AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

DeLighT: Deep and Light-weight Transformer

Sachin Mehta, Marjan Ghazvininejad, Srinivasan Iyer, Luke Zettlemoyer, Hannaneh Hajishirzi

ICLR 2021 · Architectures and efficiency › Attention and transformers · approach: Transformer block and depth design · 46 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Replace the Transformer with DeLighTMachine translation models · WMT'14 En-Fr testBLEU: 39.2 -> 40.5 (+1.3)Improved
Replace the Transformer with DeLighTMachine translation models · WMT'16 En-Ro testBLEU: 34.3 -> 34.7 (+0.4)Improved
Replace the Transformer with DeLighTMachine translation models · WMT'14 En-De testBLEU: 27.7 -> 27.6 (-0.1)No clear effect
Replace Transformer-XL with DeLighTLanguage modeling · WikiText-103 testperplexity: 24.03 -> 24.14No clear effect
Use block-wise depth scaling instead of uniform scaling (ablation)DeLighT · WikiText-103 validationperplexity: 48.58 -> 46.30Improved
Increase width multiplier w_m from 2 to 4 with uniform depth (ablation)DeLighT · WikiText-103 validationperplexity: 48.58 -> 45.10Improved
Increase block-wise depth range from N_min=4, N_max=8 to N_min=4, N_max=12 (ablation)DeLighT · WikiText-103 validationperplexity: 46.30 -> 41.18Improved
Increase width multiplier w_m from 2 to 3 (ablation)DeLighT · WikiText-103 validationperplexity: 41.18 -> 39.92Improved
Increase model dimension d_m from 256 to 384 (ablation)DeLighT · WikiText-103 validationperplexity: 41.18 -> 35.14Improved
Reverse block scaling, placing deeper and wider blocks near input (ablation)DeLighT · WikiText-103 validationperplexity: 41.18 -> 43.10Hurt
Move DeLighT transformation after attention (ablation)DeLighT block · WikiText-103 validationperplexity: 34.20 -> 39.02Hurt
Move DeLighT transformation after attention and retain light-weight FFN (ablation)DeLighT block · WikiText-103 validationperplexity: 34.20 -> 39.43Hurt

Later papers that compared with this method

2 later accepted papers report Delight as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy