AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

TRAC: Tensor-Train based Across-layer Compression for Parameter-Efficient Fine-Tuning

Bangguo Ye, Yuanwei Zhang, Xiaoqun Zhang

ICLR 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Low-rank compression and merging

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 vector controllers on shared TT cores (ablation)TRAC · CoLA (RoBERTa-Base)Matthews corr.: 61.63 -> 65.2Improved
make tensor cores trainable instead of frozen (ablation)TRAC · CoLA (RoBERTa-Base)Matthews corr.: 38.98 -> 65.2Improved
make tensor cores trainable instead of frozen (ablation)TRAC · CIFAR-100 (ViT-Base)accuracy: 67.47 -> 84.21Improved
make all cores independently trainable instead of freeze+share (ablation)TRAC · CoLA (RoBERTa-Base)Matthews corr.: 65.2 -> 66.17No clear effect
freeze-share strategy vs all trainable at same TT rank (ablation)TRAC · RoBERTa-Base MRPCaccuracy: 88.92 -> 88.63No clear effect
use TT normal initialization instead of TT uniform (ablation)TRAC · CoLA (RoBERTa-Base)Matthews corr.: 62.4 -> 65.2Improved
use TT normal instead of Kaiming normal initialization (ablation)TRAC · MRPC (RoBERTa-Base)accuracy: 87.9 -> 89.0Improved
use asymmetric rank allocation (more to B) instead of symmetric (ablation)TRAC · CoLA (RoBERTa-Base)Matthews corr.: 62.6 -> 65.2Improved
use asymmetric rank allocation instead of symmetric (ablation)TRAC · E2E (GPT-2)BLEU: 69.4 -> 70.3Improved
use 3rd-order instead of 4th-order tensor (ablation)TRAC · RTE (RoBERTa-Base)accuracy: 75.3 -> 80.1Improved
increase tensor order from 3rd to 5th (ablation)TRAC · training time per epochseconds: 776.3 -> 873.7Hurt
replace LoRA with TRACLoRA on ViT-Base · image classification averageaccuracy: 90.27 -> 90.85Improved

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