[{"_1":2,"_161":-5,"_162":-5},"loaderData",{"_3":4,"_17":18,"_25":26,"_140":141},"root",{"_5":-5,"_6":-7,"_7":8,"_9":-7,"_10":11,"_12":-7,"_13":14,"_15":16},"currentUser","primaryEmail","siteVerification","6LeIxAcTAAAAAJcZVRqyHh71UMIEGNQ_MXjiZKhI","toast","enableDev",false,"incidentStatus","theme","system","contrast","normal","routes/_layout",{"_19":-5,"_20":21,"_22":21,"_23":11,"_24":21},"username","isAskGordonEnabled",true,"isUsingNewTemplate","showMinimalLanding","showHoneypotTrap","routes/_layout.r.$namespace.$reponame",{"_27":28,"_29":-5,"_30":11,"_31":11,"_32":33,"_34":35,"_36":-5,"_37":11,"_38":39,"_88":-5,"_89":90,"_129":-5,"_130":11,"_131":21,"_132":11,"_133":-5,"_134":11,"_135":-5,"_136":137,"_138":91,"_139":-7},"canonicalUrl","https://hub.docker.com/r/jmn319/deepseek-ocr/tags","mcpServer","canSeeDockerfile","canSeeAutobuilds","namespace","jmn319","reponame","deepseek-ocr","cloudBuildSource","isAutobuildsAvailable","repo",{"_40":41,"_42":43,"_45":46,"_47":48,"_49":11,"_50":51,"_52":53,"_54":55,"_57":35,"_32":33,"_58":59,"_60":44,"_61":62,"_63":64,"_65":66,"_67":-5,"_68":62,"_69":70,"_71":11,"_72":33,"_73":74,"_79":11,"_80":81,"_85":33,"_86":87},"categories",[],"content_types",[44],"image","date_registered","2025-12-28T14:23:24.015053Z","description","A version of https://github.com/Bogdanovich77/DeekSeek-OCR---Dockerized-API","is_private","last_updated","2026-01-22T02:23:15.430343Z","last_modified","2026-01-22T02:23:30.915336Z","media_types",[56],"application/vnd.docker.container.image.v1+json","name","pull_count",3404,"repository_type","star_count",0,"status",1,"status_description","active","source","collaborator_count","full_description","# DeepSeek-OCR: PDF to Markdown Converter\n\nA powerful OCR solution that converts PDF documents to Markdown format using DeepSeek-OCR with FastAPI backend. This project provides both a batch processing script and a REST API for flexible document conversion.\n\n## 🚀 Quick Start\n\n### Option 1: Batch Processing with pdf_to_markdown_processor.py\n\n1. Place your PDF files in the `data/` directory\n2. Ensure the DeepSeek-OCR API is running (see Docker setup below)\n3. Run the processor:\n\n```bash\npython pdf_to_markdown_processor.py\n```\n\n### Option 2: REST API with Docker Backend\n\n1. Build and start the Docker container\n2. Use the API endpoints to process documents\n3. Integrate with your applications\n\n---\n\n## 📋 Prerequisites\n\n### Hardware Requirements\n- **NVIDIA GPU** with CUDA 11.8+ support\n- **GPU Memory**: Minimum 12GB VRAM (Model takes ~9GB)\n- **System RAM**: Minimum 32GB (recommended: 64GB+)\n- **Storage**: 50GB+ free space for model and containers\n\n### Software Requirements\n- **Python 3.8+** (for local processing)\n- **Docker** 20.10+ with GPU support\n- **Docker Compose** 2.0+\n- **NVIDIA Container Toolkit** installed\n- **CUDA 11.8** compatible drivers\n\n---\n\n## 🐳 Docker Backend Setup\n\n### 1. Download Model Weights\n\nCreate a directory for model weights and download the DeepSeek-OCR model:\n\n```bash\n# Create models directory\nmkdir -p models\n\n# Download using Hugging Face CLI\npip install huggingface_hub\nhuggingface-cli download deepseek-ai/DeepSeek-OCR --local-dir models/deepseek-ai/DeepSeek-OCR\n\n# Or using git\ngit clone https://huggingface.co/deepseek-ai/DeepSeek-OCR models/deepseek-ai/DeepSeek-OCR\n```\n\n### 2. Build and Run the Docker Container\n\n#### Windows Users\n\n```cmd\nREM Build the Docker image\nbuild.bat\n\nREM Start the service\ndocker-compose up -d\n\nREM Check