Intermediate · Systems
Graph execution
Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.
TL;DR. A mode of operation in TensorFlow where a computation graph is first defined and then executed, enabling optimizations and portability.
Technical Definition
A mode of operation in TensorFlow where a computation graph is first defined and then executed, enabling optimizations and portability.
How it works
Graph execution, a core concept in frameworks like TensorFlow 1.x, involves defining a computation graph that represents the workflow of operations. This graph is then compiled and executed, often on specific devices. This approach allows for various optimizations, such as parallel execution and efficient memory management, before the actual computations begin. It contrasts with eager execution, where operations are performed immediately as they are called.
Related Concepts
- Deep Learning — A subset of machine learning using neural networks with many layers to learn hierarchical representations from large datasets.
- Eager execution — A TensorFlow execution mode where operations are performed immediately as they are called.