ICML 2019 Expo Talk

July 12, 2020

Expo 2019 Schedule »

Texar: An open-source modularized, versatile, and extensible toolkit for Text Generation and Beyond

Sponsor: Petuum Inc.

Organizers:
Zhiting Hu (Petuum Inc.)

Presenters:
Zhiting Hu (Petuum Inc.)

https://texar.io/index.html
Abstract:

We are excited to present Texar, an open-source, general-purpose toolkit that supports a broad set of machine learning (ML) applications with a focus on text generation tasks. Texar is an open-source toolkit based on Tensorflow, aiming to support a broad set of ML especially text generation tasks, such as machine translation, dialog, summarization, content manipulation, language modeling, and so on. It extracts common patterns underlying the diverse tasks and methodologies within text generation and creates a library of highly reusable modules and functionalities, and facilitates arbitrary model architectures and algorithmic paradigms.

For example:

  • Encoder(s) to decoder(s), sequential- and self-attentions, memory, hierarchical models, classifiers...
  • Maximum likelihood learning, reinforcement learning, adversarial learning, probabilistic modeling, ...

With Texar, cutting-edge complex models can be easily constructed, freely enriched with best modeling/training practices, readily fitted into standard training/evaluation pipelines, rapidly made ready for experimentation, and updated. (e.g., plugging-in and swapping-out different modules).

Texar emphasizes well-structured, highly-readable code with uniform design patterns and consistent styles, along with clean documentation and rich tutorial examples. Texar is currently supporting several research and engineering projects at Petuum, Inc.