Visualizing Dataflow Graphs of Deep Learning Models in TensorFlow
Visualizing Dataflow Graphs of Deep Learning Models in TensorFlow
复制标题
DOI：
10.1109/tvcg.2017.2744878
复制
发表时间：
2018-01-01
影响因子：
5.2
通讯作者：
Wattenberg, Martin
中科院分区：
文献类型：
--
作者：
Wongsuphasawat, Kanit;Smilkov, Daniel;Wattenberg, Martin
本文介绍了TensorFlow机器智能平台中TensorFlow图形可视化工具的设计研究。该工具通过可视化底层数据流图来帮助用户理解复杂的机器学习体系结构。该工具通过应用一系列图形转换来工作，这些图形转换使标准布局技术能够生成清晰的交互式图表。以使图表更加整洁。我们将非关键节点从布局中分离出来。为了提供概述，我们使用源代码中注释的层次结构构建聚集图。支持按需探索嵌套结构。我们执行边缘捆绑，以实现稳定且响应迅速的集群扩展。最后，我们检测和突出重复结构，以强调模型的模块化组成。为了演示可视化工具的实用性，我们描述了示例使用场景并报告了用户反馈。总体而言，用户发现可视化工具对于理解、调试和共享他们模型的结构很有用。
We present a design study of the TensorFlow Graph Visualizer, part of the TensorFlow machine intelligence platform. This tool helps users understand complex machine learning architectures by visualizing their underlying dataflow graphs. The tool works by applying a series of graph transformations that enable standard layout techniques to produce a legible interactive diagram. To declutter the graph. we decouple non-critical nodes from the layout. To provide an overview, we build a clustered graph using the hierarchical structure annotated in the source code. To support exploration of nested structure on demand. we perform edge bundling to enable stable and responsive cluster expansion. Finally, we detect and highlight repeated structures to emphasize a model's modular composition. To demonstrate the utility of the visualizer, we describe example usage scenarios and report user feedback. Overall, users find the visualizer useful for understanding, debugging, and sharing the structures of their models.