paper-with-me

홈 › Papers

ResidualTransformer: Residual Low-Rank Learning with Weight-Sharing for Transformer Layers

2023-10-03 · Yiming Wang, Jinyu Li

Memory constraint of always-on devices is one of the major concerns when deploying speech processing models on these devices. While larger models trained with sufficiently large amount of data generally perform better, making them fit in the device memory is a demanding challenge. In this paper, we aim to reduce model size by reparameterizing model weights across Transformer encoder layers and assuming a special weight composition and structure. More specifically, inspired by ResNet and the more recent LoRA work, we propose an approach named ResidualTransformer, where each weight matrix in a Transformer layer comprises 1) a shared full-rank component with its adjacent layers, and 2) a unique low-rank component to itself. The low-rank matrices only account for a small amount of model size increase. In addition, we add diagonal weight matrices to improve modeling capacity of the low-rank matrices. Experiments of our 10k-hour speech recognition and speech translation tasks show that the Transformer encoder size can be reduced by ~3X with very slight performance degradation.

📄 PDF Abstract BibTeX arXiv:2310.02489

Code (0)

등록된 구현이 없습니다.

Tasks

speech-recognitionSpeech Recognition

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Average Pooling 설명 없음
Kaiming Initialization 설명 없음
1x1 Convolution A 1 x 1 Convolution is a convolution with some special properties in that it can be used for dimensionality reduction,…
Batch Normalization 설명 없음
ReLU How Do I Communicate to Expedia? How Do I Communicate to Expedia? – Call ☎️ +1-(888) 829 (0881) or +1-805-330-4056 or +1-805-330-4056 for Live Support & Special Travel…
Convolution A convolution is a type of matrix operation, consisting of a kernel, a small matrix of weights, that slides over input data performing element-wise multiplication with the…
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…

Similar Papers 제목 키워드 기반

On the Residual Scaling of Looped Transformers: Stability and Transferability

2026-06-16 · Shaowen Wang, Bingrui Li, Ge Zhang, Wenhao Huang 외 arxiv

Looped (weight-tied) Transformers apply a shared residual block $N$ times ($h \leftarrow h + \varepsilon\,f(h)$, same $f$ at each step), increasing effective depth without adding parameters. Prior depth-scaling analyses …

Low-Rank Key Value Attention

2026-01-16 · James O'Neill, Robert Clancy, Mariia Matskevichus, Fergal Reid arxiv

The key-value (KV) cache is a primary memory bottleneck in Transformers. We propose Low-Rank Key-Value (LRKV) attention, which reduces KV cache memory by exploiting redundancy across attention heads, while being compute …

Share Your Attention: Transformer Weight Sharing via Matrix-based Dictionary Learning

2025-08-06 · Magauiya Zhussip, Dmitriy Shopkhoev, Ammar Ali, Stamatios Lefkimmiatis arxiv

Large language models have revolutionized AI applications, yet their high computational and memory demands hinder their widespread deployment. Existing compression techniques focus on intra-block optimizations (e.g., low…

Image Classification

Efficient Storage of Fine-Tuned Models via Low-Rank Approximation of Weight Residuals

2023-05-28 · Simo Ryu, Seunghyun Seo, Jaejun Yoo

In this paper, we present an efficient method for storing fine-tuned models by leveraging the low-rank properties of weight residuals. Our key observation is that weight residuals in large overparameterized models exhibi…

Quantization

ResViT: Residual vision transformers for multi-modal medical image synthesis

2021-06-30 · Onat Dalmaz, Mahmut Yurt, Tolga Çukur

Generative adversarial models with convolutional neural network (CNN) backbones have recently been established as state-of-the-art in numerous medical image synthesis tasks. However, CNNs are designed to perform local pr…

DiversityImage GenerationImage-to-Image TranslationInductive Bias