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.

LOTUS: Improving Transformer Efficiency with Sparsity Pruning and Data Lottery Tickets

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

Early Transformers: A study on Efficient Training of Transformer Models through Early-Bird Lottery Tickets

Sparse then Prune: Toward Efficient Vision Transformers

Data Level Lottery Ticket Hypothesis for Vision Transformers

A Survey on Efficient Vision Transformers: Algorithms, Techniques, and Performance Benchmarking

Learned Thresholds Token Merging and Pruning for Vision Transformers

Peeling the Onion: Hierarchical Reduction of Data Redundancy for Efficient Vision Transformer Training.

Vision Transformer with Sparse Scan Prior

Patch Slimming for Efficient Vision Transformers

Chasing Sparsity in Vision Transformers: An End-to-End Exploration

Width & Depth Pruning for Vision Transformers

Three things everyone should know about Vision Transformers

Improve Vision Transformers Training by Suppressing Over-smoothing

An Attention-Based Token Pruning Method for Vision Transformers

Vision Transformer Slimming: Multi-Dimension Searching in Continuous Optimization Space

Boost Vision Transformer with GPU-Friendly Sparsity and Quantization

Beyond 2:4: exploring V:N:M sparsity for efficient transformer inference on GPUs

CP-ViT: Cascade Vision Transformer Pruning via Progressive Sparsity Prediction

Scattering Vision Transformer: Spectral Mixing Matters

Enhancing Vision Transformer: Amplifying Non-Linearity in Feedforward Network Module