paper-with-me

홈 › Papers

Compression of Generative Pre-trained Language Models via Quantization

2022-03-21 · ACL 2022 5 · Chaofan Tao, Lu Hou, Wei zhang, Lifeng Shang, Xin Jiang, Qun Liu, Ping Luo, Ngai Wong

The increasing size of generative Pre-trained Language Models (PLMs) has greatly increased the demand for model compression. Despite various methods to compress BERT or its variants, there are few attempts to compress generative PLMs, and the underlying difficulty remains unclear. In this paper, we compress generative PLMs by quantization. We find that previous quantization methods fail on generative tasks due to the \textit{homogeneous word embeddings} caused by reduced capacity, and \textit{varied distribution of weights}. Correspondingly, we propose a token-level contrastive distillation to learn distinguishable word embeddings, and a module-wise dynamic scaling to make quantizers adaptive to different modules. Empirical results on various tasks show that our proposed method outperforms the state-of-the-art compression methods on generative PLMs by a clear margin. With comparable performance with the full-precision models, we achieve 14.4x and 13.4x compression rates on GPT-2 and BART, respectively.

📄 PDF Abstract BibTeX arXiv:2203.10705

Code (0)

등록된 구현이 없습니다.

Tasks

Model CompressionQuantizationWord Embeddings

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Cosine Annealing Cosine Annealing is a type of learning rate schedule that has the effect of starting with a large learning rate that is relatively rapidly decreased to a minimum value before…
Linear Warmup With Cosine Annealing Linear Warmup With Cosine Annealing is a learning rate schedule where we increase the learning rate linearly for $n$ updates and then anneal according to a cosine schedule…
Attention Dropout Attention Dropout is a type of dropout used in attention-based architectures, where elements are randomly dropped out of the…
Linear Warmup With Linear Decay Linear Warmup With Linear Decay is a learning rate schedule in which we increase the learning rate linearly for $n$ updates and then linearly decay afterwards.
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

S-PRESSO: Ultra Low Bitrate Sound Effect Compression With Diffusion Autoencoders And Offline Quantization

2026-02-16 · Zineb Lahrichi, Gaëtan Hadjeres, Gaël Richard, Geoffroy Peeters arxiv

Neural audio compression models have recently achieved extreme compression rates, enabling efficient latent generative modeling. Conversely, latent generative models have been applied to compression, pushing the limits o…

GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers

2022-10-31 · Elias Frantar, Saleh Ashkboos, Torsten Hoefler, Dan Alistarh

Generative Pre-trained Transformer models, known as GPT or OPT, set themselves apart through breakthrough performance across complex language modelling tasks, but also by their extremely high computational and storage co…

GPULanguage ModellingModel CompressionQuantization

Bridging the Gap between Gaussian Diffusion Models and Universal Quantization for Image Compression

2025-04-03 · CVPR 2025 1 · Lucas Relic, Roberto Azevedo, Yang Zhang, Markus Gross 외

Generative neural image compression supports data representation at extremely low bitrate, synthesizing details at the client and consistently producing highly realistic images. By leveraging the similarities between qua…

Image CompressionQuantization

Rate-Adaptive Quantization: A Multi-Rate Codebook Adaptation for Vector Quantization-based Generative Models

2024-05-23 · Jiwan Seo, Joonhyuk Kang

Learning discrete representations with vector quantization (VQ) has emerged as a powerful approach in various generative models. However, most VQ-based models rely on a single, fixed-rate codebook, requiring extensive re…

Data CompressionImage GenerationImage ReconstructionQuantization

Layer-Wise Data-Free CNN Compression

2020-11-18 · Maxwell Horton, Yanzi Jin, Ali Farhadi, Mohammad Rastegari

We present a computationally efficient method for compressing a trained neural network without using real data. We break the problem of data-free network compression into independent layer-wise compressions. We show how …

Quantization