AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

StochasT: Learning with Stochastic Turn Depth for Visual Instruction Tuning

Yuan Qing, Chengzhi Mao, Boqing Gong

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: VLM capability 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
use StochasT instead of MultiT trainingLLaVA-1.5-7B · Avg of 4 VIT datasets (MultiT eval)accuracy: 68.52 -> 70.63Improved
use StochasT instead of MultiT trainingQwen2.5-VL-3B · Avg of 4 VIT datasets (MultiT eval)accuracy: 77.39 -> 80.16Improved
use StochasT instead of MultiT trainingLLaVA-1.5-7B · Avg of 4 VIT datasets (SingleT eval)accuracy: 61.51 -> 67.16Improved
use StochasT instead of MultiT trainingQwen2.5-VL-3B · Avg of 4 VIT datasets (SingleT eval)accuracy: 71.20 -> 77.28Improved
use StochasT instead of MultiT trainingLLaVA-1.5-7B · MMDUscore: 12.68 -> 13.10Improved
use StochasT instead of MultiT trainingQwen2.5-VL-3B · MMDUscore: 57.90 -> 59.80Improved
use StochasT instead of MultiT trainingLLaVA-1.5-7B · Avg of 4 VIT datasets (BLS eval)CRA / CRA+: CRA 61.82 -> 67.69; CRA+ 41.19 -> 50.89Improved
use StochasT instead of SingleT trainingLLaVA-1.5-7B · Avg of 4 VIT datasets (BLS eval)CRA+: 53.15 -> 50.89Hurt
use StochasT instead of SingleT trainingQwen2.5-VL-3B · Avg of 4 VIT datasets (SingleT eval)accuracy: 76.71 -> 77.28No clear effect
use Beta(2,2) instead of Beta(0.5,0.5) for dropout rates (ablation)StochasT · CoralVQA + iNat-Plant avgCRA+: 69.05 -> 73.30Improved
use Beta(5,1) instead of Beta(1,5) for dropout rates (ablation)StochasT · CoralVQA + iNat-Plant avgCRA+: 69.81 -> 72.69Improved
replace StochasT with random-fixed cutoff at matched expected depth (ablation)Qwen2.5-VL-3B · iNat-Plant + PathVQA avgCRA: 79.24 -> 70.61Hurt

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