Zero Shot Composed Image Retrieval Zs Cir On 2

Metrics

(Recall@10+Recall@50)/2

Results

Performance results of various models on this benchmark

Model Name	(Recall@10+Recall@50)/2	Paper Title	Repository
RTD + LinCIR (CLIP G/14)	56.74	An Efficient Post-hoc Framework for Reducing Task Discrepancy of Text Encoders for Composed Image Retrieval	-
WeiMoCIR (CLIP G/14)	47.16	Training-free Zero-shot Composed Image Retrieval via Weighted Modality Fusion and Similarity	-
SEARLE (CLIP B/32)	32.71	Zero-Shot Composed Image Retrieval with Textual Inversion	-
SEARLE-XL-OTI (CLIP L/14)	37.76	Zero-Shot Composed Image Retrieval with Textual Inversion	-
LinCIR (CLIP G/14)	55.40	Language-only Efficient Training of Zero-shot Composed Image Retrieval	-
iSEARLE-XL (CLIP L/14)	38.24	iSEARLE: Improving Textual Inversion for Zero-Shot Composed Image Retrieval	-
CompoDiff (CLIP G/14)	45.37	CompoDiff: Versatile Composed Image Retrieval With Latent Diffusion	-
SEARLE-XL (CLIP L/14)	35.90	Zero-Shot Composed Image Retrieval with Textual Inversion	-
Context-I2W (CLIP L/14)	38.35	Context-I2W: Mapping Images to Context-dependent Words for Accurate Zero-Shot Composed Image Retrieval	-
WeiMoCIR (CLIP H/14)	44.58	Training-free Zero-shot Composed Image Retrieval via Weighted Modality Fusion and Similarity	-
MagicLens (CoCa B)	45.3	MagicLens: Self-Supervised Image Retrieval with Open-Ended Instructions	-
WeiMoCIR (CLIP L/14)	41.27	Training-free Zero-shot Composed Image Retrieval via Weighted Modality Fusion and Similarity	-
WeiMoCIR (CLIP B/32)	39.84	Training-free Zero-shot Composed Image Retrieval via Weighted Modality Fusion and Similarity	-
MagicLens (CoCa L)	48.1	MagicLens: Self-Supervised Image Retrieval with Open-Ended Instructions	-
CIReVL (CLIP L/14)	38.56	Vision-by-Language for Training-Free Compositional Image Retrieval	-
CoLLM (Pretrained - CLIP-L/14)	39.8	CoLLM: A Large Language Model for Composed Image Retrieval	-
OSrCIR (CLIP B/32)	42.87	Reason-before-Retrieve: One-Stage Reflective Chain-of-Thoughts for Training-Free Zero-Shot Composed Image Retrieval	-
PALAVRA	28.51	"This is my unicorn, Fluffy": Personalizing frozen vision-language representations	-
CoVR-BLIP-2	48.3	CoVR-2: Automatic Data Construction for Composed Video Retrieval	-
OSrCIR (CLIP G/14)	47.34	Reason-before-Retrieve: One-Stage Reflective Chain-of-Thoughts for Training-Free Zero-Shot Composed Image Retrieval	-

0 of 40 row(s) selected.