# Challenges in Migrating Imperative Deep Learning Programs to Graph Execution: An Empirical Study

arXiv: https://arxiv.org/abs/2201.09953 · HTML: https://mlsanity.com/paper/2201.09953/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 110 references that are in the map.
- First posted on arXiv 24 Jan 2022.

## Builds on (most cited first, 19 of 19)

- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems](https://mlsanity.com/paper/1512.01274/) (2015)
- [Taxonomy of Real Faults in Deep Learning Systems](https://mlsanity.com/paper/1910.11015/) (2019)
- [A Comprehensive Study on Deep Learning Bug Characteristics](https://mlsanity.com/paper/1906.01388/) (2019)
- [Software Engineering Challenges of Deep Learning](https://mlsanity.com/paper/1810.12034/) (2018)
- [Repairing Deep Neural Networks: Fix Patterns and Challenges](https://mlsanity.com/paper/2005.00972/) (2020)
- [Silent Bugs in Deep Learning Frameworks: An Empirical Study of Keras and TensorFlow](https://mlsanity.com/paper/2112.13314/) (2021)
- [TensorFlow Eager: A Multi-Stage, Python-Embedded DSL for Machine Learning](https://mlsanity.com/paper/1903.01855/) (2019)
- [An Empirical Study on Deployment Faults of Deep Learning Based Mobile Applications](https://mlsanity.com/paper/2101.04930/) (2021)
- [What Do Developers Ask About ML Libraries? A Large-scale Study Using Stack Overflow](https://mlsanity.com/paper/1906.11940/) (2019)
- [Design Smells in Deep Learning Programs: An Empirical Study](https://mlsanity.com/paper/2107.02279/) (2021)
- [AutoGraph: Imperative-style Coding with Graph-based Performance](https://mlsanity.com/paper/1810.08061/) (2018)
- [Analysis and Detection of Information Types of Open Source Software Issue Discussions](https://mlsanity.com/paper/1902.07093/) (2019)
- [Documentation of Machine Learning Software](https://mlsanity.com/paper/2001.11956/) (2020)
- [ArguLens: Anatomy of Community Opinions On Usability Issues Using Argumentation Models](https://mlsanity.com/paper/2001.06067/) (2020)
- [Mining Software Repositories with a Collaborative Heuristic Repository](https://mlsanity.com/paper/2103.01722/) (2021)
- [Proactive Empirical Assessment of New Language Feature Adoption via Automated Refactoring: The Case of Java 8 Default Methods](https://mlsanity.com/paper/1803.10198/) (2018)
- [A Deep Dive on the Impact of COVID-19 in Software Development](https://mlsanity.com/paper/2008.07048/) (2020)
- [Automated Evolution of Feature Logging Statement Levels Using Git Histories and Degree of Interest](https://mlsanity.com/paper/2104.07736/) (2021)

## Built on by (most cited first, 2 of 2)

- [Deep Learning Library Testing: Definition, Methods and Challenges](https://mlsanity.com/paper/2404.17871/) (2024)
- [Speculative Automated Refactoring of Imperative Deep Learning Programs to Graph Execution](https://mlsanity.com/paper/2504.05424/) (2025)

## Abstract (arXiv)

Efficiency is essential to support responsiveness w.r.t. ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code that supports symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development tends to produce DL code that is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, less error-prone imperative DL frameworks encouraging eager execution have emerged but at the expense of run-time performance. While hybrid approaches aim for the "best of both worlds," the challenges in applying them in the real world are largely unknown. We conduct a data-driven analysis of challenges -- and resultant bugs -- involved in writing reliable yet performant imperative DL code by studying 250 open-source projects, consisting of 19.7 MLOC, along with 470 and 446 manually examined code patches and bug reports, respectively. The results indicate that hybridization: (i) is prone to API misuse, (ii) can result in performance degradation -- the opposite of its intention, and (iii) has limited application due to execution mode incompatibility. We put forth several recommendations, best practices, and anti-patterns for effectively hybridizing imperative DL code, potentially benefiting DL practitioners, API designers, tool developers, and educators.