Uber has announced Neuropod, an open source abstraction layer that offers a uniform interface from which to run deep learning (DL) models.
Uber’s Advanced Technologies Group (ATG) leverages deep learning in the development of their self-driving technology. In a recent blog post, Vivek Panyam said, “Using deep learning, we can build and train models to handle tasks such as processing sensor input, identifying objects, and predicting where those objects might go.”
According to the project page on GitHub, “Neuropod is a library that provides a uniform interface to run deep learning models from multiple frameworks in C++ and Python.” The tool makes various DL frameworks appear the same when running a model and lets developers build models in the framework of their choice. It works by wrapping existing models in a neuropod package containing the original model as well as metadata, test data, and custom ops.
Neuropod currently supports TensorFlow, PyTorch, Keras, and TorchScript and also simplifies the process of adding support for new frameworks.
You can learn more about Neuropod on the project website.