LATTE: Low-Precision Approximate Attention with Head-wise Trainable Threshold for Efficient Transformer

• URL: http://arxiv.org/abs/2404.07519v1
• Date: Thu, 11 Apr 2024 07:23:19 GMT
• Title: LATTE: Low-Precision Approximate Attention with Head-wise Trainable Threshold for Efficient Transformer
• Authors: Jiing-Ping Wang, Ming-Guang Lin, An-Yeu, Wu,
• Abstract summary: We propose Low-Precision Approximate Attention with Head-wise Trainable Threshold for Efficient Transformer (LATTE) LATTE employs a headwise threshold-based filter with the low-precision dot product to reduce the computation of Multi-Head Attention (MHA) Experimental results indicate LATTE can smoothly adapt to both NLP and CV tasks, offering significant computation savings.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: With the rise of Transformer models in NLP and CV domain, Multi-Head Attention has been proven to be a game-changer. However, its expensive computation poses challenges to the model throughput and efficiency, especially for the long sequence tasks. Exploiting the sparsity in attention has been proven to be an effective way to reduce computation. Nevertheless, prior works do not consider the various distributions among different heads and lack a systematic method to determine the threshold. To address these challenges, we propose Low-Precision Approximate Attention with Head-wise Trainable Threshold for Efficient Transformer (LATTE). LATTE employs a headwise threshold-based filter with the low-precision dot product and computation reuse mechanism to reduce the computation of MHA. Moreover, the trainable threshold is introduced to provide a systematic method for adjusting the thresholds and enable end-to-end optimization. Experimental results indicate LATTE can smoothly adapt to both NLP and CV tasks, offering significant computation savings with only a minor compromise in performance. Also, the trainable threshold is shown to be essential for the leverage between the performance and the computation. As a result, LATTE filters up to 85.16% keys with only a 0.87% accuracy drop in the CV task and 89.91% keys with a 0.86 perplexity increase in the NLP task.

Related papers

• Memory-Efficient Vision Transformers: An Activation-Aware Mixed-Rank Compression Strategy [5.699098817569033]
  This paper introduces an activation-aware model compression methodology that uses selective low-rank weight tensor approximations of different layers to reduce the parameter count of ViTs. The presented method significantly reduces the parameter count of DeiT-B by 60% with less than 1% accuracy drop on the ImageNet dataset. In addition to this, the presented compression technique can compress large DeiT/ViT models to have about the same model size as smaller DeiT/ViT variants while yielding up to 1.8% accuracy gain.
  arXiv  Detail & Related papers  (2024-02-08T19:01:14Z)
• Parameter and Computation Efficient Transfer Learning for Vision-Language Pre-trained Models [79.34513906324727]
  In this paper, we aim at parameter and efficient transfer learning (PCETL) for vision-language pre-trained models. We propose a novel dynamic architecture skipping (DAS) approach towards effective PCETL.
  arXiv  Detail & Related papers  (2023-09-04T09:34:33Z)
• Approximated Prompt Tuning for Vision-Language Pre-trained Models [54.326232586461614]
  In vision-language pre-trained models, prompt tuning often requires a large number of learnable tokens to bridge the gap between the pre-training and downstream tasks. We propose a novel Approximated Prompt Tuning (APT) approach towards efficient VL transfer learning.
  arXiv  Detail & Related papers  (2023-06-27T05:43:47Z)
• Constraint-aware and Ranking-distilled Token Pruning for Efficient Transformer Inference [18.308180927492643]
  ToP is a ranking-distilled token distillation technique, which distills effective token rankings from the final layer of unpruned models to early layers of pruned models. ToP reduces the average FLOPs of BERT by 8.1x while achieving competitive accuracy on GLUE, and provides a real latency speedup of up to 7.4x on an Intel CPU.
  arXiv  Detail & Related papers  (2023-06-26T03:06:57Z)
• QuaLA-MiniLM: a Quantized Length Adaptive MiniLM [5.36703735486629]
  Limited computational budgets often prevent transformers from being used in production and from having their high accuracy utilized. A knowledge distillation approach addresses the computational efficiency by self-distilling BERT into a smaller transformer representation having fewer layers and smaller internal embedding. Dynamic-TinyBERT tackles both limitations by partially implementing the Length Adaptive Transformer (LAT) technique onto TinyBERT, achieving x3 speedup over BERT-base with minimal accuracy loss. We use MiniLM distillation jointly with the LAT method, and we further enhance the efficiency by applying low-bit quantization.
  arXiv  Detail & Related papers  (2022-10-31T07:42:52Z)
• Effective Pre-Training Objectives for Transformer-based Autoencoders [97.99741848756302]
  We study trade-offs between efficiency, cost and accuracy of Transformer encoders. We combine features of common objectives and create new effective pre-training approaches.
  arXiv  Detail & Related papers  (2022-10-24T18:39:44Z)
• Efficient Decoder-free Object Detection with Transformers [75.00499377197475]
  Vision transformers (ViTs) are changing the landscape of object detection approaches. We propose a decoder-free fully transformer-based (DFFT) object detector. DFFT_SMALL achieves high efficiency in both training and inference stages.
  arXiv  Detail & Related papers  (2022-06-14T13:22:19Z)
• Identifying Suitable Tasks for Inductive Transfer Through the Analysis of Feature Attributions [78.55044112903148]
  We use explainability techniques to predict whether task pairs will be complementary, through comparison of neural network activation between single-task models. Our results show that, through this approach, it is possible to reduce training time by up to 83.5% at a cost of only 0.034 reduction in positive-class F1 on the TREC-IS 2020-A dataset.
  arXiv  Detail & Related papers  (2022-02-02T15:51:07Z)
• FasterPose: A Faster Simple Baseline for Human Pose Estimation [65.8413964785972]
  We propose a design paradigm for cost-effective network with LR representation for efficient pose estimation, named FasterPose. We study the training behavior of FasterPose, and formulate a novel regressive cross-entropy (RCE) loss function for accelerating the convergence. Compared with the previously dominant network of pose estimation, our method reduces 58% of the FLOPs and simultaneously gains 1.3% improvement of accuracy.
  arXiv  Detail & Related papers  (2021-07-07T13:39:08Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.