiModalFieldConfig.batched("image2"),
+            images_crop=MultiModalFieldConfig.batched("image"),
+        )
+
+    def _get_prompt_updates(
+        self,
+        mm_items: MultiModalDataItems,
+        hf_processor_mm_kwargs: Mapping[str, object],
+        out_mm_kwargs: MultiModalKwargs,
+    ) -> Sequence[PromptUpdate]:
+        hf_processor = self.info.get_hf_processor(**hf_processor_mm_kwargs)
+
+        image_token_id = hf_processor.image_token_id
+        assert isinstance(image_token_id, int)
+
+        def get_replacement_deepseek_vl2(item_idx: int):
+            images = mm_items.get_items(
+                "image", (ImageEmbeddingItems, ImageProcessorItems))
+
+
+
+            if isinstance(images, ImageEmbeddingItems):
+                num_image_tokens = images.get_feature_size(item_idx)
+            else:
+
+                
+                width = images[0][-1][0][0]
+                height = images[0][-1][0][1]
+
+                num_image_tokens = self.info.get_num_image_tokens(
+                    image_width=width,
+                    image_height=height,
+                    # flag = True,
+                    cropping=CROP_MODE,
+                )
+            return [image_token_id] * num_image_tokens
+
+        return [
+            PromptReplacement(
+                modality="image",
+                target=[image_token_id],
+                replacement=get_replacement_deepseek_vl2,
+            )
+        ]
+
+    def _cached_apply_hf_processor(
+        self,
+        prompt: Union[str, list[int]],
+        mm_data_items: MultiModalDataItems,
+        hf_processor_mm_kwargs: Mapping[str, object],
+    ) -> tuple[list[int], MultiModalKwargs, bool]:
+        # The processor logic is different for len(images) <= 2 vs > 2
+        # Since the processing cache assumes that the processor output is
+        # invariant of how many images are passed per prompt, we only
+        # perform caching for the most common case
+        if mm_data_items.get_count("image", strict=False) > 2:
+            # This code path corresponds to the cache being disabled
+            return self._apply_hf_processor_main(
+                prompt=prompt,
+                mm_items=mm_data_items,
+                hf_processor_mm_kwargs=hf_processor_mm_kwargs,
+                enable_hf_prompt_update=True,
+            )
+
+        return super()._cached_apply_hf_processor(
+            prompt=prompt,
+            mm_data_items=mm_data_items,
+            hf_processor_mm_kwargs=hf_processor_mm_kwargs,
+        )
+
+
+@MULTIMODAL_REGISTRY.register_processor(
+    DeepseekOCRMultiModalProcessor,
+    info=DeepseekOCRProcessingInfo,
+    dummy_inputs=DeepseekOCRDummyInputsBuilder)
+class DeepseekOCRForCausalLM(nn.Module, SupportsMultiModal, SupportsPP):
+
+    hf_to_vllm_mapper = WeightsMapper(orig_to_new_prefix={
+        "language.": "language_model.",
+    })
+
+    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
+        super().__init__()
+
+        config: DeepseekVLV2Config = vllm_config.model_config.hf_config
+        quant_config = vllm_config.quant_config
+        multimodal_config = vllm_config.model_config.multimodal_config
+
+        # config.model_type ='deepseek_vl_v2'
+
+        self.config = config
+        self.multimodal_config = multimodal_config
+
+
+        self.vision_config = config.vision_config
+        self.projector_config = config.projector_config
+        self.text_config = config.text_config
+
+        model_config = vllm_config.model_config
+        tokenizer = cached_tokenizer_from_config(model_config)
+        self.image_token_id = tokenizer.vocab[_IMAGE_TOKEN]
+
+        self.sam_model = build_sam_vit_b()
+        self.vision_model = build_clip_l()
+
+        n_embed = 1280
+        self.projector =  MlpProjector(Dict(projector_type="linear", input_dim=2048, n_embed=n_embed))
+        self.tile_tag = config.tile_tag
+        self.global_view_pos = config.global_view_pos
+    
+        # self.sam_model = torch.compile(self.sam_model, mode="reduce-overhead")
+        # self.vision_model = torch.compile(self.vision_model, mode="reduce-overhead")
+        # self.projector = torch.compile(self.projector, mode="max-autotune")
+
+
+
+
+        # special token for image token sequence format
+        embed_std = 1 / torch.sqrt(torch.tensor(n_embed, dtype=torch.float32))
+        if self.tile_tag == "2D":
+            # <|view_separator|>, <|\n|>
+            self.image_newline = nn.Parameter(torch.randn(n_embed) * embed_std)
+            self.view_seperator = nn.Parameter(torch.randn(n_embed) * embed_std)
+        else:
+            raise ValueError(
+                f"Only 2D tile_tag is supported currently, got: {self.tile_tag}"
+            )
+
+        if self.text_config.topk_method == "noaux_tc":
+            architectures = ["DeepseekV3ForCausalLM"]
+        elif not self.text_config.use_mla:
+            architectures = ["DeepseekForCausalLM"]
+        else:
+            architectures = ["DeepseekV2ForCausalLM"]
+
+        self.language_model = init_vllm_registered_model(
+            vllm_config=vllm_config,
+            hf_config=self.text_config,
+            prefix=maybe_prefix(prefix, "language"),
+            architectures=architectures,
+        )
+
+        self.make_empty_intermediate_tensors = (
+            self.language_model.make_empty_intermediate_tensors)
+
+
+
+    def _parse_and_validate_image_input(
+            self, **kwargs: object):
+        
+        pixel_values = kwargs.pop("pixel_values", None)
+        images_spatial_crop = kwargs.pop("images_spatial_crop", None)
+        images_crop = kwargs.pop("images_crop", None)
+
+
+        if pixel_values is None or torch.sum(pixel_values).item() == 0:
+            return None
+
+        if pixel_values is not None:
+            if not isinstance(pixel_values, (torch.Tensor, list)):
+                raise ValueError("Incorrect type of pixel values. "
+                                 f"Got type: {type(pixel_values)}")
+
+            if not isinstance(images_spatial_crop, (torch.Tensor, list)):
+                raise ValueError("Incorrect type of image sizes. "
+                                 f"Got type: {type(images_spatial_crop)}")
+            
+            if not isinstance(images_crop, (torch.Tensor, list)):
+                raise ValueError("Incorrect type of image crop. "
+                                 f"Got type: {type(images_crop)}")
+
+            return [pixel_values, images_crop, images_spatial_crop]
+
+
+        raise AssertionError("This line should be unreachable.")
