AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

MODEL SOUPS NEED ONLY ONE INGREDIENT

Alireza Abdollahpoorrostam, Nikolaos Dimitriadis, Adam Hazimeh, Pascal Frossard

ICML 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Parameter-efficient transfer analysis · 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
apply MonoSoup (R=0.8) to the best-OOD checkpointFT CLIP ViT-B/32 · ImageNet avg OODaccuracy: 50.67% -> 51.60%Improved
apply MonoSoup to the worst-OOD checkpointFT CLIP ViT-B/32 · ImageNet avg OODaccuracy: 36.71% -> 44.21%Improved
apply MonoSoup to the worst-ID checkpointFT CLIP ViT-B/32 · ImageNet avg OODaccuracy: 38.64% -> 46.54%Improved
use MonoSoup on a model pair instead of ModelStockCLIP ViT-B/32 pair (ID+, OOD+) · ImageNet avg OODaccuracy: 50.53% -> 51.37%Improved
apply MonoSoup instead of LiNeSQwen3-0.6B M-1 · GSM_Plusaccuracy: 30.1 -> 30.2Improved
apply MonoSoup to the fine-tuned Qwen3-0.6BQwen3-0.6B M-1 · GSM_Plusaccuracy: 29.5 -> 30.2Improved
apply MonoSoup post-hoc to English-math fine-tuned modelQwen3-0.6B · MGSM averageaccuracy: 30.6 -> 33.8Improved
apply MonoSoup to fine-tuned ConvNeXt-Small with fixed R=0.8ConvNeXt-Small · ImageNet avg OODaccuracy: 55.85% -> 56.70%Improved
apply MonoSoup to ZH languageQwen3-0.6B · MGSM ZHaccuracy: 39.2 -> 38.8Hurt
use both spectral decay and alignment in the mixing rule instead of one signal (ablation)MonoSoup · CLIP ViT-L/14ID/OOD accuracyImproved
set variance threshold R too small (ablation)MonoSoup (fixed R) · CLIP ViT-L/14ID/OOD accuracyHurt
use MonoSoup as start point for Wise-FT interpolationWise-FT · CLIP ViT-B/32, 70 checkpointsID-OOD Pareto frontImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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