English
Related papers

Related papers: Training-Free Looped Transformers

200 papers

Large Language Models (LLMs) are predominantly deployed as dense transformers, where every parameter in every feed-forward block is activated for every token. While architecturally simple, this is computationally inefficient, since…

Machine Learning · Computer Science 2025-11-27 Ivan Novikov

Looped Transformers offer a promising alternative to purely feed-forward computation by iteratively refining latent representations, improving language modeling and reasoning. Yet recurrent architectures remain unstable to train, costly to…

Machine Learning · Computer Science 2026-05-13 Jacob Fein-Ashley , Paria Rashidinejad

Pruning is an effective way to reduce the huge inference cost of Transformer models. However, prior work on pruning Transformers requires retraining the models. This can add high training cost and high complexity to model deployment, making…

Computation and Language · Computer Science 2022-10-18 Woosuk Kwon , Sehoon Kim , Michael W. Mahoney , Joseph Hassoun , Kurt Keutzer , Amir Gholami

LLM architecture research generally aims to maximize model quality subject to fixed compute/latency budgets. However, many applications of interest such as edge and on-device deployment are further constrained by the model's memory…

Machine Learning · Computer Science 2026-04-28 Abbas Zeitoun , Lucas Torroba-Hennigen , Yoon Kim

Post-training, which elicits a pretrained Base model into the corresponding Instruct model, is widely considered to be superficial. In this work, we first reinforce this hypothesis by providing novel quantitative evidence from the weight…

Computation and Language · Computer Science 2025-09-30 Taiqiang Wu , Runming Yang , Tao Liu , Jiahao Wang , Zenan Xu , Ngai Wong

We evaluate three simple, normalization-centric changes to improve Transformer training. First, we show that pre-norm residual connections (PreNorm) and smaller initializations enable warmup-free, validation-based training with large…

Computation and Language · Computer Science 2020-01-01 Toan Q. Nguyen , Julian Salazar

We train a linear attention transformer on millions of masked-block matrix completion tasks: each prompt is masked low-rank matrix whose missing block may be (i) a scalar prediction target or (ii) an unseen kernel slice of Nystr\"om…

Machine Learning · Computer Science 2025-09-25 Patrick Lutz , Aditya Gangrade , Hadi Daneshmand , Venkatesh Saligrama

Self-supervised pre-training of large-scale transformer models on text corpora followed by finetuning has achieved state-of-the-art on a number of natural language processing tasks. Recently, Lu et al. (2021, arXiv:2103.05247) claimed that…

Machine Learning · Computer Science 2021-07-28 Danielle Rothermel , Margaret Li , Tim Rocktäschel , Jakob Foerster

Transformer-based models have delivered impressive results on many tasks, particularly vision and language tasks. In many model training situations, conventional configurations are typically adopted. For example, we often set the base model…

Machine Learning · Computer Science 2023-05-19 Fuzhao Xue , Jianghai Chen , Aixin Sun , Xiaozhe Ren , Zangwei Zheng , Xiaoxin He , Yongming Chen , Xin Jiang , Yang You

We present a framework for using transformer networks as universal computers by programming them with specific weights and placing them in a loop. Our input sequence acts as a punchcard, consisting of instructions and memory for data…

Machine Learning · Computer Science 2023-01-31 Angeliki Giannou , Shashank Rajput , Jy-yong Sohn , Kangwook Lee , Jason D. Lee , Dimitris Papailiopoulos

Large-scale transformer models have become the de-facto architectures for various machine learning applications, e.g., CV and NLP. However, those large models also introduce prohibitive training costs. To mitigate this issue, we propose a…

Computation and Language · Computer Science 2022-11-22 Zhewei Yao , Xiaoxia Wu , Conglong Li , Connor Holmes , Minjia Zhang , Cheng Li , Yuxiong He

Embedding layers in transformer-based NLP models typically account for the largest share of model parameters, scaling with vocabulary size but not yielding performance gains proportional to scale. We propose an alternative approach in which…

Computation and Language · Computer Science 2025-05-06 Henry Ndubuaku , Mouad Talhi

Transformers have achieved extraordinary success in modern machine learning due to their excellent ability to handle sequential data, especially in next-token prediction (NTP) tasks. However, the theoretical understanding of their…

Machine Learning · Computer Science 2024-10-01 Ruiquan Huang , Yingbin Liang , Jing Yang

Large language models (LLMs) have made significant strides in complex tasks, yet their widespread adoption is impeded by substantial computational demands. With hundreds of billion parameters, transformer-based LLMs necessitate months of…

Machine Learning · Computer Science 2024-08-22 Pihe Hu , Shaolong Li , Longbo Huang

We introduce ReplaceMe, a generalized training-free depth pruning method that effectively replaces transformer blocks with a linear operation, while maintaining high performance for low compression ratios. In contrast to conventional…

Computation and Language · Computer Science 2026-02-20 Dmitriy Shopkhoev , Ammar Ali , Magauiya Zhussip , Valentin Malykh , Stamatios Lefkimmiatis , Nikos Komodakis , Sergey Zagoruyko

Scaling large language models (LLMs) significantly improves performance but comes with prohibitive computational costs. Mixture-of-Experts (MoE) models offer an efficient alternative, increasing capacity without a proportional rise in…

Machine Learning · Computer Science 2024-12-16 Aditya Vavre , Ethan He , Dennis Liu , Zijie Yan , June Yang , Nima Tajbakhsh , Ashwath Aithal

Mixture-of-Experts (MoE) models improve transformer efficiency but lack a unified theoretical explanation, especially when both feed-forward and attention layers are allowed to specialize. To this end, we study the Mixture-of-Transformers…

Machine Learning · Computer Science 2025-11-03 Hongbo Li , Qinhang Wu , Sen Lin , Yingbin Liang , Ness B. Shroff

We investigate the capability of a transformer pretrained on natural language to generalize to other modalities with minimal finetuning -- in particular, without finetuning of the self-attention and feedforward layers of the residual…

Machine Learning · Computer Science 2021-07-01 Kevin Lu , Aditya Grover , Pieter Abbeel , Igor Mordatch

Solving inverse problems without any training involves using a pretrained generative model and making appropriate modifications to the generation process to avoid finetuning of the generative model. While recent methods have explored the…

Computer Vision and Pattern Recognition · Computer Science 2024-10-22 Ashwini Pokle , Matthew J. Muckley , Ricky T. Q. Chen , Brian Karrer

Training large transformer models from scratch for a target task requires lots of data and is computationally demanding. The usual practice of transfer learning overcomes this challenge by initializing the model with weights of a pretrained…