List your company

Image Captioning Evaluation on Nebula

57.5Kendall tau_cRIGEL
33.47639.71345.9552.187Dec 25, 2025Jan 25, 2026Feb 25, 2026Mar 28, 2026Apr 28, 2026May 29, 2026Jun 29, 2026

Evaluation results

MethodLinks
RIGELReference Requirement=Reference-based2026.0657.560.6
BEiTScore-L/14Size (M)=674.69, GFLOPs=296.732026.0556.7—
BEiTScore-L/14 w/o Stage 1Size (M)=674.69, GFLOPs=296.732026.0556.3—
PearlReference-based=true, Backbone=ViT-B/32, Test time [hour]=0.0432025.1255.458.4
PearlReference-based=true, Backbone=ViT-B/322025.1255.458.4
PearlReference Requirement=Reference-based2026.0655.458.4
BEiTScore-B/32Size (M)=222.28, GFLOPs=83.852026.0555.3—
DENEBReference-based=true, Backbone=ViT-B/322025.1255.158.1
DENEBReference Requirement=Reference-based2026.0655.158.1
RIGELReference Requirement=Reference-free2026.0655.158.1
PolosReference-based=true2025.125558
PolosReference Requirement=Reference-based2026.065558
EXPERTReference-based=false2025.1254.9—
EXPERTSize (M)=13350.84, GFLOPs=9317.472026.0554.9—
BEiTScore-B/32 w/o Stage 1Size (M)=222.28, GFLOPs=83.852026.0554.6—
PommsEvaluation Framework=Pomms, Ensemble Composition=(v) + Qwen2.5-VL-7B2026.0454.157
PommsEvaluation Framework=Pomms, Ensemble Composition=(vi) + Phi-3.5-Vision2026.0454.157
PommsEvaluation Framework=Pomms, Ensemble Composition=(iii) + LLaVA-OneVision-7B2026.0453.756.6
PommsEvaluation Framework=Pomms, Ensemble Composition=(ii) + Eagle2-9B2026.0453.656.6
PommsEvaluation Framework=Pomms, Ensemble Composition=(i) GPT-4o and InternVL2.5-8B2026.0453.556.4
PommsEvaluation Framework=Pomms, Ensemble Composition=(iv) + DeepSeek-VL22026.0453.556.4
GPT-4oEvaluation Framework=G-VEval2026.0453.256.1
PearlReference-based=false, Backbone=ViT-B/32, Test time [hour]=0.0432025.125355.9
PearlReference-based=false, Backbone=ViT-B/322025.125355.9
PearlReference Requirement=Reference-free2026.065355.9
FLEURSize (M)=13350.84, GFLOPs=9317.472026.0552.7—
Qwen2.5-VL-7BEvaluation Framework=G-VEval2026.0452.455.3
RefPAC-SReference-based=true, Backbone=ViT-B/322025.1251.954.7
RefPAC-SReference Requirement=Reference-based2026.0651.954.7
InternVL2.5-8BEvaluation Framework=G-VEval2026.0451.354.1
RefCLIP-SReference-based=true, Backbone=ViT-B/322025.1250.853.6
RefCLIP-SReference Requirement=Reference-based2026.0650.853.6
BLIP2ScoreReference-based=false2025.1250.753
BLIP2-ScoreReference Requirement=Reference-free2026.0650.753
RefPAC-S++Reference-based=true, Backbone=ViT-B/322025.1250.653.3
RefPAC-S++Reference Requirement=Reference-based2026.0650.653.3
SPARCSReference-based=true, Test time [hour]=< 0.012025.125053.9
VQAScore (CLIP-FlanT5-XXL)Size (M)=11459.59, GFLOPs=3742.12026.0549.4—
LongCLIP-S/ViT-BSize (M)=149.84, GFLOPs=26.242026.0548.9—
PAC-S++/ViT-LSize (M)=430.58, GFLOPs=56.262026.0548.9—
CIDErReference-based=true2025.1248.851.5
CLIP-S/ViT-LSize (M)=427.62, GFLOPs=84.362026.0548.8—
CIDErReference Requirement=Reference-based2026.0648.851.5
LongCLIP-S/ViT-LSize (M)=427.75, GFLOPs=98.92026.0548.6—
PAC-SReference-based=false, Backbone=ViT-B/322025.1248.351
PAC-SReference Requirement=Reference-free2026.0648.351
CLIP-SReference-based=false2025.1247.950.5
PAC-S++Reference-based=false, Backbone=ViT-B/322025.1247.950.5
PAC-S/ViT-LSize (M)=427.62, GFLOPs=56.262026.0547.9—
CLIP-SReference Requirement=Reference-free2026.0647.950.5
PAC-S++Reference Requirement=Reference-free2026.0647.950.5
CLIP-S/ViT-BSize (M)=151.28, GFLOPs=7.282026.0547.8—
METEORReference-based=true2025.1247.650.2
METEORReference Requirement=Reference-based2026.0647.650.2
SPICEReference-based=true2025.1247.451.5
SPICEReference Requirement=Reference-based2026.0647.451.5
PAC-S/ViT-BSize (M)=151.28, GFLOPs=4.882026.0547.2—
BERTScoreReference-based=true, Test time [hour]=0.0402025.1247.147.5
BERTScoreReference Requirement=Reference-based2026.0647.147.5
PAC-S++/ViT-BSize (M)=155.42, GFLOPs=4.882026.0546.9—
BARTScoreReference-based=true, Test time [hour]=0.682025.124547.4
BLEUReference-based=true2025.1244.146.5
BLEUReference Requirement=Reference-based2026.0644.146.5
ROUGEReference-based=true2025.1243.445.8
ROUGEReference Requirement=Reference-based2026.0643.445.8
SPECSSize (M)=151.49, GFLOPs=9.22026.0534.4—