AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

"Tracking Meets LoRA: Faster Training, Larger Model, Stronger Performance"

Liting Lin, Heng Fan, Zhipeng Zhang, Yaowei Wang, Yong Xu, Haibin Ling

ECCV 2024 · Architectures and efficiency › Pruning and efficient tuning · approach: Efficient ViT adaptation · 88 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add token type embedding with shared (sliced) positional embedding for template and search (ablation)LoRAT-L-224 · LaSOT-extSUC: 51.9 -> 52.7Improved
replace MLP-only head with convolutional head under LoRA fine-tuning (ablation)LoRAT-L-224 · LaSOTSUC: 74.2 -> 4.7Hurt
convolutional head vs MLP head under full fine-tuning (ablation)LoRAT-L-224 · LaSOTSUC: 73.0 -> 71.7Hurt
use LoRA instead of full fine-tuning (ablation)LoRAT-L-224 · LaSOTSUC: 73.0 -> 74.2Improved
add foreground indication embedding at 378 resolution (ablation)LoRAT-L-378 · LaSOT-extSUC: 55.2 -> 56.6Improved
unfreeze positional embedding (ablation)LoRAT-L-224 · LaSOTSUC: 73.5 vs 73.9No clear effect
use DINOv2 pre-training instead of MAE, CLIP or EVA-02 (ablation)LoRAT-B-224 (full fine-tuning) · LaSOTSUC: 70.3 (MAE) -> 70.9Improved
use LoRA on all linear layers (attention and MLP) rather than only q,v (ablation)LoRAT-L-224 · LaSOTSUC: 71.9 -> 74.2Improved
replace LoRA with Adapter, VPT or IA3 (ablation)LoRAT-L-224 · LaSOTSUC: 74.2 -> 55.0 (Adapter)Hurt
use truncated normal init for LoRA instead of Kaiming uniform or Gaussian (ablation)LoRAT-L-224 · LaSOTSUC: 73.7 -> 74.2Improved
replace full self-attention with S-MAM (ablation)LoRAT-L-224 · LaSOTSUC: 74.2 -> 73.5Hurt
use LoRA instead of full fine-tuning for training efficiency (ablation)LoRAT-g-378 · Training memory (8 V100)GPU memory / feasibility: Out of Memory -> 25.8 GBImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy