An universal deep learning models conversor

The general idea of this project is to convert any deep learning framework model to an other. Each deep learning framework has its ows structure representation defined by its layers, parameters and syntax. Thsi structure representation makes difficult to swap from one framework to another. This project tries to solve this problem in a straightforward manner, making it simple for the non-experienced users. This project fills this gap using an inner representation which be the bridge between all deep learning frameworks.

General Usage

You can convert from one framework to another

import RosettaStone

if __name__ == '__main__':
    rosetta = RosettaStone()
    rosetta.convert('my.caffemodel', 'my.mat', 'CaffeImporter', 'MatConvnetExporter')
    
    print 'All went OK!'

Supported formats

• Importers

  • CaffeImporter
  • MatConvnetImporter
  • LasagneImporter
  • TorchImporter
  • TensorflowImporter
  • ChainerImporter
  • TinyCNNImporter

• Exporters

  • CaffeExporter
  • MatConvnetExporter
  • LasagneExporter
  • TorcExporter
  • TensorflowExporter
  • ChainerExporter
  • TinyCNNExporter

Contributing

Developers are needed! Check our Contribution Documents.

Licence

the BSD 3-Clause Licence