SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.11932 · CVPR · 2022

Dataset Distillation by Matching Training Trajectories

Tongzhou Wang, Alexei Efros, Antonio Torralba, Jun-Yan Zhu, George Cazenavette

arXiv · PDF · Open in the Atlas

Code that ran

We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. The repositories linked to it are listed below.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Figure 1. Example distilled images from 32x32 CIFAR-100 (top), 64x64 Tiny ImageNet (middle), and 128x128 ImageNet subsets (bottom). Training a standard CNN using only such distilled images (as few as one per category) yields a trained model capable of test accuracy significantly better than previous methods of dataset distillation. Please see more results at https://georgecazenavette.github.io/mtt-distillation/.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2203.11932")
get_code_for_paper("2203.11932")
have("2203.11932")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2203.11932.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2203.11932.svg)](https://syntology.ai/paper/2203.11932)

Connect an agent — have() is free.