+    
+
+
+    def _pixel_values_to_embedding(
+        self,
+        pixel_values: torch.Tensor,
+        images_crop: torch.Tensor,
+        images_spatial_crop: torch.Tensor,
+    ) -> NestedTensors:
+
+        # Pixel_values (global view): [n_image, batch_size, 3, height, width]
+        # images_spatial_crop: [n_image, batch_size, [num_tiles_w, num_tiles_h]]
+        # images_crop (local view): [n_image, batch_size, num_pathes, 3, h, w]
+        # split the pixel and image_crop, all batch_size = 1
+
+        images_in_this_batch = []
+
+
+        # print(type(images_crop))
+
+        # print(pixel_values.shape)
+
+
+        with torch.no_grad():
+            for jdx in range(images_spatial_crop.size(0)):
+                # with torch.set_grad_enabled(False):
+                patches = images_crop[jdx][0].to(torch.bfloat16) # batch_size = 1
+                image_ori = pixel_values[jdx]
+                crop_shape = images_spatial_crop[jdx][0]
+
+                if torch.sum(patches).item() != 0:  # if all values = 0, no crop
+                    # P, C, H, W = patches.shape
+                    # crop_flag = 1
+                    local_features_1 = self.sam_model(patches)
+                    #TODO del patches 
+                    # torch.compiler.cudagraph_mark_step_begin()
+                    local_features_2 = self.vision_model(patches, local_features_1)  
+
+
+                    local_features = torch.cat((local_features_2[:, 1:], local_features_1.flatten(2).permute(0, 2, 1)), dim=-1) 
+                    local_features = self.projector(local_features)
+
+
+                    global_features_1 = self.sam_model(image_ori)
+                    global_features_2 = self.vision_model(image_ori, global_features_1) 
+                    global_features = torch.cat((global_features_2[:, 1:], global_features_1.flatten(2).permute(0, 2, 1)), dim=-1) 
+                    global_features = self.projector(global_features)
+
+                    if PRINT_NUM_VIS_TOKENS:
+                        print('=====================')
+                        print('BASE: ', global_features.shape)
+                        print('PATCHES: ', local_features.shape)
+                        print('=====================')
+
+                    _, hw, n_dim = global_features.shape
+                    h = w = int(hw ** 0.5)
+
+                    _2, hw2, n_dim2 = local_features.shape
+                    h2 = w2 = int(hw2 ** 0.5)
+
+                    width_crop_num, height_crop_num = crop_shape[0], crop_shape[1]
+
+                    global_features = global_features.view(h, w, n_dim)
+
+                    global_features = torch.cat(
+                        [global_features, self.image_newline[None, None, :].expand(h, 1, n_dim)], dim=1
+                    )
+
+                    global_features = global_features.view(-1, n_dim)
+
+
+                    local_features = local_features.view(height_crop_num, width_crop_num, h2, w2, n_dim2).permute(0, 2, 1, 3, 4).reshape(height_crop_num*h2, width_crop_num*w2, n_dim2)
+                    local_features = torch.cat(
+                        [local_features, self.image_newline[None, None, :].expand(height_crop_num * h2, 1, n_dim2)], dim=1
+                    )
+                    local_features = local_features.view(-1, n_dim2)
+
+                    global_local_features = torch.cat([local_features, global_features, self.view_seperator[None, :]], dim=0)
+                
+                else:
+                    global_features_1 = self.sam_model(image_ori)
+                    global_features_2 = self.vision_model(image_ori, global_features_1) 
+                    global_features = torch.cat((global_features_2[:, 1:], global_features_1.flatten(2).permute(0, 2, 1)), dim=-1) 
+                    global_features = self.projector(global_features)
+
+                    if PRINT_NUM_VIS_TOKENS:
+                        print('=====================')
+                        print('BASE: ', global_features.shape)
+                        print('NO PATCHES')
+                        print('=====================')
+
+                    _, hw, n_dim = global_features.shape
+                    h = w = int(hw ** 0.5)
+
+                    global_features = global_features.view(h, w, n_dim)
+
+                    global_features = torch.cat(
+                        [global_features, self.image_newline[None, None, :].expand(h, 1, n_dim)], dim=1
+                    )
+
+                    global_features = global_features.view(-1, n_dim)
+
+                    global_local_features = torch.cat([global_features, self.view_seperator[None, :]], dim=0)
+
+                images_in_this_batch.append(global_local_features)
+
+        return images_in_this_batch
+
+    def _process_image_input(
+            self, image_input) -> torch.Tensor:
+        
+
+        # image_input: [pixel_values, images_crop, images_spatial_crop]
+    
+        pixel_values = image_input[0].to(torch.bfloat16)
+        # print(image_input[1][0].shape)
+        # print(type(image_input[1]))
+        # exit()
+
+        # images_crop = image_input[1].to(torch.bfloat16)
+        images_crop = image_input[1]
+        # images_crop = image_input[1]
+        images_spatial_crop = image_input[2].to(dtype=torch.long)
+
+        # local_start = time.time()
+        vision_features = self._pixel_values_to_embedding(
+            pixel_values=pixel_values, images_crop = images_crop,  images_spatial_crop=images_spatial_crop)
+
+        # local_total_time = time.time() - local_start
+
+        # print('encoder_time: ', local_total_time)
+        # exit()
+        return vision_features
+
+    def get_language_model(self) -> torch.nn.Module:
+        return self.language_model
+
+    def get_multimodal_embeddings(
+            self, **kwargs: object) -> Optional[MultiModalEmbeddings]:
+        image_input = self._parse_and_validate_image_input(**kwargs)
+        if image_input is None:
+            return None
+        vision_embeddings = self._process_image_input(image_input)
+        return vision_embeddings
+    
+
+
+    def get_input_embeddings(
+        self,
+        input_ids: torch.Tensor,
+        multimodal_embeddings: Optional[MultiModalEmbeddings] = None,
+    ) -> torch.Tensor:
+        
+
+
+        inputs_embeds = self.language_model.get_input_embeddings(input_ids)
+
+
+        if multimodal_embeddings is not None:
+            inputs_embeds = merge_multimodal_embeddings(
+                input_ids, inputs_embeds, multimodal_embeddings,
+                self.image_token_id)
+            # print(len(multimodal_embeddings))
+            # print(input_ids.shape)
+            # print(type(inputs_embeds))
+            # print(inputs_embeds.shape)
+            
+        return inputs_embeds
+
+    def forward(self,
+                input_ids: torch.Tensor,
+                positions: torch.Tensor,
+                intermediate_tensors: Optional[IntermediateTensors] = None,
+                inputs_embeds: Optional[torch.Tensor] = None,
+                **kwargs: object):
+
+        if intermediate_tensors is not None:
+            inputs_embeds = None
+
+        # NOTE: In v1, inputs_embeds is always generated at model runner, this
+        # condition is for v0 compatibility
+        elif inputs_embeds is None:
+            vision_embeddings = self.get_multimodal_embeddings(**kwargs)
+            inputs_embeds = self.get_input_embeddings(input_ids,
+                                                      vision_embeddings)
+            input_ids = None
+
+        hidden_states = self.language_model(input_ids,
+                                            positions,
+                                            intermediate_tensors,
+                                            inputs_embeds=inputs_embeds)
+
+        return hidden_states
+
+    def compute_logits(
+        self,
+        hidden_states: torch.Tensor,
+        sampling_metadata: SamplingMetadata,
+    ) -> Optional[torch.Tensor]:
+        return self.language_model.compute_logits(hidden_states,
+                                                  sampling_metadata)
+
+
+    def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]) -> Set[str]:
+        processed_weights = []
+        
+        for name, tensor in weights:
+            if 'sam_model' in name or 'vision_model' in name or 'projector' in name or 'image_newline' in name or 'view_seperator' in name:
+                new_name = name.replace('model.', '', 1)
+            else:
+                new_name = 'language.' + name
+
+            processed_weights.append((new_name, tensor))
+        
+        loader = AutoWeightsLoader(self)
+        autoloaded_weights = loader.load_weights(processed_weights, mapper=self.hf_to_vllm_mapper)
+
+
+
+
+
+        return autoloaded_weights

main.py

@@ -0,0 +1,538 @@
+#!/usr/bin/env python3
+"""
+DeepSeek-OCR Dataset Processing with vLLM
+
+This script processes image datasets through DeepSeek-OCR using vLLM for efficient batch processing.
