AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

DiffRate : Differentiable Compression Rate for Efficient Vision Transformers

Mengzhao Chen, Wenqi Shao, Peng Xu, Mingbao Lin, Kaipeng Zhang, Fei Chao, Rongrong Ji, Yu Qiao, Ping Luo

ICCV 2023 · Architectures and efficiency › Attention and transformers · approach: Vision transformer pruning · 60 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
learn layer-wise compression rates with differentiable proxy instead of handcrafted scheduletoken compression on off-the-shelf DeiT/MAE/LV-ViT · ImageNet-1k valtop-1 accuracy: ToMe 77.99 -> DiffRate 78.76 (ViT-S, 2.3 GFLOPs)Improved
combine pruning then merging instead of pruning only (ablation)DiffRate on ViT-B (DeiT) · ImageNet-1k valtop-1 accuracy: 80.83 -> 81.50Improved
combine pruning then merging instead of merging only (ablation)DiffRate on ViT-B (DeiT) · ImageNet-1k valtop-1 accuracy: 81.14 -> 81.50Improved
merge-then-prune order instead of prune-then-merge (ablation)DiffRate on ViT-B (DeiT) · ImageNet-1k valtop-1 accuracy: 81.50 -> 81.18Hurt
sort tokens by class attention instead of random rank (ablation)DiffRate on ViT-B (DeiT) · ImageNet-1k valtop-1 accuracy: 79.72 -> 81.50Improved
use A_c times value norm instead of class attention A_c (ablation)DiffRate sorting metric · ImageNet-1k val ViT-B (DeiT)top-1 accuracy: 81.50 -> 81.53No clear effect
use only 1,000 images to search compression rate instead of all training data (ablation)DiffRate · ImageNet-1k val ViT-B (DeiT)top-1 accuracy: 81.50 -> 81.40No clear effect
extend search from 3 to 30 epochs (ablation)DiffRate · ImageNet-1k val ViT-B (DeiT)top-1 accuracy: 81.50 -> 81.52No clear effect
search for 3 epochs rather than 1 epoch (ablation)DiffRate · ImageNet-1k val ViT-B (DeiT)top-1 accuracy: 81.32 -> 81.50Improved
add latency and power constraints (DiffRate-M) to the FLOPs constraint (ablation)DiffRate on ViT-S (DeiT) · ImageNet-1k val, Gemmini acceleratorlatency, power, accuracy: latency 40.1 -> 37.6 ms, power 98 -> 90 mW, accuracy 79.83 -> 79.80Improved
fine-tune for 30 epochs after searching compression rate (ablation)DiffRate on ViT-S (DeiT) · ImageNet-1k valtop-1 accuracy: 79.58 -> 79.83Improved
apply off-the-shelf DiffRate on ViT-L (MAE) versus ToMe trainedDiffRate · ImageNet-1k val, ViT-L (MAE)top-1 accuracy: 85.59 (ToMe trained) vs 85.56No clear effect

Later papers that compared with this method

5 later accepted papers report DiffRate as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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