List your company
ResearchFormerly WizWand

Interleaved Image Multimodal Understanding on Mantis

64.2ScoreFull Model
-2.56814.76632.149.434Apr 14, 2025Jun 13, 2025Aug 13, 2025Oct 13, 2025Dec 13, 2025Feb 12, 2026Apr 14, 2026
Updated 6mo ago

Evaluation results

MethodLinks
Full ModelSparsity Ratio=0%, Base Model=LLaVA-OneVision2025.0464.2
Full ModelBackbone=LLaVA-OneVision, Sparsity Ratio=0%2026.0464.2
TopoVLMBackbone=LLaVA-OneVision, Sparsity Ratio=60%2026.0461
LLM-StreamlineBackbone=LLaVA-OneVision, Sparsity Ratio=60%2026.0458.9
ECOFLaPSparsity Ratio=60%, Base Model=LLaVA-OneVision2025.0458.1
ECoFLaPBackbone=LLaVA-OneVision, Sparsity Ratio=60%2026.0458.1
TAMPSparsity Ratio=60%, Base Model=LLaVA-OneVision2025.0457.1
TAMPBackbone=LLaVA-OneVision, Sparsity Ratio=60%2026.0457.1
LLM-PrunerBackbone=LLaVA-OneVision, Sparsity Ratio=60%2026.0455.4
SparseGPTSparsity Ratio=60%, Base Model=LLaVA-OneVision2025.0451.9
SparseGPTBackbone=LLaVA-OneVision, Sparsity Ratio=60%2026.0451.9
WandaSparsity Ratio=60%, Base Model=LLaVA-OneVision2025.0447.9
WandaBackbone=LLaVA-OneVision, Sparsity Ratio=60%2026.0447.9
OWLSparsity Ratio=60%, Base Model=LLaVA-OneVision2025.0443.8
OWLBackbone=LLaVA-OneVision, Sparsity Ratio=60%2026.0443.8
MagnitudeSparsity Ratio=60%, Base Model=LLaVA-OneVision2025.040
MagnitudeBackbone=LLaVA-OneVision, Sparsity Ratio=60%2026.040