| hide | |
|---|---|
|
The HTR (Handwritten Text Recognition) Pipeline is a package that takes an image as input, runs preprocessing algorithms, feeds it into an inference engine, and outputs the results in either XML or JSON format. The inference engine can use models from a variety of third-party libraries such as OpenMMLab, Hugging Face, or others.
Installation can be done using pypi:
pip install htr
- Step-by-Step Pipeline: The HTR pipeline can be executed step-by-step, allowing you to visualize and analyze the results at each stage.
- Model Agnostic: The inference engine is framework-agnostic and can use any model that conforms to a specified interface.
- Postprocessing Steps: Each inference step includes postprocessing steps to enhance the model outputs.
- Visualization: Each step can be visualized for easier analysis and debugging.
- Easy Model Swapping: You can easily swap out the models at any stage of the pipeline to experiment and compare results.
Provide instructions for how to install your package. This might involve running pip install, cloning a GitHub repository, etc.
Provide instructions for how to use your package. This might involve initializing the HTREngine, loading models, running the pipeline, etc.
We welcome contributions! Please see CONTRIBUTING.md for details on how to contribute.
Provide instructions for how to run the test suite.
Optionally provide an overview of the code structure and key classes.
This project is licensed under the MIT License.
If you have any questions, please contact Your Contact Information.