+"""
+
+import argparse
+import asyncio
+import json
+import logging
+import os
+import sys
+import time
+from datetime import datetime
+from typing import List
+
+import torch
+from datasets import load_dataset
+from huggingface_hub import DatasetCard, login
+from PIL import Image, ImageOps
+from tqdm.auto import tqdm
+from transformers import AutoTokenizer
+from vllm import AsyncLLMEngine, SamplingParams
+from vllm.engine.arg_utils import AsyncEngineArgs
+from vllm.model_executor.models.registry import ModelRegistry
+
+# Import DeepSeek-OCR modules
+import config
+from deepseek_ocr import DeepseekOCRForCausalLM
+from process.image_process import DeepseekOCRProcessor
+from process.ngram_norepeat import NoRepeatNGramLogitsProcessor
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+# Resolution mode presets
+RESOLUTION_MODES = {
+    "tiny": {"base_size": 512, "image_size": 512, "crop_mode": False},
+    "small": {"base_size": 640, "image_size": 640, "crop_mode": False},
+    "base": {"base_size": 1024, "image_size": 1024, "crop_mode": False},
+    "large": {"base_size": 1280, "image_size": 1280, "crop_mode": False},
+    "gundam": {"base_size": 1024, "image_size": 640, "crop_mode": True},
+}
+
+
+def check_cuda_availability():
+    """Check if CUDA is available and exit if not."""
+    if not torch.cuda.is_available():
+        logger.error("CUDA is not available. This script requires a GPU.")
+        logger.error("Please run on a machine with a CUDA-capable GPU.")
+        sys.exit(1)
+    else:
+        logger.info(f"CUDA is available. GPU: {torch.cuda.get_device_name(0)}")
+
+
+def setup_config(resolution_mode: str):
+    """Set up global config based on resolution mode."""
+    if resolution_mode not in RESOLUTION_MODES:
+        raise ValueError(
+            f"Invalid resolution mode: {resolution_mode}. "
+            f"Choose from {list(RESOLUTION_MODES.keys())}"
+        )
+
+    mode_config = RESOLUTION_MODES[resolution_mode]
+    config.BASE_SIZE = mode_config["base_size"]
+    config.IMAGE_SIZE = mode_config["image_size"]
+    config.CROP_MODE = mode_config["crop_mode"]
+
+    logger.info(
+        f"Resolution mode: {resolution_mode} "
+        f"(BASE_SIZE={config.BASE_SIZE}, IMAGE_SIZE={config.IMAGE_SIZE}, "
+        f"CROP_MODE={config.CROP_MODE})"
+    )
+
+
+async def process_images_async(
+    images: List[Image.Image],
+    engine: AsyncLLMEngine,
+    processor: DeepseekOCRProcessor,
+    sampling_params: SamplingParams,
+    prompt: str,
+) -> List[str]:
+    """Process a batch of images asynchronously with vLLM."""
+    results = []
+
+    for image in images:
+        # Preprocess image
+        image = image.convert("RGB")
+        image_features = processor.tokenize_with_images(
+            images=[image], bos=True, eos=True, cropping=config.CROP_MODE
+        )
+
+        # Generate async
+        request_id = f"request-{int(time.time() * 1000)}"
+        request = {"prompt": prompt, "multi_modal_data": {"image": image_features}}
+
+        output_text = ""
+        async for request_output in engine.generate(request, sampling_params, request_id):
+            if request_output.outputs:
+                output_text = request_output.outputs[0].text
+
+        results.append(output_text.strip())
+
+    return results
+
+
+def create_dataset_card(
+    source_dataset: str,
+    model: str,
+    num_samples: int,
+    processing_time: str,
+    resolution_mode: str,
+    base_size: int,
+    image_size: int,
+    crop_mode: bool,
+    max_model_len: int,
+    max_tokens: int,
+    gpu_memory_utilization: float,
+    image_column: str = "image",
+    split: str = "train",
+) -> str:
+    """Create a dataset card documenting the OCR process."""
+    return f"""---
+tags:
+- ocr
+- document-processing
+- deepseek
+- deepseek-ocr
+- markdown
+- vllm
+- generated
+---
+
+# Document OCR using DeepSeek-OCR (vLLM)
+
+This dataset contains markdown-formatted OCR results from images in [{source_dataset}](https://huggingface.co/datasets/{source_dataset}) using DeepSeek-OCR with vLLM.
