English
Related papers

Related papers: Masked Structural Growth for 2x Faster Language Mo…

200 papers

As Large Language Models (LLMs) gain widespread practical application, offering model families with varying parameter sizes has become standard practice to accommodate diverse computational requirements. Traditionally, each model in the…

Computation and Language · Computer Science 2026-03-17 Kazuki Yano , Sho Takase , Sosuke Kobayashi , Shun Kiyono , Jun Suzuki

Autoregressive language models are the currently dominant paradigm for text generation, but they have some fundamental limitations that cannot be remedied by scale-for example inherently sequential and unidirectional generation. While…

Computation and Language · Computer Science 2024-08-01 Yuchen Li , Alexandre Kirchmeyer , Aashay Mehta , Yilong Qin , Boris Dadachev , Kishore Papineni , Sanjiv Kumar , Andrej Risteski

Recent advancements in large language models (LLMs) have remarkably enhanced performances on a variety of tasks in multiple languages. However, tokenizers in LLMs trained primarily on English-centric corpora often overly fragment a text…

Computation and Language · Computer Science 2024-08-07 Jimin Hong , Gibbeum Lee , Jaewoong Cho

Modern generative pre-trained language models excel at open-ended text generation, yet continue to underperform on structure-related tasks such as NER, relation extraction, and semantic role labeling, especially when compared to…

Computation and Language · Computer Science 2025-12-23 Minho Lee , Junghyun Min , Yerang Kim , Woochul Lee , Yeonsoo Lee

The pre-training of masked language models (MLMs) consumes massive computation to achieve good results on downstream NLP tasks, resulting in a large carbon footprint. In the vanilla MLM, the virtual tokens, [MASK]s, act as placeholders and…

Computation and Language · Computer Science 2022-11-16 Baohao Liao , David Thulke , Sanjika Hewavitharana , Hermann Ney , Christof Monz

The rapid advancement in Large Language Models has been met with significant challenges in their training processes, primarily due to their considerable computational and memory demands. This research examines parallelization techniques…

Distributed, Parallel, and Cluster Computing · Computer Science 2024-05-27 Ishan Patwardhan , Shubham Gandhi , Om Khare , Amit Joshi , Suraj Sawant

Unstructured pruning reduces the memory footprint in deep neural networks (DNNs). Recently, researchers proposed different types of structural pruning intending to reduce also the computation complexity. In this work, we first suggest a new…

Artificial Intelligence · Computer Science 2021-10-22 Itay Hubara , Brian Chmiel , Moshe Island , Ron Banner , Seffi Naor , Daniel Soudry

Large language models (LLMs) are typically developed through large-scale pre-training followed by task-specific fine-tuning. Recent advances highlight the importance of an intermediate mid-training stage, where models undergo multiple…

Computation and Language · Computer Science 2025-10-09 Kaixiang Mo , Yuxin Shi , Weiwei Weng , Zhiqiang Zhou , Shuman Liu , Haibo Zhang , Anxiang Zeng

Despite their wide adoption, the underlying training and memorization dynamics of very large language models is not well understood. We empirically study exact memorization in causal and masked language modeling, across model sizes and…

Computation and Language · Computer Science 2022-11-04 Kushal Tirumala , Aram H. Markosyan , Luke Zettlemoyer , Armen Aghajanyan

The increasing size and complexity of pre-trained language models have demonstrated superior performance in many applications, but they usually require large training datasets to be adequately trained. Insufficient training sets could…

Computation and Language · Computer Science 2025-02-03 Yaping Chai , Haoran Xie , Joe S. Qin

LLMs are computationally expensive to pre-train due to their large scale. Model growth emerges as a promising approach by leveraging smaller models to accelerate the training of larger ones. However, the viability of these model growth…

Computation and Language · Computer Science 2024-10-23 Wenyu Du , Tongxu Luo , Zihan Qiu , Zeyu Huang , Yikang Shen , Reynold Cheng , Yike Guo , Jie Fu

Recent Large Language Models (LLMs) have significantly advanced natural language processing and automated decision-making. However, these models still encounter difficulties when performing complex reasoning tasks involving logical…

Computation and Language · Computer Science 2025-06-26 Yubo Dong , Hehe Fan

Large Language Models (LLMs) employ three popular training approaches: Masked Language Models (MLM), Causal Language Models (CLM), and Sequence-to-Sequence Models (seq2seq). However, each approach has its strengths and limitations, and…

Machine Learning · Computer Science 2025-02-18 Xuefeng Liu , Songhao Jiang , Bo Li , Rick Stevens

Large Language Models (LLMs) have demonstrated exceptional performance in natural language processing tasks, yet their massive size makes serving them inefficient and costly. Semi-structured pruning has emerged as an effective method for…

Machine Learning · Computer Science 2025-06-25 Hongyi Liu , Rajarshi Saha , Zhen Jia , Youngsuk Park , Jiaji Huang , Shoham Sabach , Yu-Xiang Wang , George Karypis

With the rapid scaling of large language models (LLMs), structured pruning has become a widely used technique to learn efficient, smaller models from larger ones, delivering superior performance compared to training similarly sized models…

Computation and Language · Computer Science 2025-06-04 Bairu Hou , Qibin Chen , Jianyu Wang , Guoli Yin , Chong Wang , Nan Du , Ruoming Pang , Shiyu Chang , Tao Lei

We show how bidirectional transformers trained for masked token prediction can be applied to neural image compression to achieve state-of-the-art results. Such models were previously used for image generation by progressivly sampling groups…

Image and Video Processing · Electrical Eng. & Systems 2023-04-18 Fabian Mentzer , Eirikur Agustsson , Michael Tschannen

The development of state-of-the-art large language models is commonly understood as a two-stage process involving pre-training and post-training. We point out the need for an additional intermediate stage called reinforcement mid-training…

Computation and Language · Computer Science 2025-09-30 Yijun Tian , Shaoyu Chen , Zhichao Xu , Yawei Wang , Jinhe Bi , Peng Han , Wei Wang

Increasing the batch size during training -- a ''batch ramp'' -- is a promising strategy to accelerate large language model pretraining. While for SGD, doubling the batch size can be equivalent to halving the learning rate, the optimal…

Machine Learning · Computer Science 2025-10-17 Alexandru Meterez , Depen Morwani , Jingfeng Wu , Costin-Andrei Oncescu , Cengiz Pehlevan , Sham Kakade

Most works on transformers trained with the Masked Language Modeling (MLM) objective use the original BERT model's fixed masking rate of 15%. We propose to instead dynamically schedule the masking rate throughout training. We find that…

Computation and Language · Computer Science 2024-02-13 Zachary Ankner , Naomi Saphra , Davis Blalock , Jonathan Frankle , Matthew L. Leavitt

Pre-trained encoder-only and sequence-to-sequence (seq2seq) models each have advantages, however training both model types from scratch is computationally expensive. We explore recipes to improve pre-training efficiency by initializing one…

Computation and Language · Computer Science 2023-06-16 Saleh Soltan , Andy Rosenbaum , Tobias Falke , Qin Lu , Anna Rumshisky , Wael Hamza