AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

SP-VLA: A Joint Model Scheduling and Token Pruning Approach for VLA Model Acceleration

Ye Li, Yuan Meng, Zewen Sun, Kangye Ji, Chen Tang, Jiajun Fan, Xinzhu Ma, Shu-Tao Xia, Zhi Wang, Wenwu Zhu

ICLR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Vision-language-action foundation models

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 action-type aware model scheduling (ablation)SP-VLA · LIBERO averagesuccess rate (%): 71.52 -> 74.90Improved
add spatio-semantic token pruning (ablation)SP-VLA · LIBERO averagespeedup: 1.27 -> 1.35Improved
add Canny edge (contour) tokens to token pruning (ablation)SP-VLA · LIBERO averagesuccess rate (%): 23.93 -> 74.90Improved
preserve relative token positions and add edge tokens to baseline pruners (+R, +S)FastVLM, VisionZip · LIBERO averagesuccess rate (%): 74.60 -> 66.65 (FastVLM), 33.05 (VisionZip)No clear effect
apply VLM token pruning (PruMerge + R) to OpenVLAOpenVLA · LIBERO averagesuccess rate (%): 74.60 -> 0.00Hurt
random token droppingCogACT · SimplerEnv Visual Matchingsuccess rate (%): 74.80 -> 20.90Hurt
apply SP-VLA to CogACTCogACT · SimplerEnv Visual Aggregationsuccess rate (%): 61.30 -> 65.16Improved
apply SP-VLA to CogACTCogACT · SimplerEnv WindowXsuccess rate (%): 51.30 -> 57.29Improved
apply SP-VLA to CogACTCogACT · SimplerEnv Visual Matchingfrequency (Hz): 3.77 -> 8.06Improved
apply SP-VLA to OpenVLAOpenVLA · LIBEROfrequency (Hz): 3.18 -> 4.43Improved
increase action buffer size n from 6 to 8 (ablation)SP-VLA · LIBERO-Objectsuccess rate (%): 82.4 -> 68.4Hurt
lower tau from 0.5 to 3/8 (ablation)SP-VLA · LIBERO-Objectsuccess rate (%): 82.4 -> 64.4Hurt

Later papers that compared with this method

1 later accepted paper reports + SP-VLA (128) as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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