HyperAI초신경

홈 뉴스 연구 논문 튜토리얼 데이터셋 백과사전 SOTA LLM 모델 GPU 랭킹 컨퍼런스

한국어

HyperAI초신경

Image Retrieval On Flickr30K

평가 지표

Recall@1

Recall@10

Recall@5

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름	Recall@1	Recall@10	Recall@5	Paper Title	Repository
HADA	81.36	98.02	95.94	HADA: A Graph-based Amalgamation Framework in Image-text Retrieval
LGSGM	57.4	90.2	84.1	A Deep Local and Global Scene-Graph Matching for Image-Text Retrieval
BLIP-2 ViT-L (zero-shot, 1K test set)	88.6	98.9	97.6	BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models
MaMMUT (ours)	82.5	98	96	MaMMUT: A Simple Architecture for Joint Learning for MultiModal Tasks
VisualSparta	57.4	88.1	82.0	VisualSparta: An Embarrassingly Simple Approach to Large-scale Text-to-Image Search with Weighted Bag-of-words
ALBEF	79.76	97.72	95.3	HADA: A Graph-based Amalgamation Framework in Image-text Retrieval
UNITER	75.56	96.76	94.08	HADA: A Graph-based Amalgamation Framework in Image-text Retrieval
BLIP-2 ViT-G (zero-shot, 1K test set)	89.7	98.9	98.1	BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models
GSMN	-	89	82.3	A Deep Local and Global Scene-Graph Matching for Image-Text Retrieval

0 of 9 row(s) selected.