BabyHGRN: Exploring RNNs for Sample-Efficient Training of Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Haller, Patrick, Golde, Jonas, Akbik, Alan
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908619337891840
author Haller, Patrick
Golde, Jonas
Akbik, Alan
author_facet Haller, Patrick
Golde, Jonas
Akbik, Alan
contents This paper explores the potential of recurrent neural networks (RNNs) and other subquadratic architectures as competitive alternatives to transformer-based models in low-resource language modeling scenarios. We utilize HGRN2 (Qin et al., 2024), a recently proposed RNN-based architecture, and comparatively evaluate its effectiveness against transformer-based baselines and other subquadratic architectures (LSTM, xLSTM, Mamba). Our experimental results show that BABYHGRN, our HGRN2 language model, outperforms transformer-based models in both the 10M and 100M word tracks of the challenge, as measured by their performance on the BLiMP, EWoK, GLUE and BEAR benchmarks. Further, we show the positive impact of knowledge distillation. Our findings challenge the prevailing focus on transformer architectures and indicate the viability of RNN-based models, particularly in resource-constrained environments.
format Preprint
id arxiv_https___arxiv_org_abs_2412_15978
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle BabyHGRN: Exploring RNNs for Sample-Efficient Training of Language Models
Haller, Patrick
Golde, Jonas
Akbik, Alan
Computation and Language
This paper explores the potential of recurrent neural networks (RNNs) and other subquadratic architectures as competitive alternatives to transformer-based models in low-resource language modeling scenarios. We utilize HGRN2 (Qin et al., 2024), a recently proposed RNN-based architecture, and comparatively evaluate its effectiveness against transformer-based baselines and other subquadratic architectures (LSTM, xLSTM, Mamba). Our experimental results show that BABYHGRN, our HGRN2 language model, outperforms transformer-based models in both the 10M and 100M word tracks of the challenge, as measured by their performance on the BLiMP, EWoK, GLUE and BEAR benchmarks. Further, we show the positive impact of knowledge distillation. Our findings challenge the prevailing focus on transformer architectures and indicate the viability of RNN-based models, particularly in resource-constrained environments.
title BabyHGRN: Exploring RNNs for Sample-Efficient Training of Language Models
topic Computation and Language
url https://arxiv.org/abs/2412.15978