model_converter

PyTorch model conversion toolbox

This toolbox supports model conversion to one of the following formats:

• onnx
• keras
• tflite
• coreml

Currently, two main conversion pipelines are supported:

1. PyTorch --> ONNX --> Keras --> TFLite
2. PyTorch --> TorchScript --> CoreML

Installation

Requirements

• python 3.9

Install

It can be installed with the pip:

pip install git+ssh://[email protected]/opencv-ai/model_converter

Get started

To use converter in your project:

1. Import converter:

   import model_converter

2. Create an instance of a convertor:

   my_converter = model_converter.Converter(save_dir=<path to your output directory>, 
                                            simplify_exported_model=False
                                           )

   Use simplify_exported_model=True key to simplify onnx model.

3. Run conversion of your model:

   converted_model = my_converter.convert(
       torch_model, # model for conversion
       torch_weights, # path to model checkpoint
       batch_size, # batch size
       input_size, # input size in [height, width] format
       channels, # number of input channels
       fmt, # output format for conversion - one of 'onnx', 'keras', 'tflite', 'coreml', 'tflite_coreml'
       force # set to `True` to rebuild all intermediate steps
   ) 

Model outputs wrapping (CoreML conversion)

You can wrap the output of your PyTorch model in a NamedTuple as shown below. By doing this, the Converter will be able to assign the correct names to the output in the resulting CoreML model.

class Model(nn.Module):
    """ 
    Parameters
    ----------
    nn : [nn.Module]
        Core feature extractor model that takes as input images and outputs feature 
        vector, e.g. of dimension Bx2048x7x7 
    """
    Output = collections.namedtuple('output', ['cls',])

    def __init__(self,
                 core: nn.Module):
        super().__init__()
        self.core = core

    def forward(self, x):
        return self.Output(cls=self.core(x))

Development

Poetry

Want to know more about Poetry? Check its documentation.

Details about Poetry

Poetry's commands are very intuitive and easy to learn, like:

• poetry add numpy@latest
• poetry run pytest
• poetry publish --build

etc

Building and releasing your package

Building a new version of the application contains steps:

• Bump the version of your package poetry version <version>. You can pass the new version explicitly, or a rule such as major, minor, or patch. For more details, refer to the Semantic Versions standard.
• Make a commit to GitHub.
• Create a GitHub release.
• And... publish 🙂 poetry publish --build

🚀 Features

Development features

• Supports for Python 3.9 and higher.
• Poetry as the dependencies manager. See configuration in pyproject.toml and setup.cfg.
• Automatic codestyle with black, isort and pyupgrade.
• Ready-to-use pre-commit hooks with code-formatting.
• Type checks with mypy; docstring checks with darglint; security checks with safety and bandit
• Testing with pytest.
• Ready-to-use .editorconfig, .dockerignore, and .gitignore. You don't have to worry about those things.

Deployment features

• GitHub integration: issue and pr templates.
• Github Actions with predefined build workflow as the default CI/CD.
• Everything is already set up for security checks, codestyle checks, code formatting, testing, linting, docker builds, etc with Makefile. More details in makefile-usage.
• Dockerfile for your package.
• Always up-to-date dependencies with @dependabot. You will only enable it.
• Automatic drafts of new releases with Release Drafter. You may see the list of labels in release-drafter.yml. Works perfectly with Semantic Versions specification.

Open source community features

• Ready-to-use Pull Requests templates and several Issue templates.
• Files such as: LICENSE, CONTRIBUTING.md, CODE_OF_CONDUCT.md, and SECURITY.md are generated automatically.
• Stale bot that closes abandoned issues after a period of inactivity. (You will only need to setup free plan). Configuration is here.
• Semantic Versions specification with Release Drafter.

Installation

pip install -U model_converter

or install with Poetry

poetry add model_converter

Makefile usage

Makefile contains a lot of functions for faster development.

1. Download and remove Poetry

To download and install Poetry run:

make poetry-download

To uninstall

make poetry-remove

2. Install all dependencies and pre-commit hooks

Install requirements:

make install

Pre-commit hooks coulb be installed after git init via

make pre-commit-install

3. Codestyle

Automatic formatting uses pyupgrade, isort and black.

make codestyle

# or use synonym
make formatting

Codestyle checks only, without rewriting files:

make check-codestyle

Note: check-codestyle uses isort, black and darglint library

Update all dev libraries to the latest version using one comand

make update-dev-deps

4. Code security

make check-safety

This command launches Poetry integrity checks as well as identifies security issues with Safety and Bandit.

make check-safety

5. Type checks

Run mypy static type checker

make mypy

6. Tests with coverage badges

Run pytest

make test

7. All linters

Of course there is a command to rule run all linters in one:

make lint

the same as:

make test && make check-codestyle && make mypy && make check-safety

8. Docker

make docker-build

which is equivalent to:

make docker-build VERSION=latest

Remove docker image with

make docker-remove

More information about docker.

9. Cleanup

Delete pycache files

make pycache-remove

Remove package build

make build-remove

Delete .DS_STORE files

make dsstore-remove

Remove .mypycache

make mypycache-remove

Or to remove all above run:

make cleanup

📈 Releases

You can see the list of available releases on the GitHub Releases page.

We follow Semantic Versions specification.

We use Release Drafter. As pull requests are merged, a draft release is kept up-to-date listing the changes, ready to publish when you’re ready. With the categories option, you can categorize pull requests in release notes using labels.

List of labels and corresponding titles

Label	Title in Releases
enhancement, feature	🚀 Features
bug, refactoring, bugfix, fix	🔧 Fixes & Refactoring
build, ci, testing	📦 Build System & CI/CD
breaking	💥 Breaking Changes
documentation	📝 Documentation
dependencies	⬆️ Dependencies updates

You can update it in release-drafter.yml.

GitHub creates the bug, enhancement, and documentation labels for you. Dependabot creates the dependencies label. Create the remaining labels on the Issues tab of your GitHub repository, when you need them.

🛡 License

This project is licensed under the terms of the MIT license. See LICENSE for more details.

📃 Citation

@misc{model_converter,
  author = {OpenCV.AI},
  title = {PyTorch model conversion to different formats},
  year = {2023},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/opencv-ai/model_converter}}
}

Credits

This project was generated with python-package-template