+
+## Processing Details
+
+- **Source Dataset**: [{source_dataset}](https://huggingface.co/datasets/{source_dataset})
+- **Model**: [{model}](https://huggingface.co/{model})
+- **Number of Samples**: {num_samples:,}
+- **Processing Time**: {processing_time}
+- **Processing Date**: {datetime.now().strftime("%Y-%m-%d %H:%M UTC")}
+
+### Configuration
+
+- **Image Column**: `{image_column}`
+- **Output Column**: `markdown`
+- **Dataset Split**: `{split}`
+- **Resolution Mode**: {resolution_mode}
+- **Base Size**: {base_size}
+- **Image Size**: {image_size}
+- **Crop Mode**: {crop_mode}
+- **Max Model Length**: {max_model_len:,} tokens
+- **Max Output Tokens**: {max_tokens:,}
+- **GPU Memory Utilization**: {gpu_memory_utilization:.1%}
+- **Implementation**: vLLM AsyncEngine (batch processing)
+
+## Model Information
+
+DeepSeek-OCR is a state-of-the-art document OCR model that excels at:
+- 📐 **LaTeX equations** - Mathematical formulas preserved in LaTeX format
+- 📊 **Tables** - Extracted and formatted as HTML/markdown
+- 📝 **Document structure** - Headers, lists, and formatting maintained
+- 🖼️ **Image grounding** - Spatial layout and bounding box information
+- 🔍 **Complex layouts** - Multi-column and hierarchical structures
+- 🌍 **Multilingual** - Supports multiple languages
+
+### Resolution Modes
+
+- **Tiny** (512×512): Fast processing, 64 vision tokens
+- **Small** (640×640): Balanced speed/quality, 100 vision tokens
+- **Base** (1024×1024): High quality, 256 vision tokens
+- **Large** (1280×1280): Maximum quality, 400 vision tokens
+- **Gundam** (dynamic): Adaptive multi-tile processing for large documents
+
+## Dataset Structure
+
+The dataset contains all original columns plus:
+- `markdown`: The extracted text in markdown format with preserved structure
+- `inference_info`: JSON list tracking all OCR models applied to this dataset
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+# Load the dataset
+dataset = load_dataset("{{{{output_dataset_id}}}}", split="{split}")
+
+# Access the markdown text
+for example in dataset:
+    print(example["markdown"])
+    break
+```
+
+## Reproduction
+
+This dataset was generated using the DeepSeek-OCR vLLM Space:
+
+```bash
+hf jobs run --flavor l4x1 \\
+    --secrets HF_TOKEN \\
+    hf.co/spaces/davanstrien/deepseek-ocr \\
+    python main.py \\
+    --input-dataset {source_dataset} \\
+    --output-dataset <output-dataset> \\
+    --resolution-mode {resolution_mode} \\
+    --image-column {image_column}
+```
+
+## Performance
+
+- **Processing Speed**: ~{num_samples / (float(processing_time.split()[0]) * 60) if processing_time.split()[0].replace('.','').isdigit() else 'N/A':.1f} images/second
+- **Processing Method**: Async batch processing with vLLM (optimized for throughput)
+
+Generated with 🤖 [DeepSeek-OCR Space](https://huggingface.co/spaces/davanstrien/deepseek-ocr)
+"""
+
+
+async def main_async(
+    input_dataset: str,
+    output_dataset: str,
+    image_column: str = "image",
+    model: str = "deepseek-ai/DeepSeek-OCR",
+    resolution_mode: str = "gundam",
+    max_model_len: int = 8192,
+    max_tokens: int = 8192,
+    gpu_memory_utilization: float = 0.75,
+    prompt: str = "<image>\n<|grounding|>Convert the document to markdown.",
+    hf_token: str = None,
+    split: str = "train",
+    max_samples: int = None,
+    private: bool = False,
+    shuffle: bool = False,
+    seed: int = 42,
+):
+    """Process images from HF dataset through DeepSeek-OCR model with vLLM."""
+
+    # Check CUDA availability
+    check_cuda_availability()
+
+    # Track processing start time
+    start_time = datetime.now()
+
+    # Enable HF_TRANSFER for faster downloads
+    os.environ["HF_HUB_ENABLE_HF_TRANSFER"] = "1"
+
+    # Login to HF if token provided
+    HF_TOKEN = hf_token or os.environ.get("HF_TOKEN")
+    if HF_TOKEN:
+        login(token=HF_TOKEN)
+
+    # Set up config for resolution mode
+    setup_config(resolution_mode)
+
+    # Initialize tokenizer
+    logger.info(f"Loading tokenizer from {model}...")
+    config.TOKENIZER = AutoTokenizer.from_pretrained(model, trust_remote_code=True)
+    config.MODEL_PATH = model
+    config.PROMPT = prompt
+
+    # Load dataset
+    logger.info(f"Loading dataset: {input_dataset}")
+    dataset = load_dataset(input_dataset, split=split)
+
+    # Validate image column
+    if image_column not in dataset.column_names:
+        raise ValueError(
+            f"Column '{image_column}' not found. Available: {dataset.column_names}"
+        )
+
+    # Shuffle if requested
+    if shuffle:
+        logger.info(f"Shuffling dataset with seed {seed}")
+        dataset = dataset.shuffle(seed=seed)
+
+    # Limit samples if requested
+    if max_samples:
+        dataset = dataset.select(range(min(max_samples, len(dataset))))
+        logger.info(f"Limited to {len(dataset)} samples")
+
+    # Register custom model
+    logger.info("Registering custom DeepSeek-OCR model...")
+    ModelRegistry.register_model("DeepseekOCRForCausalLM", DeepseekOCRForCausalLM)
+
+    # Initialize vLLM AsyncEngine
+    logger.info(f"Initializing vLLM AsyncEngine with model: {model}")
+    logger.info("This may take a few minutes on first run...")
+
+    engine_args = AsyncEngineArgs(
+        model=model,
+        hf_overrides={"architectures": ["DeepseekOCRForCausalLM"]},
+        block_size=256,
+        max_model_len=max_model_len,
+        enforce_eager=False,
+        trust_remote_code=True,
+        tensor_parallel_size=1,
+        gpu_memory_utilization=gpu_memory_utilization,
+    )
+    engine = AsyncLLMEngine.from_engine_args(engine_args)
+
+    # Set up sampling params
+    logits_processors = [
+        NoRepeatNGramLogitsProcessor(
+            ngram_size=30, window_size=90, whitelist_token_ids={128821, 128822}
+        )
+    ]
+
+    sampling_params = SamplingParams(
+        temperature=0.0,
+        max_tokens=max_tokens,
+        logits_processors=logits_processors,
+        skip_special_tokens=False,
+    )
+
+    # Initialize processor
+    processor = DeepseekOCRProcessor()
+
+    logger.info(f"Processing {len(dataset)} images with vLLM AsyncEngine")
+
+    # Process images one at a time (async but sequential for simplicity)
+    all_markdown = []
+    for idx in tqdm(range(len(dataset)), desc="DeepSeek-OCR processing"):
+        image = dataset[idx][image_column]
+
+        # Convert to PIL if needed
+        if not isinstance(image, Image.Image):
+            image = Image.open(image) if isinstance(image, str) else image
+
+        try:
+            image = ImageOps.exif_transpose(image.convert("RGB"))
+
+            # Process single image
+            results = await process_images_async(
+                [image], engine, processor, sampling_params, prompt
+            )
+            all_markdown.append(results[0])
+
+        except Exception as e:
+            logger.error(f"Error processing image {idx}: {e}")
+            all_markdown.append("[OCR FAILED]")
+
+    # Calculate processing time
+    processing_duration = datetime.now() - start_time
+    processing_time_str = f"{processing_duration.total_seconds() / 60:.1f} min"
+
+    # Add markdown column to dataset
+    logger.info("Adding markdown column to dataset")
+    dataset = dataset.add_column("markdown", all_markdown)
+
+    # Handle inference_info tracking
+    logger.info("Updating inference_info...")
