List your company

Visual Question Answering on InstaOrder Depth (val)

66.1AccuracySNs (mean)
63.29264.02164.7565.479Mar 11, 2026

Evaluation results

MethodLinks
SNs (mean)Backbone=LLaVA-v1.5-7b2026.0366.16368.875.972.2
SNs (maj. voting)Backbone=LLaVA-v1.5-7b2026.0365.26366.480.972.9
LLaVA-v1.5-7b2026.0365—66.480.272.6
SNs (maj. voting)Backbone=Qwen3-VL-4b-Instruct2026.0364.962.669.271.170.1
SNs (mean)Backbone=Qwen3-VL-4b-Instruct2026.0363.662.670.164.867.4
Qwen3-VL-4b-Instruct2026.0363.4—69.36667.6