logs\ndocker-compose logs -f deepseek-ocr\n```\n\n#### Linux/macOS Users\n\n```bash\n# Build the Docker image\ndocker-compose build\n\n# Start the service\ndocker-compose up -d\n\n# Check logs\ndocker-compose logs -f deepseek-ocr\n```\n\n### 3. Verify Installation\n\n```bash\n# Health check\ncurl http://localhost:8000/health\n\n# Expected response:\n{\n \"status\": \"healthy\",\n \"model_loaded\": true,\n \"model_path\": \"/app/models/deepseek-ai/DeepSeek-OCR\",\n \"cuda_available\": true,\n \"cuda_device_count\": 1\n}\n```\n\n---\n\n## 📄 PDF Processing Scripts\n\nThis project provides several PDF processing scripts, each designed for different use cases. All scripts scan the `data/` directory for PDF files and convert them to Markdown format with different prompts and post-processing options.\n\n### Output Naming Convention\n\nAll processors append a suffix to the output filename to indicate the processing method used:\n- **-MD.md**: Markdown conversion (preserves document structure)\n- **-OCR.md**: Plain OCR extraction (raw text without formatting)\n- **-CUSTOM.md**: Custom prompt processing (uses prompt from YAML file)\n\nFor example, processing `document.pdf` will create:\n- `document-MD.md` (markdown processors)\n- `document-OCR.md` (OCR processor)\n- `document-CUSTOM.md` (custom prompt processors)\n\n---\n\n### 1. pdf_to_markdown_processor.py\n\n**Purpose**: Basic PDF to Markdown conversion using the standard markdown prompt\n\n**Features**:\n- Uses prompt: `'\\n<|grounding|>Convert the document to markdown.'`\n- Converts PDFs to structured Markdown format\n- Simple processing without image extraction\n- Outputs files with `-MD.md` suffix\n\n**Usage**:\n```bash\n# Place PDF files in the data directory\ncp your_document.pdf data/\n\n# Run the processor\npython pdf_to_markdown_processor.py\n\n# Check results\nls data/*-MD.md\n```\n\n---\n\n### 2. pdf_to_markdown_processor_enhanced.py\n\n**Purpose**: Enhanced PDF to Markdown conversion with post-processing\n\n**Features**:\n- Uses the same markdown prompt as the basic version\n- **Post-processing features**:\n - Image extraction and saving to `data/images/` folder\n - Special token cleanup\n - Reference processing for layout information\n - Content cleaning and formatting\n- Outputs files with `-MD.md` suffix\n\n**Usage**:\n```bash\n# Place PDF files in the data directory\ncp your_document.pdf data/\n\n# Run the enhanced processor\npython pdf_to_markdown_processor_enhanced.py\n\n# Check results (including extracted images)\nls data/*-MD.md\nls data/images/\n```\n\n---\n\n### 3. pdf_to_ocr_enhanced.py\n\n**Purpose**: Plain OCR text extraction without markdown formatting\n\n**Features**:\n- Uses OCR prompt: `'\\nFree OCR.'`\n- Extracts raw text without markdown structure\n- Includes the same post-processing features as the enhanced markdown processor\n- Outputs files with `-OCR.md` suffix\n\n**Usage**:\n```bash\n# Place PDF files in the data directory\ncp your_document.pdf data/\n\n# Run the OCR processor\npython pdf_to_ocr_enhanced.py\n\n# Check results\nls data/*-OCR.md\n```\n\n---\n\n### 4. pdf_to_custom_prompt.py\n\n**Purpose**: PDF processing with custom prompts (raw output)\n\n**Features**:\n- Uses custom prompt loaded from `custom_prompt.yaml`\n- Returns raw model response without post-processing\n- Ideal for testing and debugging different prompts\n- Outputs files with `-CUSTOM.md` suffix\n\n**Configuration**:\nEdit `custom_prompt.yaml` to customize the prompt:\n```yaml\n# Custom prompt for PDF processing\nprompt: '\\n<|grounding|>Convert the document to markdown.'\n```\n\n**Usage**:\n```bash\n# Edit the prompt in custom_prompt.yaml\nnano custom_prompt.yaml\n\n# Place PDF files in the data directory\ncp your_document.pdf data/\n\n# Run the custom prompt processor\npython pdf_to_custom_prompt.py\n\n# Check results\nls data/*-CUSTOM.md\n```\n\n---\n\n### 5. pdf_to_custom_prompt_enhanced.py\n\n**Purpose**: PDF processing with custom prompts and full post-processing\n\n**Features**:\n- Uses custom prompt loaded from `custom_prompt.yaml`\n- Includes all post-processing features (image extraction, content cleaning, etc.)\n- Combines custom prompts with enhanced output processing\n- Outputs files with `-CUSTOM.md` suffix\n\n**Configuration**:\nSame as `pdf_to_custom_prompt.py` - edit `custom_prompt.yaml` to customize the prompt.\n\n**Usage**:\n```bash\n# Edit the prompt in custom_prompt.yaml\nnano custom_prompt.yaml\n\n# Place PDF files in the data directory\ncp your_document.pdf data/\n\n# Run the enhanced custom prompt processor\npython pdf_to_custom_prompt_enhanced.py\n\n# Check results (including extracted images)\nls data/*-CUSTOM.md\nls data/images/\n```\n\n---\n\n## 📊 Comparison of Processors\n\n| Processor | Prompt | Post-Processing | Image Extraction | Output Suffix | Use Case |\n|-----------|--------|-----------------|------------------|---------------|----------|\n| `pdf_to_markdown_processor.py` | Markdown | ❌ | ❌ | `-MD.md` | Quick markdown conversion |\n| `pdf_to_markdown_processor_enhanced.py` | Markdown | ✅ | ✅ | `-MD.md` | Full-featured markdown with images |\n| `pdf_to_ocr_enhanced.py` | Free OCR | ✅ | ✅ | `-OCR.md` | Raw text extraction |\n| `pdf_to_custom_prompt.py` | Custom (YAML) | ❌ | ❌ | `-CUSTOM.md` | Testing custom prompts |\n| `pdf_to_custom_prompt_enhanced.py` | Custom (YAML) | ✅ | ✅ | `-CUSTOM.md` | Custom prompts with full features |\n\n---\n\n## 📋 Common Usage Patterns\n\n### Comparing Different Processing Methods\n\nTo compare how different processors handle the same document:\n\n```bash\n# Place a PDF in the data directory\ncp test_document.pdf data/\n\n# Run all processors\npython pdf_to_markdown_processor.py\npython pdf_to_markdown_processor_enhanced.py\npython pdf_to_ocr_enhanced.py\npython pdf_to_custom_prompt.py\n\n# Compare outputs\nls data/test_document-*.md\n```\n\n### Processing with Custom Prompts\n\n1. Edit `custom_prompt.yaml` with your desired prompt:\n```yaml\nprompt: '\\nExtract all tables and format as CSV.'\n```\n\n2. Run the custom processor:\n```bash\npython pdf_to_custom_prompt_enhanced.py\n```\n\n3. Check the specialized output:\n```bash\ncat data/your_document-CUSTOM.md\n```\n\n---\n\n---\n\n## 🔌 REST API Usage\n\nThe FastAPI backend provides several endpoints for document processing.\n\n### API Endpoints\n\n#### Health Check\n```bash\nGET http://localhost:8000/health\n```\n\n#### Process Single Image\n```bash\ncurl -X POST \"http://localhost:8000/ocr/image\" \\\n -H \"accept: application/json\" \\\n -H \"Content-Type: multipart/form-data\" \\\n -F \"file=@your_image.jpg\"\n```\n\n#### Process PDF\n```bash\ncurl -X POST \"http://localhost:8000/ocr/pdf\" \\\n -H \"accept: application/json\" \\\n -H \"Content-Type: multipart/form-data\" \\\n -F \"file=@your_document.pdf\"\n```\n\n#### Batch Processing\n```bash\ncurl -X POST \"http://localhost:8000/ocr/batch\" \\\n -H \"accept: application/json\" \\\n -H \"Content-Type: multipart/form-data\" \\\n -F \"files=@image1.jpg\" \\\n -F \"files=@document.pdf\" \\\n -F \"files=@image2.png\"\n```\n\n### Response Formats\n\n#### Single Image Response\n```json\n{\n \"success\": true,\n \"result\": \"# Document Title\\n\\nThis is the OCR result in markdown format...\",\n \"page_count\": 1\n}\n```\n\n#### PDF Response\n```json\n{\n \"success\": true,\n \"results\": [\n {\n \"success\": true,\n \"result\": \"# Page 1 Content\\n...\",\n \"page_count\": 1\n },\n {\n \"success\": true,\n \"result\": \"# Page 2 Content\\n...\",\n \"page_count\": 2\n }\n ],\n \"total_pages\": 2,\n \"filename\": \"document.pdf\"\n}\n```\n\n---\n\n## 💻 Client Integration Examples\n\n### Python Client\n\n```python\nimport requests\n\nclass DeepSeekOCRClient:\n def __init__(self, base_url=\"http://localhost:8000\"):\n self.base_url = base_url\n \n def process_image(self, image_path):\n with open(image_path, 'rb') as f:\n response = requests.post(\n f\"{self.base_url}/ocr/image\",\n files={\"file\": f}\n )\n return response.json()\n \n def process_pdf(self, pdf_path):\n with open(pdf_path, 'rb') as f:\n response = requests.post(\n f\"{self.base_url}/ocr/pdf\",\n files={\"file\": f}\n )\n return response.json()\n\n# Usage\nclient = DeepSeekOCRClient()\nresult = client.process_pdf(\"document.pdf\")\n\nif result[\"success\"]:\n for page_result in result[\"results\"]:\n print(f\"Page {page_result['page_count']}:\")\n print(page_result[\"result\"])\n print(\"---\")\n```\n\n### JavaScript Client\n\n```javascript\nclass DeepSeekOCR {\n constructor(baseUrl = 'http://localhost:8000') {\n this.baseUrl = baseUrl;\n }\n \n async processImage(file) {\n const formData = new FormData();\n formData.append('file', file);\n \n const response = await fetch(`${this.baseUrl}/ocr/image`, {\n method: 'POST',\n body: formData\n });\n \n return await response.json();\n }\n \n async processPDF(file) {\n const formData = new FormData();\n formData.append('file', file);\n \n const response = await fetch(`${this.baseUrl}/ocr/pdf`, {\n method: 'POST',\n body: formData\n });\n \n return await response.json();\n }\n}\n\n// Usage in browser\nconst ocr = new DeepSeekOCR();\ndocument.getElementById('fileInput').addEventListener('change', async (e) => {\n const file = e.target.files[0];\n const result = await ocr.processPDF(file);\n \n if (result.success) {\n result.results.forEach(page => {\n console.log(`Page ${page.page_count}:`, page.result);\n });\n }\n});\n```\n\n---\n\n## ⚙️ Configuration\n\n### Custom Configuration and Critical Fixes\n\nThis project includes custom files that replace the original DeepSeek-OCR library code to fix critical issues and provide enhanced functionality. These replacements are applied transparently during the Docker build process.\n\n#### 🚨 Critical Prompt Parameter Fix\n\n**Issue**: The original DeepSeek-OCR library has a bug where the `tokenize_with_images()` method is called without the required `prompt` parameter during model initialization, causing server startup failures.\n\n**Solution**: Custom run scripts have been created to properly handle the prompt parameter and prevent startup errors.\n\n#### Custom Files and Their Purpose\n\nThe following custom files in the project root replace their counterparts during Docker build:\n\n- **`custom_config.py`**: Custom configuration with customizable default prompt and settings\n- **`custom_image_process.py`**: Fixed version of the image processing module that handles the prompt parameter correctly\n- **`custom_run_dpsk_ocr_pdf.py`**: Enhanced PDF script that accepts `--prompt` argument and fixes the initialization issue\n- **`custom_run_dpsk_ocr_image.py`**: Enhanced image script that accepts `--prompt` argument and fixes the initialization issue\n- **`custom_run_dpsk_ocr_eval_batch.py`**: Enhanced batch script that accepts `--prompt` argument and fixes the initialization issue\n\nThese custom files are automatically copied over the original library files during the Docker build process, ensuring the fixes are applied without requiring manual intervention.