GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding

Code

• github.com/ofa-sys/ofa
  pytorch★ 2,557
• github.com/alibaba/EasyNLP
  jax★ 2,181
• github.com/jsalt18-sentence-repl/jiant
  In paperpytorch★ 1,674
• github.com/benzakenelad/BitFit
  pytorch★ 143
• github.com/ashi-ta/speechglue
  none★ 13
• github.com/smallbenchnlp/benchmark
  none★ 11
• github.com/colinzhaoust/intrinsic_fewshot_hardness
  none★ 4
• github.com/nyu-mll/GLUE-baselines
  In paperpytorch★ 0
• github.com/kainoj/run_glue
  pytorch★ 0
• github.com/nvshrao/Pytorch-GLUE
  pytorch★ 0

Abstract

For natural language understanding (NLU) technology to be maximally useful, both practically and as a scientific object of study, it must be general: it must be able to process language in a way that is not exclusively tailored to any one specific task or dataset. In pursuit of this objective, we introduce the General Language Understanding Evaluation benchmark (GLUE), a tool for evaluating and analyzing the performance of models across a diverse range of existing NLU tasks. GLUE is model-agnostic, but it incentivizes sharing knowledge across tasks because certain tasks have very limited training data. We further provide a hand-crafted diagnostic test suite that enables detailed linguistic analysis of NLU models. We evaluate baselines based on current methods for multi-task and transfer learning and find that they do not immediately give substantial improvements over the aggregate performance of training a separate model per task, indicating room for improvement in developing general and robust NLU systems.

Tasks

Benchmark Results

Reproductions