SymJAX: symbolic CPU/GPU/TPU programming
This work provides a tool for researchers and practitioners in machine learning to streamline development on CPUs, GPUs, and TPUs, but it is incremental as it builds on existing frameworks like JAX and Theano.
The authors introduced SymJAX, a symbolic programming framework for machine and deep learning that simplifies graph operations and supports multiple hardware platforms, offering a user experience similar to Theano with fast optimization and Lasagne-like functionalities.
SymJAX is a symbolic programming version of JAX simplifying graph input/output/updates and providing additional functionalities for general machine learning and deep learning applications. From an user perspective SymJAX provides a la Theano experience with fast graph optimization/compilation and broad hardware support, along with Lasagne-like deep learning functionalities.