+
+    if "inference_info" in dataset.column_names:
+        try:
+            existing_info = json.loads(dataset[0]["inference_info"])
+            if not isinstance(existing_info, list):
+                existing_info = [existing_info]
+        except (json.JSONDecodeError, TypeError):
+            existing_info = []
+        dataset = dataset.remove_columns(["inference_info"])
+    else:
+        existing_info = []
+
+    # Add new inference info
+    new_info = {
+        "column_name": "markdown",
+        "model_id": model,
+        "processing_date": datetime.now().isoformat(),
+        "resolution_mode": resolution_mode,
+        "base_size": config.BASE_SIZE,
+        "image_size": config.IMAGE_SIZE,
+        "crop_mode": config.CROP_MODE,
+        "prompt": prompt,
+        "max_tokens": max_tokens,
+        "gpu_memory_utilization": gpu_memory_utilization,
+        "max_model_len": max_model_len,
+        "script": "main.py",
+        "script_version": "1.0.0",
+        "space_url": "https://huggingface.co/spaces/davanstrien/deepseek-ocr",
+        "implementation": "vllm-async (optimized)",
+    }
+    existing_info.append(new_info)
+
+    # Add updated inference_info column
+    info_json = json.dumps(existing_info, ensure_ascii=False)
+    dataset = dataset.add_column("inference_info", [info_json] * len(dataset))
+
+    # Push to hub
+    logger.info(f"Pushing to {output_dataset}")
+    dataset.push_to_hub(output_dataset, private=private, token=HF_TOKEN)
+
+    # Create and push dataset card
+    logger.info("Creating dataset card...")
+    card_content = create_dataset_card(
+        source_dataset=input_dataset,
+        model=model,
+        num_samples=len(dataset),
+        processing_time=processing_time_str,
+        resolution_mode=resolution_mode,
+        base_size=config.BASE_SIZE,
+        image_size=config.IMAGE_SIZE,
+        crop_mode=config.CROP_MODE,
+        max_model_len=max_model_len,
+        max_tokens=max_tokens,
+        gpu_memory_utilization=gpu_memory_utilization,
+        image_column=image_column,
+        split=split,
+    )
+
+    card = DatasetCard(card_content)
+    card.push_to_hub(output_dataset, token=HF_TOKEN)
+    logger.info("✅ Dataset card created and pushed!")
+
+    logger.info("✅ OCR conversion complete!")
+    logger.info(f"Dataset available at: https://huggingface.co/datasets/{output_dataset}")
+    logger.info(f"Processing time: {processing_time_str}")
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(
+        description="OCR images to markdown using DeepSeek-OCR (vLLM AsyncEngine)",
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        epilog="""
+Resolution Modes:
+  tiny      512×512 pixels, fast processing (64 vision tokens)
+  small     640×640 pixels, balanced (100 vision tokens)
+  base      1024×1024 pixels, high quality (256 vision tokens)
+  large     1280×1280 pixels, maximum quality (400 vision tokens)
+  gundam    Dynamic multi-tile processing (adaptive)
+
+Examples:
+  # Basic usage with default Gundam mode
+  python main.py input-dataset output-dataset
+
+  # High quality processing
+  python main.py input-dataset output-dataset --resolution-mode large
+
+  # Fast processing for testing
+  python main.py input-dataset output-dataset --resolution-mode tiny --max-samples 100
+
+  # With HF Jobs
+  hf jobs run --flavor l4x1 --secrets HF_TOKEN \\
+      hf.co/spaces/davanstrien/deepseek-ocr \\
+      python main.py input-dataset output-dataset --resolution-mode gundam
+        """,
+    )
+
+    parser.add_argument("input_dataset", help="Input dataset ID from Hugging Face Hub")
+    parser.add_argument("output_dataset", help="Output dataset ID for Hugging Face Hub")
+    parser.add_argument(
+        "--image-column",
+        default="image",
+        help="Column containing images (default: image)",
+    )
+    parser.add_argument(
+        "--model",
+        default="deepseek-ai/DeepSeek-OCR",
+        help="Model to use (default: deepseek-ai/DeepSeek-OCR)",
+    )
+    parser.add_argument(
+        "--resolution-mode",
+        default="gundam",
+        choices=list(RESOLUTION_MODES.keys()),
+        help="Resolution mode preset (default: gundam)",
+    )
+    parser.add_argument(
+        "--max-model-len",
+        type=int,
+        default=8192,
+        help="Maximum model context length (default: 8192)",
+    )
+    parser.add_argument(
+        "--max-tokens",
+        type=int,
+        default=8192,
+        help="Maximum tokens to generate (default: 8192)",
+    )
+    parser.add_argument(
+        "--gpu-memory-utilization",
+        type=float,
+        default=0.75,
+        help="GPU memory utilization (default: 0.75)",
+    )
+    parser.add_argument(
+        "--prompt",
+        default="<image>\n<|grounding|>Convert the document to markdown.",
+        help="Prompt for OCR (default: grounding markdown conversion)",
+    )
+    parser.add_argument("--hf-token", help="Hugging Face API token")
+    parser.add_argument(
+        "--split", default="train", help="Dataset split to use (default: train)"
+    )
+    parser.add_argument(
+        "--max-samples",
+        type=int,
+        help="Maximum number of samples to process (for testing)",
+    )
+    parser.add_argument(
+        "--private", action="store_true", help="Make output dataset private"
+    )
+    parser.add_argument(
+        "--shuffle",
+        action="store_true",
+        help="Shuffle the dataset before processing (useful for random sampling)",
+    )
+    parser.add_argument(
+        "--seed",
+        type=int,
+        default=42,
+        help="Random seed for shuffling (default: 42)",
+    )
+
+    args = parser.parse_args()
+
+    # Run async main
+    asyncio.run(
+        main_async(
+            input_dataset=args.input_dataset,
+            output_dataset=args.output_dataset,
+            image_column=args.image_column,
+            model=args.model,
+            resolution_mode=args.resolution_mode,
+            max_model_len=args.max_model_len,
+            max_tokens=args.max_tokens,
+            gpu_memory_utilization=args.gpu_memory_utilization,
+            prompt=args.prompt,
+            hf_token=args.hf_token,
+            split=args.split,
+            max_samples=args.max_samples,
+            private=args.private,
+            shuffle=args.shuffle,
+            seed=args.seed,
+        )
+    )

process/image_process.py

@@ -0,0 +1,502 @@
+import math
+from typing import List, Tuple
+
+import torch
+import torchvision.transforms as T
+from PIL import Image, ImageOps
+from transformers import AutoProcessor, BatchFeature, LlamaTokenizerFast
+from transformers.processing_utils import ProcessorMixin
+from config import IMAGE_SIZE, BASE_SIZE, CROP_MODE, MIN_CROPS, MAX_CROPS, PROMPT, TOKENIZER
+
+def find_closest_aspect_ratio(aspect_ratio, target_ratios, width, height, image_size):
+    best_ratio_diff = float('inf')
+    best_ratio = (1, 1)
+    area = width * height
+    for ratio in target_ratios:
+        target_aspect_ratio = ratio[0] / ratio[1]
+        ratio_diff = abs(aspect_ratio - target_aspect_ratio)
+        if ratio_diff < best_ratio_diff:
+            best_ratio_diff = ratio_diff
+            best_ratio = ratio
+        elif ratio_diff == best_ratio_diff:
+            if area > 0.5 * image_size * image_size * ratio[0] * ratio[1]:
+                best_ratio = ratio
+    # print(f'width: {width}, height: {height}, best_ratio: {best_ratio}')
+    return best_ratio
+
+
+def count_tiles(orig_width, orig_height, min_num=MIN_CROPS, max_num=MAX_CROPS, image_size=640, use_thumbnail=False):
+    aspect_ratio = orig_width / orig_height
+
+    # calculate the existing image aspect ratio
+    target_ratios = set(
+        (i, j) for n in range(min_num, max_num + 1) for i in range(1, n + 1) for j in range(1, n + 1) if
+        i * j <= max_num and i * j >= min_num)
+    # print(target_ratios)
+    target_ratios = sorted(target_ratios, key=lambda x: x[0] * x[1])
+
+    # find the closest aspect ratio to the target
+    target_aspect_ratio = find_closest_aspect_ratio(
+        aspect_ratio, target_ratios, orig_width, orig_height, image_size)
+
+    return target_aspect_ratio
+
+
+def dynamic_preprocess(image, min_num=MIN_CROPS, max_num=MAX_CROPS, image_size=640, use_thumbnail=False):
+    orig_width, orig_height = image.size
+    aspect_ratio = orig_width / orig_height
+
+    # calculate the existing image aspect ratio
+    target_ratios = set(
+        (i, j) for n in range(min_num, max_num + 1) for i in range(1, n + 1) for j in range(1, n + 1) if
+        i * j <= max_num and i * j >= min_num)
+    # print(target_ratios)
+    target_ratios = sorted(target_ratios, key=lambda x: x[0] * x[1])
+
+    # find the closest aspect ratio to the target
+    target_aspect_ratio = find_closest_aspect_ratio(
+        aspect_ratio, target_ratios, orig_width, orig_height, image_size)
+
+    # print(target_aspect_ratio)
+    # calculate the target width and height
+    target_width = image_size * target_aspect_ratio[0]
+    target_height = image_size * target_aspect_ratio[1]
+    blocks = target_aspect_ratio[0] * target_aspect_ratio[1]
+
+    # resize the image
+    resized_img = image.resize((target_width, target_height))
+    processed_images = []
+    for i in range(blocks):
+        box = (
+            (i % (target_width // image_size)) * image_size,
+            (i // (target_width // image_size)) * image_size,
+            ((i % (target_width // image_size)) + 1) * image_size,
+            ((i // (target_width // image_size)) + 1) * image_size
+        )
+        # split the image
+        split_img = resized_img.crop(box)
+        processed_images.append(split_img)
+    assert len(processed_images) == blocks
+    if use_thumbnail and len(processed_images) != 1:
+        thumbnail_img = image.resize((image_size, image_size))
+        processed_images.append(thumbnail_img)
+    return processed_images, target_aspect_ratio
+
+
+
+
+
+class ImageTransform:
+
+    def __init__(self,
+                 mean: Tuple[float, float, float] = (0.5, 0.5, 0.5),
+                 std: Tuple[float, float, float] = (0.5, 0.5, 0.5),
+                 normalize: bool = True):
+        self.mean = mean
+        self.std = std
+        self.normalize = normalize
+
+        transform_pipelines = [T.ToTensor()]
+
+        if normalize:
+            transform_pipelines.append(T.Normalize(mean, std))
+
+        self.transform = T.Compose(transform_pipelines)
+
+    def __call__(self, pil_img: Image.Image):
+        x = self.transform(pil_img)
+        return x
+
+
+class DeepseekOCRProcessor(ProcessorMixin):
+    tokenizer_class = ("LlamaTokenizer", "LlamaTokenizerFast")
+    attributes = ["tokenizer"]
+
+    def __init__(
+        self,
+        tokenizer: LlamaTokenizerFast = TOKENIZER,
+        candidate_resolutions: Tuple[Tuple[int, int]] = [[1024, 1024]],
+        patch_size: int = 16,
+        downsample_ratio: int = 4,
+        image_mean: Tuple[float, float, float] = (0.5, 0.5, 0.5),
+        image_std: Tuple[float, float, float] = (0.5, 0.5, 0.5),
+        normalize: bool = True,
+        image_token: str = "<image>",
+        pad_token: str = "<｜▁pad▁｜>",
+        add_special_token: bool = False,
+        sft_format: str = "deepseek",
+        mask_prompt: bool = True,
+        ignore_id: int = -100,
+        **kwargs,
+    ):
+
+        # self.candidate_resolutions = candidate_resolutions # placeholder no use
+        self.image_size = IMAGE_SIZE
+        self.base_size = BASE_SIZE
+        # self.patch_size = patch_size
+        self.patch_size = 16 
+        self.image_mean = image_mean
+        self.image_std = image_std
+        self.normalize = normalize
+        # self.downsample_ratio = downsample_ratio
+        self.downsample_ratio = 4
+
+        self.image_transform = ImageTransform(mean=image_mean, std=image_std, normalize=normalize)
+
+
+        self.tokenizer = tokenizer
+        # self.tokenizer = add_special_token(tokenizer)
+        self.tokenizer.padding_side = 'left'  # must set this，padding side with make a difference in batch inference
+
+        # add the pad_token as special token to use 'tokenizer.pad_token' and 'tokenizer.pad_token_id'
+        if self.tokenizer.pad_token is None:
+            self.tokenizer.add_special_tokens({'pad_token': pad_token})
+
+        # add image token
+        # image_token_id = self.tokenizer.vocab.get(image_token)
+        # if image_token_id is None:
+        #     special_tokens = [image_token]
+        #     special_tokens_dict = {"additional_special_tokens": special_tokens}
+        #     self.tokenizer.add_special_tokens(special_tokens_dict)
+        self.image_token_id = self.tokenizer.vocab.get(image_token)
+
+        # add five special tokens for grounding-related tasks
+        # <|ref|>, <|/ref|>, <|det|>, <|/det|>, <|grounding|>
+        # special_tokens = ['<|ref|>', '<|/ref|>', '<|det|>', '<|/det|>', '<|grounding|>']
+        # special_tokens_dict = {"additional_special_tokens": special_tokens}
+
+        # special_tokens = ['<image>','<|ref|>', '<|/ref|>', '<|det|>', '<|/det|>', '<|grounding|>', '<td>', '</td>', '<tr>', '</tr>']
+        # special_tokens_dict = {"additional_special_tokens": special_tokens}
+        # self.tokenizer.add_special_tokens(special_tokens_dict)
+
+        # # add special tokens for SFT data
+        # special_tokens = ["<|User|>", "<|Assistant|>"]
+        # special_tokens_dict = {"additional_special_tokens": special_tokens}
+        # self.tokenizer.add_special_tokens(special_tokens_dict)
+
+        self.image_token = image_token
+        self.pad_token = pad_token
+        self.add_special_token = add_special_token
+        self.sft_format = sft_format
+        self.mask_prompt = mask_prompt
+        self.ignore_id = ignore_id
+
+        super().__init__(
+            tokenizer,
+            **kwargs,
+        )
+
+
+    
+
+    # def select_best_resolution(self, image_size):
+    #     # used for cropping
+    #     original_width, original_height = image_size
+    #     best_fit = None
+    #     max_effective_resolution = 0
+    #     min_wasted_resolution = float("inf")
+
+    #     for width, height in self.candidate_resolutions:
+    #         scale = min(width / original_width, height / original_height)
+    #         downscaled_width, downscaled_height = int(
+    #             original_width * scale), int(original_height * scale)
