AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

TinyViT: Fast Pretraining Distillation for Small Vision Transformers

Kan Wu, Jinnian Zhang, Houwen Peng, Mengchen Liu, Bin Xiao, Jianlong Fu, Lu Yuan

ECCV 2022 · Architectures and efficiency › Attention and transformers · approach: Hybrid and interpretable ViT designs

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 pretraining distillation on IN-21k (ablation)DeiT-Ti · ImageNet-1k valtop-1 accuracy (%): 73.0 -> 74.4Improved
add pretraining distillation on IN-21k (ablation)DeiT-S · ImageNet-1k valtop-1 accuracy (%): 80.5 -> 82.0Improved
add pretraining distillation on IN-21k (ablation)Swin-T · ImageNet-1k valtop-1 accuracy (%): 81.9 -> 83.4Improved
pretrain on IN-21k without distillation instead of training on IN-1k (ablation)Swin-T · ImageNet-1k valtop-1 accuracy (%): 81.2 -> 81.9Improved
add IN-21k pretraining without distillation (ablation)TinyViT-21M · ImageNet-1k valtop-1 accuracy (%): 83.1 -> 83.8Improved
pretrain on cleaned IN-21k instead of original (ablation)TinyViT-21M · ImageNet-1k valtop-1 accuracy (%): 83.8 -> 84.2Improved
distill from Florence instead of cleaning data (ablation)TinyViT-21M · ImageNet-1k valtop-1 accuracy (%): 84.2 -> 84.8Improved
use CLIP-ViT-L/14 teacher instead of BEiT-L (ablation)TinyViT-21M · ImageNet-1k valtop-1 accuracy (%): 84.1 -> 84.8Improved
use Florence teacher instead of CLIP-ViT-L/14 (ablation)TinyViT-21M · ImageNet-Realtop-1 accuracy (%): 88.9 -> 89.0No clear effect
remove ground-truth labels from distillation (Florence teacher) (ablation)TinyViT-21M variant · ImageNet-1k valtop-1 accuracy (%): 84.2 -> 84.9Improved
remove ground-truth labels from distillation (CLIP teacher) (ablation)TinyViT-21M variant · ImageNet-1k valtop-1 accuracy (%): 84.3 -> 84.5Improved
replace Swin-T backbone with TinyViT-21MCascade Mask R-CNN · COCO valAP: 48.1 -> 49.6Improved

Later papers that compared with this method

1 later accepted paper reports TinyViT-5M as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

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