Abstract:Vision Transformers (ViTs) have recently made a splash in computer vision domain and achieved state-of-the-art in many vision tasks. Nevertheless, due to their vast model size and high computational costs, rare transformer-based models are adopted in real-world applications. Since the computational costs of attention operation is the square of the input size, some compression methods for the Multi-Head Self-Attention (MHSA) module have been proposed, reducing its FLOPs successfully but almost without parameters reduction. Meanwhile, the number of parameters and computational costs in the Feed-Forward Network (FFN) module exceeds the MHSA larger, while its compression technologies have not been delved deeper. Consequently, we focus our insight on the compression of FFN layer and present a pruning method named Multi-Dimension Compression of Feed-Forward Network in Vision Transformers(MCF), which greatly reduces the model’s parameters and computational costs. Firstly, we identify the critical elements in the output of the FFN module and then employ them to guide the irregular sparsity of this layer, recognizing insignificant elements of FFN layer that have less impact on the output. Successively, to discard the insignificant elements, we transform the irregular sparsity into regular sparsity and prune them, thus reducing the models’ parameters and getting a substantial speed-up during inference. Extensive results on ImageNet-1K validate the effectiveness of our proposed method, which obtains significant parameters and computational costs reduction with almost unimpaired generalization. For example, we compress DeiT-Tiny with 42% reduction in FLOPs and 33% reduction in parameters, almost without losing accuracy on the ImageNet dataset. Further, we verify the effectiveness of our method in the downstream task, using the pruned DeiT-Small as the backbone for the object detection task on the COCO dataset, gaining revenue without compromising its performance.

Compressed representation learning method based on tensor decomposition

Multi-Dimension Compression of Feed-Forward Network in Vision Transformers

Hybrid Tensor Decomposition in Neural Network Compression

Semi-tensor Product-based TensorDecomposition for Neural Network Compression

Deep neural network compression by Tucker decomposition with nonlinear response

Deep Convolutional Neural Network Compression Method: Tensor Ring Decomposition with Variational Bayesian Approach

Convolutional Neural Network Compression Based on Low-Rank Decomposition

TDLC: Tensor decomposition‐based direct learning‐compression algorithm for DNN model compression

Tensor Decomposition for Compressing Recurrent Neural Network

Low-rank Tensor Decomposition for Compression of Convolutional Neural Networks Using Funnel Regularization

Convolutional neural networks compression with low rank and sparse tensor decompositions

Tensor Decomposition for Model Reduction in Neural Networks: A Review

Low-Rank+Sparse Tensor Compression for Neural Networks

Stable Low-rank Tensor Decomposition for Compression of Convolutional Neural Network

On Compressing Deep Models by Low Rank and Sparse Decomposition.

Towards Efficient Tensor Decomposition-Based DNN Model Compression with Optimization Framework

Holistic CNN Compression Via Low-Rank Decomposition with Knowledge Transfer.

DNN Compression Approach Based on Bayesian Optimization Tensor Ring Decomposition

Nonlinear Tensor Train Format for Deep Neural Network Compression

STN: Scalable Tensorizing Networks via Structure-Aware Training and Adaptive Compression

Kronecker CP Decomposition with Fast Multiplication for Compressing RNNs