AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

LIFT and PLACE: A Simple, Stable, and Effective Knowledge Distillation Framework for Lightweight Diffusion Models

Hyunsoo Han, Sangyeop Yeo, Jaejun Yoo

CVPR 2026 · Label-efficient and robust learning › Knowledge distillation · approach: Rethinking distillation objectives · 0 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 adaptive coarse-to-fine weight instead of linear schedule (ablation)LIFT · CelebA 90% prunedFID: 18.09 -> 15.73Improved
add adaptive coarse-to-fine weight instead of cosine schedule (ablation)LIFT · CelebA 90% prunedFID: 17.45 -> 15.73Improved
add PLACE to LIFT with FeatKD (ablation)LIFT · CelebA 90% pruned, 78.7M teacherFID: 18.49 -> 15.73Improved
add PLACE to LIFT without FeatKD (ablation)LIFT · CelebA 90% pruned, 78.7M teacherFID: 19.07 -> 16.67Improved
add FeatKD to LIFT+PLACE (ablation)LIFT+PLACE · CelebA 90% pruned, 78.7M teacherFID: 16.67 -> 15.73Improved
replace OutKD+FeatKD with LIFT/PLACE at 90% pruningKD · CelebA 90% prunedFID: 211.23 -> 15.73Improved
use OutKD instead of fine-tuning without KD at 50% and 70% pruningOutKD · CelebAFID: 5.30 -> 5.53 and 6.32 -> 6.46Hurt
replace OutKD+FeatKD with LIFT/PLACE at 30% pruningKD · LSUN-BedroomFID: 23.35 -> 16.57Improved
use strongest 78.7M teacher with conventional OutKD+FeatKD instead of 9.2M teacher (ablation)OutKD+FeatKD · CelebA 1.3M studentFID: 40.88 -> 193.56Hurt
use strongest 78.7M teacher with LIFT/PLACE (ablation)LIFT+PLACE · CelebA 1.3M studentFID: 25.38 -> 17.03Improved
use too small group size (K=2^3) or too large (K=2^6) in PLACE (ablation)PLACE · CelebA 90% prunedFID, precision, recallHurt
replace OutKD with LIFT/PLACE for BK-Base-v2BK-SDM · LAION SD 2.1FID: 15.85 -> 16.72Hurt

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