# On Multilingual Encoder Language Model Compression for Low-Resource Languages

arXiv: https://arxiv.org/abs/2505.16956 · HTML: https://mlsanity.com/paper/2505.16956/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Knowledge distillation for LLMs, one of 12 directions in Knowledge distillation.
- Builds on 26 of its 43 references that are in the map.
- First posted on arXiv 22 May 2025.

## Placed in

- Knowledge distillation / [Knowledge distillation for LLMs](https://mlsanity.com/distill/knowledge-distillation-for-llms/)

## Builds on (most cited first, 26 of 26)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [Unsupervised Cross-lingual Representation Learning at Scale](https://mlsanity.com/paper/1911.02116/) (2019)
- [No Language Left Behind: Scaling Human-Centered Machine Translation](https://mlsanity.com/paper/2207.04672/) (2022)
- [The State and Fate of Linguistic Diversity and Inclusion in the NLP World](https://mlsanity.com/paper/2004.09095/) (2020)
- [XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization](https://mlsanity.com/paper/2003.11080/) (2020)
- [CCNet: Extracting High Quality Monolingual Datasets from Web Crawl Data](https://mlsanity.com/paper/1911.00359/) (2019)
- [Patient Knowledge Distillation for BERT Model Compression](https://mlsanity.com/paper/1908.09355/) (2019)
- [MAD-X: An Adapter-Based Framework for Multi-Task Cross-Lingual Transfer](https://mlsanity.com/paper/2005.00052/) (2020)
- [DynaBERT: Dynamic BERT with Adaptive Width and Depth](https://mlsanity.com/paper/2004.04037/) (2020)
- [SIB-200: A Simple, Inclusive, and Big Evaluation Dataset for Topic Classification in 200+ Languages and Dialects](https://mlsanity.com/paper/2309.07445/) (2023)
- [Comparing Kullback-Leibler Divergence and Mean Squared Error Loss in Knowledge Distillation](https://mlsanity.com/paper/2105.08919/) (2021)
- [Glot500: Scaling Multilingual Corpora and Language Models to 500 Languages](https://mlsanity.com/paper/2305.12182/) (2023)
- [AfriSenti: A Twitter Sentiment Analysis Benchmark for African Languages](https://mlsanity.com/paper/2302.08956/) (2023)
- [BUFFET: Benchmarking Large Language Models for Few-shot Cross-lingual Transfer](https://mlsanity.com/paper/2305.14857/) (2023)
- [SemEval-2023 Task 12: Sentiment Analysis for African Languages (AfriSenti-SemEval)](https://mlsanity.com/paper/2304.06845/) (2023)
- [Load What You Need: Smaller Versions of Multilingual BERT](https://mlsanity.com/paper/2010.05609/) (2020)
- [FastFormers: Highly Efficient Transformer Models for Natural Language Understanding](https://mlsanity.com/paper/2010.13382/) (2020)
- [MasakhaPOS: Part-of-Speech Tagging for Typologically Diverse African Languages](https://mlsanity.com/paper/2305.13989/) (2023)
- [An Efficient Multilingual Language Model Compression through Vocabulary Trimming](https://mlsanity.com/paper/2305.15020/) (2023)
- [GlotCC: An Open Broad-Coverage CommonCrawl Corpus and Pipeline for Minority Languages](https://mlsanity.com/paper/2410.23825/) (2024)
- [Distilling Efficient Language-Specific Models for Cross-Lingual Transfer](https://mlsanity.com/paper/2306.01709/) (2023)
- [Which Student is Best? A Comprehensive Knowledge Distillation Exam for Task-Specific BERT Models](https://mlsanity.com/paper/2201.00558/) (2022)
- [The Privileged Students: On the Value of Initialization in Multilingual Knowledge Distillation](https://mlsanity.com/paper/2406.16524/) (2024)
- [Extracting General-use Transformers for Low-resource Languages via Knowledge Distillation](https://mlsanity.com/paper/2501.12660/) (2025)

## Abstract (arXiv)

In this paper, we combine two-step knowledge distillation, structured pruning, truncation, and vocabulary trimming for extremely compressing multilingual encoder-only language models for low-resource languages. Our novel approach systematically combines existing techniques and takes them to the extreme, reducing layer depth, feed-forward hidden size, and intermediate layer embedding size to create significantly smaller monolingual models while retaining essential language-specific knowledge. We achieve compression rates of up to 92% while maintaining competitive performance, with average drops of 2-10% for moderate compression and 8-13% at maximum compression in four downstream tasks, including sentiment analysis, topic classification, named entity recognition, and part-of-speech tagging, across three low-resource languages. Notably, the performance degradation correlates with the amount of language-specific data in the teacher model, with larger datasets resulting in smaller performance losses. Additionally, we conduct ablation studies to identify the best practices for multilingual model compression using these techniques.