\n\n#### Using Custom Configuration\n\n1. **Edit the Default Prompt**:\n ```python\n # Edit custom_config.py\n PROMPT = '\\n<|grounding|>Your custom default prompt here.'\n ```\n\n2. **Use Custom Prompts with Direct Scripts**:\n ```bash\n # Using default prompt from custom_config.py\n python custom_run_dpsk_ocr_pdf.py --input your_file.pdf --output output_dir\n \n # Using custom prompt via command line\n python custom_run_dpsk_ocr_pdf.py --prompt \"\\n<|grounding|>Extract tables as CSV.\" --input your_file.pdf\n ```\n\n3. **Use Custom Prompts with API**:\n ```bash\n # Using default prompt\n curl -X POST \"http://localhost:8000/ocr/pdf\" -F \"file=@your_file.pdf\"\n \n # Using custom prompt\n curl -X POST \"http://localhost:8000/ocr/pdf\" -F \"file=@your_file.pdf\" -F \"prompt=\\n<|grounding|>Your custom prompt here.\"\n ```\n\n4. **Build and Run**:\n ```bash\n # Rebuild with custom configuration and fixes\n docker-compose build\n \n # Start the container\n docker-compose up -d\n ```\n\n#### Docker Build Process\n\nThe Dockerfile automatically applies the custom files during the build process:\n\n```dockerfile\n# Copy custom files to replace the originals (transparent replacement approach)\nCOPY custom_config.py ./DeepSeek-OCR-vllm/config.py\nCOPY custom_image_process.py ./DeepSeek-OCR-vllm/process/image_process.py\n\n# Copy custom run scripts to replace the originals\nCOPY custom_run_dpsk_ocr_pdf.py ./DeepSeek-OCR-vllm/run_dpsk_ocr_pdf.py\nCOPY custom_run_dpsk_ocr_image.py ./DeepSeek-OCR-vllm/run_dpsk_ocr_image.py\nCOPY custom_run_dpsk_ocr_eval_batch.py ./DeepSeek-OCR-vllm/run_dpsk_ocr_eval_batch.py\n```\n\nThis transparent replacement approach ensures that:\n- The critical prompt parameter fix is applied\n- Custom configuration options are available\n- No manual modification of the original library code is required\n- The fixes persist across container rebuilds\n\nFor detailed documentation, see **`CUSTOM_CONFIG_README.md`**.\n\n### Environment Variables\n\nEdit `docker-compose.yml` to adjust these settings:\n\n```yaml\nenvironment:\n - CUDA_VISIBLE_DEVICES=0 # GPU device to use\n - MODEL_PATH=/app/models/deepseek-ai/DeepSeek-OCR # Model path\n - MAX_CONCURRENCY=50 # Max concurrent requests\n - GPU_MEMORY_UTILIZATION=0.85 # GPU memory usage (0.1-1.0)\n```\n\n### Performance Tuning\n\n#### For High-Throughput Processing\n```yaml\nenvironment:\n - MAX_CONCURRENCY=100\n - GPU_MEMORY_UTILIZATION=0.95\n```\n\n#### For Memory-Constrained Systems\n```yaml\nenvironment:\n - MAX_CONCURRENCY=10\n - GPU_MEMORY_UTILIZATION=0.7\n```\n\n---\n\n## 🔧 Troubleshooting\n\n### Common Issues\n\n#### 1. Out of Memory Errors\n```bash\n# Reduce concurrency and GPU memory usage\n# Edit docker-compose.yml:\nenvironment:\n - MAX_CONCURRENCY=10\n - GPU_MEMORY_UTILIZATION=0.7\n```\n\n#### 2. Model Loading Issues\n```bash\n# Check model directory structure\nls -la models/deepseek-ai/DeepSeek-OCR/\n\n# Verify model files are present\ndocker-compose exec deepseek-ocr ls -la /app/models/deepseek-ai/DeepSeek-OCR/\n```\n\n#### 3. CUDA Errors\n```bash\n# Check GPU