AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

Pre-training Vision Transformers with Very Limited Synthesized Images

Ryo Nakamura, Hirokatsu Kataoka, Sora Takashima, Edgar Josafat Martinez Noriega, Rio Yokota, Nakamasa Inoue

ICCV 2023 · Architectures and efficiency › Pruning and efficient tuning · approach: Efficient ViT adaptation · 17 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
replace 1,000 instances per category with a single instance plus data augmentation (OFDB vs FractalDB)FDSL pre-training of ViT-T · CIFAR-100accuracy: 81.6 -> 84.0Improved
add random pattern augmentation (ablation)2D-OFDB-1k · CIFAR-100accuracy: 84.0 -> 85.3Improved
add IFS parameter fluctuation augmentation (ablation)2D-OFDB-1k · CIFAR-100accuracy: 84.0 -> 81.6Hurt
add rotation augmentation (ablation)2D-OFDB-1k · CIFAR-100accuracy: 84.0 -> 84.1No clear effect
scale 2D-OFDB to 21k categoriesViT-B · ImageNet-1k fine-tuningaccuracy: 0.4 points higher than ImageNet-21kImproved
scale 2D-OFDB to 21k categoriesViT-T · ImageNet-1k fine-tuningaccuracy: 73.8 vs 74.1 (ImageNet-21k)Hurt
reduce instances per category from 1000 to 1 (ablation)ImageNet-1k pre-training (supervised) · CIFAR-100accuracyHurt
use fixed 30-degree yaw-only viewpoints for 3D fractals instead of random multi-axis viewpoints (ablation)3D-OFDB · CIFAR-100accuracyImproved
balanced easy:hard (50:50) data pruning for category selection (ablation)2D-OFDB-1k from 21k · CIFAR-100accuracyImproved
convert one-instance ImageNet to binary or Canny images (ablation)ImageNet-1k subset pre-training · CIFAR-100accuracyImproved

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