List your company

Image Captioning Evaluation on Flickr8k-CF

56.6Kendall-b Correlation (tau_b)ITIScore
23.11231.80640.549.194May 10, 2023Nov 17, 2023May 26, 2024Dec 3, 2024Jun 12, 2025Dec 20, 2025Jun 29, 2026

Evaluation results

MethodLinks
ITIScoreZero-shot=true2026.0456.657.2
PAC-SZero-shot=true2026.0453.954.3
MIDZero-shot=true2026.0453.554.9
ChatGPT-5Zero-shot=true2026.045355.8
SPICEZero-shot=true2026.0451.744.9
CLIP-SZero-shot=true2026.0451.151.2
Gemini-3.0-FlashZero-shot=true2026.0450.852.5
ChatGPT-4oZero-shot=true2026.044849.5
UMICZero-shot=true2026.0447.346.8
Gemini-2.5-FlashZero-shot=true2026.0445.546.8
CIDErZero-shot=true2026.0443.643.9
InternVL3.5 (7B)Zero-shot=true2026.0442.844.1
METEORZero-shot=true2026.0441.541.8
RefDISCODE-INFF=true2025.1240.520.9
RefDISCODE-LVFF=true2025.1240.420.9
RIGELReference Requirement=Reference-based2026.0640.420.9
RIGELReference Requirement=Reference-free2026.0640.420.9
DISCODEBackbone=LLaVA, Reference Requirement=Reference-free2026.0640.220.8
DISCODEBackbone=InternVL, Reference Requirement=Reference-free2026.0640.120.8
BEiTScore-L/14Size (M)=674.69, GFLOPs=296.732026.0539.6—
EXPERTSize (M)=13350.84, GFLOPs=9317.472026.0539.3—
EXPERTReference Requirement=Reference-free2026.0639.3—
HarmonicEvalReference protocol=Reference-free2024.1239.2—
Qwen3-VL (7B)Zero-shot=true2026.0439.140
GPT-FLEURReference protocol=Reference-free2024.1239—
RefPAC-S++ViT-LFF=false2025.1238.8—
RefFLEURFF=true2025.1238.8—
RefFLEURReference protocol=Reference-based2024.1238.8—
Ref-FLEURReference Requirement=Reference-based2026.0638.8—
G-VEvalReference protocol=Reference-free2024.1238.7—
G-VEvalReference Requirement=Reference-free2026.0638.720.2
PearlReference-based=true, Backbone=ViT-B/32, Test time [hour]=0.0432025.1238.620
FLEURReference protocol=Reference-free2024.1238.6—
FLEURSize (M)=13350.84, GFLOPs=9317.472026.0538.6—
PearlReference Requirement=Reference-based2026.0638.620
FLEURReference Requirement=Reference-free2026.0638.6—
BERT-SZero-shot=true2026.0438.539.2
PAC-S++/ViT-LSize (M)=430.58, GFLOPs=56.262026.0538.5—
BEiTScore-L/14 w/o Stage 1Size (M)=674.69, GFLOPs=296.732026.0538.5—
BEiTScore-B/32Size (M)=222.28, GFLOPs=83.852026.0538.5—
DENEB (ViT-L)FF=false2025.1238.3—
CLAIRFF=true2025.1238.2—
G-VEval-C (GPT-4o)FF=true2025.1238.219.9
Ref-HICEScoreReference Requirement=Reference-based2026.0638.219.8
CLAIRReference Requirement=Reference-based2026.0638.217
G-VEvalReference Requirement=Reference-based2026.0638.219.9
DENEBReference Requirement=Reference-based2026.063819.6
RefPAC-S++ViT-B2025.1237.919.6
RefPAC-S++Reference Requirement=Reference-based2026.0637.919.6
PearlReference-based=false, Backbone=ViT-B/32, Test time [hour]=0.0432025.1237.819.5
PolosFF=false2025.1237.8—
PolosReference protocol=Reference-based2024.1237.8—
BEiTScore-B/32 w/o Stage 1Size (M)=222.28, GFLOPs=83.852026.0537.8—
PolosReference Requirement=Reference-based2026.0637.819.5
PearlReference Requirement=Reference-free2026.0637.819.5
RefPAC-S2025.1237.619.5
RefPAC-SReference Requirement=Reference-based2026.0637.619.5
MIDFF=true2025.1237.3—
HICEScoreReference Requirement=Reference-free2026.0637.219.2
PAC-S/ViT-LSize (M)=427.62, GFLOPs=56.262026.0537—
PAC-S++/ViT-BSize (M)=155.42, GFLOPs=4.882026.0537—
PAC-S++Reference Requirement=Reference-free2026.063719.1
BLIP2-ScoreReference Requirement=Reference-free2026.0636.719
CLIP-StuneType=w/o ref2023.0536.6—
InfoMetIC+Type=w/o ref2023.0536.6—
InfoMetIC+Reference protocol=Reference-free2024.1236.6—
VQAScore (CLIP-FlanT5-XXL)Size (M)=11459.59, GFLOPs=3742.12026.0536.5—
CLIP-S/ViT-LSize (M)=427.62, GFLOPs=84.362026.0536.5—
RefCLIP-SType=w/ ref2023.0536.4—
RefCLIP-SFF=true2025.1236.418.8
RefCLIP-ScoreReference protocol=Reference-based2024.1236.4—
RefCLIP-SReference Requirement=Reference-based2026.0636.418.8
InfoMetICType=w/o ref2023.0536.3—
BRIDGEReference-based=false2025.1236.319
InfoMetICReference protocol=Reference-free2024.1236.3—
BRIDGEReference Requirement=Reference-free2026.0636.319
PAC-S/ViT-BSize (M)=151.28, GFLOPs=4.882026.0536—
PAC-SReference Requirement=Reference-free2026.063618.6
LongCLIP-S/ViT-LSize (M)=427.75, GFLOPs=98.92026.0535.9—
FAIEr-4Type=w/ ref2023.0535.4—
FAIEr-4Reference protocol=Reference-based2024.1235.4—
CLIP-S/ViT-BSize (M)=151.28, GFLOPs=7.282026.0535.3—
LongCLIP-S/ViT-BSize (M)=149.84, GFLOPs=26.242026.0535.2—
CLIP-SType=w/o ref2023.0534.4—
CLIP-SReference protocol=Reference-free2024.1234.4—
CLIP-SReference Requirement=Reference-free2026.0634.417.7
FAIEr-rType=w/o ref2023.0532.4—
FAIEr-rReference protocol=Reference-free2024.1232.4—
BLEUZero-shot=true2026.0432.232.3
ROUGEZero-shot=true2026.0432.132.3
UMICReference protocol=Reference-free2024.1230.1—
LEIC2025.1229.5—
InfoCLIPtuneType=w/o ref2023.0527.7—
SPARCSReference-based=true, Test time [hour]=< 0.012025.1225.213
CIDErType=w/ ref2023.0524.6—
CIDErFF=false2025.1224.612.7
CIDErReference protocol=Reference-based2024.1224.6—
CIDErReference Requirement=Reference-based2026.0624.612.7
SPICEType=w/ ref2023.0524.4—
SPICEFF=false2025.1224.412