availability\nnvidia-smi\n\n# Check Docker GPU support\ndocker run --rm --gpus all nvidia/cuda:11.8-base-ubuntu20.04 nvidia-smi\n```\n\n#### 4. API Connection Errors\n```bash\n# Check if the API is running\ncurl http://localhost:8000/health\n\n# Check container logs\ndocker-compose logs -f deepseek-ocr\n\n# Restart the service\ndocker-compose restart deepseek-ocr\n```\n\n#### 5. PDF Processing Errors\n```bash\n# Check if PDF files are valid\nfile data/your_document.pdf\n\n# Try processing a single PDF manually\ncurl -X POST \"http://localhost:8000/ocr/pdf\" \\\n -H \"accept: application/json\" \\\n -H \"Content-Type: multipart/form-data\" \\\n -F \"file=@data/your_document.pdf\"\n```\n\n#### 6. Prompt Parameter Error (Fixed)\nIf you encounter this error during server startup:\n```\nTypeError: DeepseekOCRProcessor.tokenize_with_images() missing 1 required positional argument: 'prompt'\n```\n\nThis error has been fixed with the custom files included in the Docker build. If you still see it:\n\n1. **Ensure you're using the updated Dockerfile** (includes custom run scripts)\n2. **Rebuild the container completely**:\n ```bash\n docker-compose down\n docker-compose build --no-cache\n docker-compose up -d\n ```\n3. **Verify the fix is applied**:\n ```bash\n docker-compose exec deepseek-ocr ls -la /app/DeepSeek-OCR-vllm/run_dpsk_ocr_*.py\n # These should show recent timestamps from the build\n ```\n\nThe fix ensures that the `tokenize_with_images()` method is called with the correct prompt parameter during model initialization.\n\n### Debug Mode\n\nFor debugging, you can run the container with additional tools:\n\n```bash\n# Run with shell access\ndocker-compose run --rm deepseek-ocr bash\n\n# Check model loading\npython -c \"\nimport sys\nsys.path.insert(0, '/app/DeepSeek-OCR-master/DeepSeek-OCR-vllm')\nfrom config import MODEL_PATH\nprint(f'Model path: {MODEL_PATH}')\nprint(f'Model exists: {os.path.exists(MODEL_PATH)}')\n\"\n```\n\n---\n\n## 📊 Performance Tips\n\n1. **Batch Processing**: Process multiple files at once using the `/ocr/batch` endpoint\n2. **Optimize DPI**: The default DPI of 144 provides good balance between quality and speed\n3. **GPU Utilization**: Adjust `GPU_MEMORY_UTILIZATION` based on your GPU capacity\n4. **Concurrency**: Increase `MAX_CONCURRENCY` for better throughput on powerful GPUs\n5. **File Size**: For large PDFs, consider splitting them into smaller chunks\n\n---\n\n## 🏗️ Project Structure\n\n```\nDeepSeek-OCR/\n├── README.md # This file\n├── CUSTOM_CONFIG_README.md # Custom configuration documentation\n├── pdf_to_markdown_processor.py # Basic markdown conversion\n├── pdf_to_markdown_processor_enhanced.py # Enhanced markdown with post-processing\n├── pdf_to_ocr_enhanced.py # OCR text extraction\n├── pdf_to_custom_prompt.py # Custom prompt processing (raw)\n├── pdf_to_custom_prompt_enhanced.py # Custom prompt with post-processing\n├── custom_prompt.yaml # Configuration for custom prompts\n├── custom_config.py # Custom configuration (replaces original config.py)\n├── custom_image_process.py # Fixed image processing (replaces original)\n├── custom_run_dpsk_ocr_pdf.py # Custom PDF script with prompt support (replaces original)\n├── custom_run_dpsk_ocr_image.py # Custom image script with prompt support (replaces original)\n├── custom_run_dpsk_ocr_eval_batch.py # Custom batch script with prompt support (replaces original)\n├── test_custom_config.py # Test script for custom configuration\n├── start_server.py # FastAPI server\n├── Dockerfile # Docker container