+    #         effective_resolution = min(downscaled_width * downscaled_height,
+    #                                    original_width * original_height)
+    #         wasted_resolution = (width * height) - effective_resolution
+
+    #         if effective_resolution > max_effective_resolution or (
+    #                 effective_resolution == max_effective_resolution
+    #                 and wasted_resolution < min_wasted_resolution):
+    #             max_effective_resolution = effective_resolution
+    #             min_wasted_resolution = wasted_resolution
+    #             best_fit = (width, height)
+
+    #     return best_fit
+
+    @property
+    def bos_id(self):
+        return self.tokenizer.bos_token_id
+
+    @property
+    def eos_id(self):
+        return self.tokenizer.eos_token_id
+
+    @property
+    def pad_id(self):
+        return self.tokenizer.pad_token_id
+
+    def encode(self, text: str, bos: bool = True, eos: bool = False):
+        t = self.tokenizer.encode(text, add_special_tokens=False)
+
+        if bos:
+            t = [self.bos_id] + t
+        if eos:
+            t = t + [self.eos_id]
+
+        return t
+
+    def decode(self, t: List[int], **kwargs) -> str:
+        return self.tokenizer.decode(t, **kwargs)
+
+    def process_one(
+        self,
+        prompt: str,
+        images: List,
+        inference_mode: bool = True,
+        **kwargs,
+    ):
+        """
+
+        Args:
+            prompt (str): the formatted prompt;
+            conversations (List[Dict]): conversations with a list of messages;
+            images (List[ImageType]): the list of images;
+            inference_mode (bool): if True, then remove the last eos token;
+            system_prompt (str): the system prompt;
+            **kwargs:
+
+        Returns:
+            outputs (BaseProcessorOutput): the output of the processor,
+                - input_ids (torch.LongTensor): [N + image tokens]
+                - target_ids (torch.LongTensor): [N + image tokens]
+                - pixel_values (torch.FloatTensor): [n_patches, 3, H, W]
+                - image_id (int): the id of the image token
+                - num_image_tokens (List[int]): the number of image tokens
+        """
+
+        assert (prompt is not None and images is not None
+                ), "prompt and images must be used at the same time."
+
+        sft_format = prompt
+
+        input_ids, pixel_values, images_crop, images_seq_mask, images_spatial_crop, num_image_tokens, _ = images[0]
+
+
+        return {
+            "input_ids": input_ids,
+            "pixel_values": pixel_values,
+            "images_crop": images_crop,
+            "images_seq_mask": images_seq_mask,
+            "images_spatial_crop": images_spatial_crop,
+            "num_image_tokens": num_image_tokens,
+        }
+
+
+        # prepare = BatchFeature(
+        #     data=dict(
+        #         input_ids=input_ids,
+        #         pixel_values=pixel_values,
+        #         images_crop = images_crop,
+        #         images_seq_mask=images_seq_mask,
+        #         images_spatial_crop=images_spatial_crop,
+        #         num_image_tokens=num_image_tokens,
+        #     ),
+        #     tensor_type="pt",
+        # )
+        # return prepare
+
+    def __call__(
+        self,
+        *,
+        prompt: str,
+        images: List,
+        inference_mode: bool = True,
+        **kwargs,
+    ):
+        """
+
+        Args:
+            prompt (str): the formatted prompt;
+            images (List[ImageType]): the list of images;
+            inference_mode (bool): if True, then remove the last eos token;
+            **kwargs:
+
+        Returns:
+            outputs (BaseProcessorOutput): the output of the processor,
+                - input_ids (torch.LongTensor): [N + image tokens]
+                - images (torch.FloatTensor): [n_images, 3, H, W]
+                - image_id (int): the id of the image token
+                - num_image_tokens (List[int]): the number of image tokens
+        """
+
+        prepare = self.process_one(
+            prompt=prompt,
+            images=images,
+            inference_mode=inference_mode,
+        )
+
+        return prepare
+
+    def tokenize_with_images(
+        self,
+        # conversation: str,
+        images: List[Image.Image],
+        bos: bool = True,
+        eos: bool = True,
+        cropping: bool = True,
+    ):
+        """Tokenize text with <image> tags."""
+
+        # print(conversation)
+        conversation = PROMPT
+        assert conversation.count(self.image_token) == len(images)
+        text_splits = conversation.split(self.image_token)
+        images_list, images_crop_list, images_seq_mask, images_spatial_crop = [], [], [], []
+        image_shapes = []
+        num_image_tokens = []
+        tokenized_str = []
+        # print('image: ', len(images))
+        for text_sep, image in zip(text_splits, images):
+            """encode text_sep"""
+            tokenized_sep = self.encode(text_sep, bos=False, eos=False)
+            tokenized_str += tokenized_sep
+            images_seq_mask += [False] * len(tokenized_sep)
+
+            """select best resolution for anyres"""
+            # if cropping:
+            #     best_width, best_height = self.select_best_resolution(image.size)
+            # else:
+            #     best_width, best_height = self.image_size, self.image_size
+
+            image_shapes.append(image.size)
+
+            if image.size[0] <= 640 and image.size[1] <= 640:
+                crop_ratio = [1, 1]
+            else:
+                if cropping:
+                    # print('image-size: ', image.size)
+                    # best_width, best_height = select_best_resolution(image.size, self.candidate_resolutions)
+                    # print('image ', image.size)
+                    # print('open_size:', image.size)
+                    images_crop_raw, crop_ratio = dynamic_preprocess(image, image_size=IMAGE_SIZE)
+                    # print('crop_ratio: ', crop_ratio)
+                else:
+                    # best_width, best_height = self.image_size, self.image_size
+                    crop_ratio = [1, 1]
+            # print(image.size, (best_width, best_height)) # check the select_best_resolutions func
+
+            # print(crop_ratio)
+            """process the global view"""
+
+            # if cropping
+            if self.image_size <= 640 and not cropping:
+                # print('directly resize')
+                image = image.resize((self.image_size, self.image_size))
+
+            global_view = ImageOps.pad(image, (self.base_size, self.base_size),
+                                    color=tuple(int(x * 255) for x in self.image_transform.mean))
+            images_list.append(self.image_transform(global_view))
+
+            """record height / width crop num"""
+            # width_crop_num, height_crop_num = best_width // self.image_size, best_height // self.image_size
+            num_width_tiles, num_height_tiles = crop_ratio
+            images_spatial_crop.append([num_width_tiles, num_height_tiles])
+
+
+
+
+            if num_width_tiles > 1 or num_height_tiles > 1:
+                """process the local views"""
+                # local_view = ImageOps.pad(image, (best_width, best_height),
+                #                         color=tuple(int(x * 255) for x in self.image_transform.mean))
+                # for i in range(0, best_height, self.image_size):
+                #     for j in range(0, best_width, self.image_size):
+                #         images_crop_list.append(
+                #             self.image_transform(local_view.crop((j, i, j + self.image_size, i + self.image_size))))
+                for i in range(len(images_crop_raw)):
+                    images_crop_list.append(self.image_transform(images_crop_raw[i]))
+
+            # """process the global view"""
+            # global_view = ImageOps.pad(image, (self.image_size, self.image_size),
+            #                            color=tuple(int(x * 255) for x in self.image_transform.mean))
+            # images_list.append(self.image_transform(global_view))
+
+            # """process the local views"""
+            # local_view = ImageOps.pad(image, (best_width, best_height),
+            #                           color=tuple(int(x * 255) for x in self.image_transform.mean))
+            # for i in range(0, best_height, self.image_size):
+            #     for j in range(0, best_width, self.image_size):
+            #         images_list.append(
+            #             self.image_transform(local_view.crop((j, i, j + self.image_size, i + self.image_size))))
+
+            # """add image tokens"""
+            """add image tokens"""
+            num_queries = math.ceil((self.image_size // self.patch_size) / self.downsample_ratio)
+            num_queries_base = math.ceil((self.base_size // self.patch_size) / self.downsample_ratio)
+
+
+            tokenized_image = ([self.image_token_id] * num_queries_base + [self.image_token_id]) * num_queries_base
+            tokenized_image += [self.image_token_id]
+            if num_width_tiles > 1 or num_height_tiles > 1:
+                tokenized_image += ([self.image_token_id] * (num_queries * num_width_tiles) + [self.image_token_id]) * (
+                            num_queries * num_height_tiles)
+            tokenized_str += tokenized_image
+            images_seq_mask += [True] * len(tokenized_image)
+            num_image_tokens.append(len(tokenized_image))
+
+        """process the last text split"""
+        tokenized_sep = self.encode(text_splits[-1], bos=False, eos=False)
+        tokenized_str += tokenized_sep
+        images_seq_mask += [False] * len(tokenized_sep)
+
+        """add the bos and eos tokens"""
+        if bos:
+            tokenized_str = [self.bos_id] + tokenized_str
+            images_seq_mask = [False] + images_seq_mask
+        if eos:
+            tokenized_str = tokenized_str + [self.eos_id]
+            images_seq_mask = images_seq_mask + [False]
+
+        assert len(tokenized_str) == len(
+            images_seq_mask), f"tokenize_with_images func: tokenized_str's length {len(tokenized_str)} is not equal to imags_seq_mask's length {len(images_seq_mask)}"
+        
+
+
+        masked_tokenized_str = []
+        for token_index in tokenized_str:
+            if token_index != self.image_token_id:
+                masked_tokenized_str.append(token_index)
+            else:
+                masked_tokenized_str.append(self.ignore_id)
+
+        assert len(tokenized_str) == len(images_seq_mask) == len(masked_tokenized_str), \
+            (f"tokenized_str's length {len(tokenized_str)}, input_ids' length {len(masked_tokenized_str)}, "
+             f"imags_seq_mask's length {len(images_seq_mask)}, are not equal")
+
+        input_ids = torch.LongTensor(tokenized_str)
+        target_ids = torch.LongTensor(masked_tokenized_str)
+        images_seq_mask = torch.tensor(images_seq_mask, dtype=torch.bool)
+
+        # set input_ids < 0 | input_ids == self.image_token_id as ignore_id
+        target_ids[(input_ids < 0) |
+                   (input_ids == self.image_token_id)] = self.ignore_id
+        input_ids[input_ids < 0] = self.pad_id
+
+        inference_mode = True
+
+        if inference_mode:
+            # Remove the ending eos token
+            assert input_ids[-1] == self.eos_id
+            input_ids = input_ids[:-1]
+            target_ids = target_ids[:-1]
+            images_seq_mask = images_seq_mask[:-1]
+
+        if len(images_list) == 0:
+            pixel_values = torch.zeros((1, 3, self.base_size, self.base_size))
+            images_spatial_crop = torch.zeros((1, 1), dtype=torch.long)
+            images_crop = torch.zeros((1, 3, self.image_size, self.image_size)).unsqueeze(0)
+        else:
+            pixel_values = torch.stack(images_list, dim=0)
+            images_spatial_crop = torch.tensor(images_spatial_crop, dtype=torch.long)
+            if images_crop_list:
+                images_crop = torch.stack(images_crop_list, dim=0).unsqueeze(0)
+            else:
+                images_crop = torch.zeros((1, 3, self.image_size, self.image_size)).unsqueeze(0)
+
+        input_ids = input_ids.unsqueeze(0)
+
+        
+        return [[input_ids, pixel_values, images_crop, images_seq_mask, images_spatial_crop, num_image_tokens, image_shapes]]
+
+
+AutoProcessor.register("DeepseekVLV2Processor", DeepseekOCRProcessor)

process/ngram_norepeat.py

@@ -0,0 +1,40 @@
+import torch
+from transformers import LogitsProcessor
+from transformers.generation.logits_process import _calc_banned_ngram_tokens
+from typing import List, Set
+
+
+class NoRepeatNGramLogitsProcessor(LogitsProcessor):
+
+    def __init__(self, ngram_size: int, window_size: int = 100, whitelist_token_ids: set = None):
+        if not isinstance(ngram_size, int) or ngram_size <= 0:
+            raise ValueError(f"`ngram_size` has to be a strictly positive integer, but is {ngram_size}")
+        if not isinstance(window_size, int) or window_size <= 0:
+            raise ValueError(f"`window_size` has to be a strictly positive integer, but is {window_size}")
+        self.ngram_size = ngram_size
+        self.window_size = window_size
+        self.whitelist_token_ids = whitelist_token_ids or set()
+    
+    def __call__(self, input_ids: List[int], scores: torch.FloatTensor) -> torch.FloatTensor:
+        if len(input_ids) < self.ngram_size:
+            return scores
+        
+        current_prefix = tuple(input_ids[-(self.ngram_size - 1):])
+        
+        search_start = max(0, len(input_ids) - self.window_size)
+        search_end = len(input_ids) - self.ngram_size + 1
+        
+        banned_tokens = set()
+        for i in range(search_start, search_end):
+            ngram = tuple(input_ids[i:i + self.ngram_size])
+            if ngram[:-1] == current_prefix:
+                banned_tokens.add(ngram[-1])
+        
+        banned_tokens = banned_tokens - self.whitelist_token_ids
+        
+        if banned_tokens:
+            scores = scores.clone()
+            for token in banned_tokens:
+                scores[token] = -float("inf")
+        
+        return scores

requirements.txt

@@ -0,0 +1,11 @@
+transformers==4.46.3
+tokenizers==0.20.3
+datasets
+huggingface-hub[hf_transfer]
+pillow
+numpy
+tqdm
+einops
+addict
+easydict
+matplotlib