Papers with Code · CC BY-SA 4.0

Winoground

아카이브 평가 114건 · 마지막 적재/변경 2026-10-30

평가 기록

과거 수집 기록이며 최신 순위가 아닙니다. 데이터셋 버전·split·실험 조건과 지표 방향이 검증되지 않은 결과는 직접 비교할 수 없습니다. 날짜는 논문 날짜이며 최신 평가일을 뜻하지 않습니다.

표가 화면보다 넓으면 좌우로 스크롤하여 논문과 구현 링크를 확인하세요.

원본 기록 순서 · 순위 미부여
모델	원본 지표·값	논문	구현
CLIP (SGVL)	Text Score 32.0 Image Score 14.0 Group Score 9.8	Incorporating Structured Representations into Pretrained Vision & Language Models Using Scene Graphs 2023-05-10	연결 없음
ViT-B/16 + BERT base	Text Score 31.2	ViLEM: Visual-Language Error Modeling for Image-Text Retrieval 2023-01-01	연결 없음
Gemini	Text Score 30.75 Image Score 26 Group Score 25	CoCoT: Contrastive Chain-of-Thought Prompting for Large Multimodal Models with Multiple Image Inputs 2024-01-05	코드 1
OCLIP (ViT-H/14)	Text Score 30.75 Image Score 12.75	SelfEval: Leveraging the discriminative nature of generative models for evaluation 2023-11-17	연결 없음
CLIP (ViT-B/32)	Text Score 30.75 Image Score 10.50 Group Score 8.00	Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality 2022-04-07	코드 1 · 코드 2
OFA large (ITM)	Text Score 30.75 Image Score 10.25 Group Score 7.25	Simple Token-Level Confidence Improves Caption Correctness 2023-05-11	연결 없음
KeyComp (GPT-3.5)	Text Score 30.3 Image Score 24.6 Group Score 12.4	Prompting Large Vision-Language Models for Compositional Reasoning 2024-01-20	코드 1
CLIP (ViT-L/14)	Text Score 30.25 Image Score 8.0	SelfEval: Leveraging the discriminative nature of generative models for evaluation 2023-11-17	연결 없음
ViLLA base	Text Score 30.00 Image Score 12.00 Group Score 8.00	Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality 2022-04-07	코드 1 · 코드 2
syn-CLIP	Text Score 30.00 Image Score 11.50 Group Score 9.50	Going Beyond Nouns With Vision & Language Models Using Synthetic Data 2023-03-30	코드 1
syn-CyCLIP	Text Score 30.00 Image Score 10.75 Group Score 8.25	Going Beyond Nouns With Vision & Language Models Using Synthetic Data 2023-03-30	코드 1
NegCLIP	Text Score 29.5 Image Score 10.5 Group Score 8.0	Incorporating Structured Representations into Pretrained Vision & Language Models Using Scene Graphs 2023-05-10	연결 없음
OFA large (TLC-A)	Text Score 29.25 Image Score 27.00 Group Score 17.50	Simple Token-Level Confidence Improves Caption Correctness 2023-05-11	연결 없음
ALBEF 4M	Text Score 29.2 Image Score 15.5 Group Score 11.0	Measuring Progress in Fine-grained Vision-and-Language Understanding 2023-05-12	코드 1 · 코드 2
LDM-T5 (SelfEval)	Text Score 29.00 Image Score 13.50	SelfEval: Leveraging the discriminative nature of generative models for evaluation 2023-11-17	연결 없음
CyCLIP	Text Score 28.50 Image Score 9.50 Group Score 7.25	Going Beyond Nouns With Vision & Language Models Using Synthetic Data 2023-03-30	코드 1
PDM-T5 (SelfEval)	Text Score 28.25 Image Score 12.00	SelfEval: Leveraging the discriminative nature of generative models for evaluation 2023-11-17	연결 없음
COCA ViT-L14 (f.t on COCO)	Text Score 28.25 Image Score 11.50 Group Score 8.25	What You See is What You Read? Improving Text-Image Alignment Evaluation 2023-05-17	코드 1
LLaVA-1.5-ZS-CoT	Text Score 28.0 Image Score 22.5 Group Score 12.3	Compositional Chain-of-Thought Prompting for Large Multimodal Models 2023-11-27	코드 1
BLIP (ITC)	Text Score 28.0 Image Score 9.0 Group Score 6.5	Revisiting the Role of Language Priors in Vision-Language Models 2023-06-02	코드 1
OFA large (ft SNLI-VE)	Text Score 27.70 Image Score 14.30 Group Score 9.00	What You See is What You Read? Improving Text-Image Alignment Evaluation 2023-05-17	코드 1
OFA base (ITM)	Text Score 26.75 Image Score 10.75 Group Score 6.50	Simple Token-Level Confidence Improves Caption Correctness 2023-05-11	연결 없음
CLIP RN50x64	Text Score 26.50 Image Score 13.75 Group Score 10.25	What You See is What You Read? Improving Text-Image Alignment Evaluation 2023-05-17	코드 1
LLaVA-7B (GPTScore)	Text Score 25.50 Image Score 17.00 Group Score 10.50	An Examination of the Compositionality of Large Generative Vision-Language Models 2023-08-21	코드 1
FLAVA (contrastive)	Text Score 25.25 Image Score 13.50 Group Score 9.00	Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality 2022-04-07	코드 1 · 코드 2
Random chance	Text Score 25.00 Image Score 25.00 Group Score 16.67	Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality 2022-04-07	코드 1 · 코드 2
LLaVA	Text Score 24.8 Image Score 25.0 Group Score 13.0	Incorporating Structured Representations into Pretrained Vision & Language Models Using Scene Graphs 2023-05-10	연결 없음
OFA base (TLC-A)	Text Score 24.50 Image Score 23.50 Group Score 13.75	Simple Token-Level Confidence Improves Caption Correctness 2023-05-11	연결 없음
MiniGPT-4-7B (GPTScore)	Text Score 24.50 Image Score 21.75 Group Score 11.50	An Examination of the Compositionality of Large Generative Vision-Language Models 2023-08-21	코드 1
ViLBERT base	Text Score 23.75 Image Score 7.25 Group Score 4.75	Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality 2022-04-07	코드 1 · 코드 2