definition (includes custom files)\n├── docker-compose.yml # Docker compose configuration\n├── build.bat # Windows build script\n├── data/ # Input/output directory for PDFs\n│ ├── images/ # Extracted images (when using enhanced processors)\n│ └── *.md # Generated markdown files\n├── models/ # Model weights directory\n└── DeepSeek-OCR/ # DeepSeek-OCR source code\n └── DeepSeek-OCR-master/\n └── DeepSeek-OCR-vllm/ # Original library files (replaced during build)\n```\n\n---\n\n## 📝 License\n\nThis project follows the same license as the DeepSeek-OCR project. Please refer to the original project's license file for details.\n\n---\n\n## 🤝 Support\n\nFor issues related to:\n- **Docker setup**: Check this README first\n- **DeepSeek-OCR model**: Refer to the [official repository](https://github.com/deepseek-ai/DeepSeek-OCR)\n- **vLLM**: Refer to [vLLM documentation](https://docs.vllm.ai/)\n\n---\n\n## 🔄 Usage Workflow\n\n```mermaid\ngraph TD\n A[Start] --> B{Choose Method}\n \n B -->|Batch Processing| C[Place PDFs in data/ folder]\n B -->|API Usage| D[Start Docker Container]\n \n C --> E[Run python pdf_to_markdown_processor.py]\n D --> F[Use API endpoints]\n \n E --> G[Check data/ folder for .md files]\n F --> H[Process results from API response]\n \n G --> I[Done]\n H --> I\n \n style A fill:#e1f5fe\n style I fill:#e8f5e8\n style C fill:#fff3e0\n style D fill:#fff3e0\n style E fill:#f3e5f5\n style F fill:#f3e5f5\n","has_starred","hub_user","immutable_tags_settings",{"_75":11,"_76":77},"enabled","rules",[78],".*","is_automated","permissions",{"_82":21,"_83":11,"_84":11},"read","write","admin","user","storage_size",10677966516,"orgData","recentTags",[91],{"_92":44,"_93":94,"_95":96,"_97":98,"_99":100,"_101":102,"_50":118,"_119":94,"_120":33,"_121":56,"_57":122,"_123":124,"_125":109,"_126":118,"_127":66,"_128":21},"content_type","creator",1100834,"digest","sha256:e915154eb2e3378dc3b82d47563fbb461df604b91eda8f4a8d6988e6e4833083","full_size",9277477619,"id",1038945854,"images",[103],{"_104":105,"_95":96,"_106":107,"_108":109,"_110":111,"_112":113,"_114":107,"_115":-5,"_116":98,"_63":66,"_117":-5},"architecture","amd64","features","","last_pulled","2026-10-03T00:52:21.360027877Z","last_pushed","2025-12-29T02:02:42.479902874Z","os","linux","os_features","os_version","size","variant","2026-01-22T02:23:14.267186Z","last_updater","last_updater_username","media_type","latest","repository",29598358,"tag_last_pulled","tag_last_pushed","tag_status","v2","sbxKitMetadata","isPotentialSpam","isHubSbxKitsEnabled","showSbxLogoOverride","sbxTrademarkText","displaySecurityMetadata","tagAttestations","tagsHaveAttestations",[],"selectedTag","attestationsUrl","routes/_layout.r.$namespace.$reponame.tags",{"_38":142,"_149":150,"_136":159,"_160":11},{"_40":143,"_42":144,"_45":46,"_47":48,"_49":11,"_50":51,"_52":53,"_54":145,"_57":35,"_32":33,"_58":59,"_60":44,"_61":62,"_63":64,"_65":66,"_67":-5,"_68":62,"_69":70,"_71":11,"_72":33,"_73":146,"_79":11,"_80":148,"_85":33,"_86":87},[],[44],[56],{"_75":11,"_76":147},[78],{"_82":21,"_83":11,"_84":11},"tags",{"_151":64,"_152":-5,"_153":-5,"_154":155},"count","next","previous","results",[156],{"_92":44,"_93":94,"_95":96,"_97":98,"_99":100,"_101":157,"_50":118,"_119":94,"_120":33,"_121":56,"_57":122,"_123":124,"_125":109,"_126":118,"_127":66,"_128":21},[158],{"_104":105,"_95":96,"_106":107,"_108":109,"_110":111,"_112":113,"_114":107,"_115":-5,"_116":98,"_63":66,"_117":-5},[],"canDeleteTags","actionData","errors"]

jmn319/deepseek-ocr

Tags for jmn319/deepseek-ocr