Papers › Optimizing small BERTs trained for German NER

Optimizing small BERTs trained for German NER

23 Apr 2021arXiv:2104.11559archive 2025-07-28

Jochen Zöllner, Konrad Sperfeld, Christoph Wick, Roger Labahn

Currently, the most widespread neural network architecture for training language models is the so called BERT which led to improvements in various Natural Language Processing (NLP) tasks. In general, the larger the number of parameters in a BERT model, the better the results obtained in these NLP tasks. Unfortunately, the memory consumption and the training duration drastically increases with the size of these models. In this article, we investigate various training techniques of smaller BERT models: We combine different methods from other BERT variants like ALBERT, RoBERTa, and relative positional encoding. In addition, we propose two new fine-tuning modifications leading to better performance: Class-Start-End tagging and a modified form of Linear Chain Conditional Random Fields. Furthermore, we introduce Whole-Word Attention which reduces BERTs memory usage and leads to a small increase in performance compared to classical Multi-Head-Attention. We evaluate these techniques on five public German Named Entity Recognition (NER) tasks of which two are introduced by this article.

PaperPDFCode

Code

NEISSproject/NERDatasets officialmentioned in paper report
NEISSproject/tf2_neiss_nlp officialmentioned in papertf report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

NERNamed Entity RecognitionNamed Entity Recognition (NER)named-entity-recognition

Datasets

Introduced by this paper, per the archive.

ArendtSturm

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

ALBERTAdamAttentionAttention DropoutBERTDense ConnectionsDropoutLAMBLayer NormalizationLinear LayerLinear Warmup With Linear DecayMulti-Head AttentionResidual ConnectionRoBERTaSoftmaxWeight DecayWordPiece

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections