# AutoGraph: Imperative-style Coding with Graph-based Performance

arXiv: https://arxiv.org/abs/1810.08061 · HTML: https://mlsanity.com/paper/1810.08061/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 Oct 2018.

## Built on by (most cited first, 9 of 9)

- [tf.data: A Machine Learning Data Processing Framework](https://mlsanity.com/paper/2101.12127/) (2021)
- [Torch.fx: Practical Program Capture and Transformation for Deep Learning in Python](https://mlsanity.com/paper/2112.08429/) (2021)
- [LazyTensor: combining eager execution with domain-specific compilers](https://mlsanity.com/paper/2102.13267/) (2021)
- [Challenges in Migrating Imperative Deep Learning Programs to Graph Execution: An Empirical Study](https://mlsanity.com/paper/2201.09953/) (2022)
- [Using Python for Model Inference in Deep Learning](https://mlsanity.com/paper/2104.00254/) (2021)
- [Terra: Imperative-Symbolic Co-Execution of Imperative Deep Learning Programs](https://mlsanity.com/paper/2201.09210/) (2022)
- [GraphMend: Code Transformations for Fixing Graph Breaks in PyTorch 2](https://mlsanity.com/paper/2509.16248/) (2025)
- [stratum: A System Infrastructure for Massive Agent-Centric ML Workloads](https://mlsanity.com/paper/2603.03589/) (2026)
- [Speculative Automated Refactoring of Imperative Deep Learning Programs to Graph Execution](https://mlsanity.com/paper/2504.05424/) (2025)

## Abstract (arXiv)

There is a perceived trade-off between machine learning code that is easy to write, and machine learning code that is scalable or fast to execute. In machine learning, imperative style libraries like Autograd and PyTorch are easy to write, but suffer from high interpretive overhead and are not easily deployable in production or mobile settings. Graph-based libraries like TensorFlow and Theano benefit from whole-program optimization and can be deployed broadly, but make expressing complex models more cumbersome. We describe how the use of staged programming in Python, via source code transformation, offers a midpoint between these two library design patterns, capturing the benefits of both. A key insight is to delay all type-dependent decisions until runtime, via dynamic dispatch. We instantiate these principles in AutoGraph, a software system that improves the programming experience of the TensorFlow library, and demonstrate usability improvements with no loss in performance compared to native TensorFlow graphs. We also show that our system is backend agnostic, and demonstrate targeting an alternate IR with characteristics not found in TensorFlow graphs.