Retraining-Free Depth Pruning of Vision Transformers: A Cross Domain Study of Layer Redundancy

arXiv:2610.15005 · cs.CV · Submitted 2026-10-10 · Read on arXiv

cs.CV

Submitted: 2026-10-10

Updated: 2026-10-10

Code: https://github.com/zincalex/ViT-Drop

Terminology

• Pruning — 1× in this paper · explained in 4 episodes
• Redundancy — 1× in this paper · explained in 2 episodes
• Transformer — 1× in this paper · explained in 2 episodes

Sources

• SIMPLER: Efficient Foundation Model Adaptation via Similarity-Guided Layer Pruning for Earth Observation
• DAPlankton: Benchmark Dataset for Multi-instrument Plankton Recognition via Fine-grained Domain Adaptation
• A Comprehensive Survey on Hardware-Aware Neural Architecture Search
• Variance-Based Pruning for Accelerating and Compressing Trained Networks
• Token Merging: Your ViT But Faster
• Tree Species Identification from Bark Images Using Convolutional Neural Networks
• Chasing Sparsity in Vision Transformers: An End-to-End Exploration
• GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers
• The Unreasonable Ineffectiveness of the Deeper Layers
• Learning both Weights and Connections for Efficient Neural Networks
• What Matters in Transformers? Not All Attention is Needed
• Distilling the Knowledge in a Neural Network
• Pruning Filters for Efficient ConvNets
• AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration
• MLP Can Be A Good Transformer Learner
• GradMAP: Faster Layer Pruning with Gradient Metric and Projection Compensation
• Swin Transformer V2: Scaling Up Capacity and Resolution
• Preserving Deep Representations In One-Shot Pruning: A Hessian-Free Second-Order Optimization Framework
• ShortGPT: Layers in Large Language Models are More Redundant Than You Expect
• DINOv2: Learning Robust Visual Features without Supervision

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models