Share your thoughts, 1 month free Claude Pro on usSee more

Referring Expression Grounding on RefCOCO (testB)

89.5Accuracy

MM1.5-30B

Updated 1mo ago

Evaluation Results

Method	Links
MM1.5-30B 2024.09		89.5	-	-
Ferret-V2-13B 2024.09		88.9	-	-
Ferret-V2-7B 2024.09		88.7	-	-
G-DINO 2025.05		88.2	-	-
MM1.5-7B 2024.09		86.7	-	-
MM1.5-3B-MoE 2024.09		86.4	-	-
MM1.5-3B 2024.09		86.1	-	-
Argus-X3-8B 2025.05		85.4	-	-
MiniGPT-7B 2025.05		85.3	-	-
QwenVL-7B 2025.05		85.3	-	-
MM1.5-1B-MoE 2024.09		84.8	-	-
Qwen-VL-7B 2024.09		84.5	-	-
Ferret-13B 2024.09		84.4	-	-
MiniGPT-v2-7B 2024.09		84.3	-	-
Ferret-7B 2024.09		82.5	-	-
Ferret-7B 2025.05		82.5	-	-
MM1.5-1B 2024.09		81.9	-	-
MDETR 2025.05		81.4	-	-
InternVL2-8B 2025.05		80.7	-	-
UniTAB 2025.05		80.6	-	-
Shikra-7B 2025.05		80.2	-	-
TransVG 2025.05		78.3	-	-
SPHINX-Tiny 2024.09		77.9	-	-
OFA-L 2025.05		76.4	-	-
InternVL2 2024.09		75.9	-	-
VILLA 2025.05		74.8	-	-
UNITER 2025.05		74.2	-	-
MCN 2021.03		73.09	65.69	-
MAttNet 2025.05		69.3	-	-
MAttNet 2021.03		69.28	68.33	-
Iterative Shrinking 2021.03		68.1	68.63	-
FAOAVG 2021.03		67.59	61.79	-
RCCF 2021.03		66.73	67.3	-
DGA 2021.03		65.53	65.26	-
SLR 2021.03		64.96	61.25	-
LLaVA-OneVision-7B 2024.09		61.6	-	-
Phi-3-Vision-4B 2024.09		36.1	-	-
MM1-1B 2024.09		0	-	-
MM1-3B 2024.09		0	-	-
MM1-7B 2024.09		0	-	-
MM1-30B 2024.09		0	-	-
Qwen2.5-VL-3B 2026.06		-	-	83
Sequential 2026.06		-	-	84.2
Visual Para-Thinker (3B) 2026.06		-	-	87.7
Visual Para-Thinker++ (3B) 2026.06		-	-	90.4