• Sinapsis DeepSeek OCR

  • Uses the DeepSeek OCR model for high-quality OCR
  • Supports optional grounding for bounding box extraction
  • Multiple inference modes (tiny, small, base, large, gundam)

• Sinapsis DocTR

  • Uses the DocTR library for high-quality OCR with modern deep learning models
  • Supports multiple detection and recognition architectures
  • Provides detailed text extraction with bounding boxes and confidence scores

• Sinapsis EasyOCR

  • Leverages the EasyOCR library for simple yet effective OCR
  • Supports multiple languages
  • Extracts text with bounding boxes and confidence scores

• Sinapsis GLM OCR

  • Uses Zhipu AI's GLM-OCR model for high-quality OCR
  • Supports document parsing (text, formula, table) and structured information extraction via JSON schema
  • Batch inference support for faster processing of multiple images

agent:
  name: doctr_prediction
  description: agent to run inference with DocTR, performs on images read, recognition and save

templates:
  - template_name: InputTemplate
    class_name: InputTemplate
    attributes: {}

  - template_name: FolderImageDatasetCV2
    class_name: FolderImageDatasetCV2
    template_input: InputTemplate
    attributes:
      data_dir: dataset/input

  - template_name: DocTROCRPrediction
    class_name: DocTROCRPrediction
    template_input: FolderImageDatasetCV2
    attributes:
      recognized_characters_as_labels: True

  - template_name: BBoxDrawer
    class_name: BBoxDrawer
    template_input: DocTROCRPrediction
    attributes:
      draw_confidence: True
      draw_extra_labels: True

  - template_name: ImageSaver
    class_name: ImageSaver
    template_input: BBoxDrawer
    attributes:
      save_dir: output
      root_dir: dataset

agent:
  name: easyocr_inference
  description: agent to run inference with EasyOCR, performs on images read, recognition and save

templates:
  - template_name: InputTemplate
    class_name: InputTemplate
    attributes: {}

  - template_name: FolderImageDatasetCV2
    class_name: FolderImageDatasetCV2
    template_input: InputTemplate
    attributes:
      data_dir: dataset/input

  - template_name: EasyOCR
    class_name: EasyOCR
    template_input: FolderImageDatasetCV2
    attributes: {}

  - template_name: BBoxDrawer
    class_name: BBoxDrawer
    template_input: EasyOCR
    attributes:
      draw_confidence: True
      draw_extra_labels: True

  - template_name: ImageSaver
    class_name: ImageSaver
    template_input: BBoxDrawer
    attributes:
      save_dir: output
      root_dir: dataset

agent:
  name: deepseek_ocr_inference
  description: agent to run inference with DeepSeek OCR

templates:
  - template_name: InputTemplate
    class_name: InputTemplate
    attributes: {}

  - template_name: FolderImageDatasetCV2
    class_name: FolderImageDatasetCV2
    template_input: InputTemplate
    attributes:
      data_dir: dataset/input

  - template_name: DeepSeekOCRInference
    class_name: DeepSeekOCRInference
    template_input: FolderImageDatasetCV2
    attributes:
      prompt: "Convert the document to markdown."
      enable_grounding: true
      mode: base

  - template_name: BBoxDrawer
    class_name: BBoxDrawer
    template_input: DeepSeekOCRInference
    attributes:
      draw_confidence: True
      draw_extra_labels: True

  - template_name: ImageSaver
    class_name: ImageSaver
    template_input: BBoxDrawer
    attributes:
      save_dir: output
      root_dir: dataset

agent:
  name: glm_ocr_table_agent
  description: "Agent to read images, perform GLM OCR for table recognition."

templates:
  - template_name: InputTemplate
    class_name: InputTemplate
    attributes: {}

  - template_name: FolderImageDatasetCV2
    class_name: FolderImageDatasetCV2
    template_input: InputTemplate
    attributes:
      load_on_init: True
      root_dir: "."
      data_dir: "artifacts"
      pattern: "expense.jpg"

  - template_name: GLMOCRInference
    class_name: GLMOCRInference
    template_input: FolderImageDatasetCV2
    attributes:
      prompt: "Table Recognition:"
      init_args:
        pretrained_model_name_or_path: zai-org/GLM-OCR
        torch_dtype: auto
        attn_implementation: kernels-community/flash-attn2
        device_map: auto
      generation_config:
        max_new_tokens: 8192
        do_sample: false

IMPORTANT This docker image depends on the sinapsis:base image. Please refer to the official sinapsis instructions to Build with Docker.

1. Build the sinapsis-ocr image:

docker compose -f docker/compose.yaml build

2. Start the app container:

docker compose -f docker/compose_app.yaml up

3. Check the status:

docker logs -f sinapsis-ocr-app

4. The logs will display the URL to access the webapp, e.g.:

NOTE: The url can be different, check the output of logs

Running on local URL:  http://127.0.0.1:7860

5. To stop the app:

docker compose -f docker/compose_app.yaml down

To run the webapp using the uv package manager, please:

1. Create the virtual environment and sync the dependencies:

uv sync --frozen

2. Install packages:

uv pip install sinapsis-ocr[all] --extra-index-url https://pypi.sinapsis.tech

3. Run the webapp:

uv run webapps/gradio_ocr.py

4. The terminal will display the URL to access the webapp, e.g.:

Running on local URL:  http://127.0.0.1:7860

NOTE: The url can be different, check the output of the terminal

5. To stop the app press Control + C on the terminal