AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

FRoD: Full-Rank Efficient Fine-Tuning with Rotational Degrees for Fast Convergence

Guoan Wan, Tianyu Chen, Fangzheng Feng, Haoyi Zhou, Runhua Xu

AAAI 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Adapter-based parameter-efficient tuning · 2 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 off-axis sparse rotation S to on-axis Sigma updates (ablation)FRoD · 4 vision datasets, epoch 10average accuracy: 71.26 -> 74.59Improved
add on-axis Sigma updates to off-axis S updates (ablation)FRoD · 4 vision datasets, epoch 10average accuracy: 73.12 -> 74.59Improved
increase on-axis lr(Sigma) to 5e-03 without off-axis update (ablation)FRoD on-axis only · 4 vision datasetsaverage accuracy: epoch 1: 25.89 -> 50.03; epoch 10: 70.77 -> 67.28Mixed
increase off-axis learning rate lr(S) from 1e-04 to 5e-04 (ablation)FRoD off-axis only · 4 vision datasets, epoch 10average accuracy: 73.12 -> 60.60Hurt
increase sparsity s from 0.02 to 0.1 (lr(S)=1e-04) (ablation)FRoD off-axis only · 4 vision datasets, epoch 10average accuracy: 73.12 -> 72.77No clear effect
use FRoD instead of LoRACLIP ViT-B/32 · Image classification average (7 datasets)accuracy: 77.09 -> 82.80Improved
use FRoD instead of VeRALLaMA2-7B · Commonsense reasoning average (6 datasets)accuracy: 66.70 -> 83.53Improved
use FRoD instead of full fine-tuningLLaMA2-7B · Commonsense reasoning average (6 datasets)accuracy: 84.58 -> 83.53Hurt
use FRoD instead of full fine-tuningRoBERTa-large · GLUE average (7 tasks)score: 90.42 -> 90.53No clear effect
keep tan(alpha) of off-axis to on-axis rotation in moderate range (ablation)FRoD · 4 vision datasets, epoch 10accuracyImproved

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