Accueil Actualités Articles de recherche récents Tutoriels Ensembles de données Wiki SOTA Modèles LLM Classement GPU Événements

Français

Image Captioning On Nocaps Val Near Domain

Métriques

CIDEr

Pre-train (#images)

SPICE

Résultats

Résultats de performance de divers modèles sur ce benchmark

Nom du modèle	CIDEr	Pre-train (#images)	SPICE	Paper Title	Repository
BLIP-2 ViT-G OPT 6.7B (zero-shot)	119.2	1.1B	15.3	BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models
OmniVL	108.3	14M	14.9	OmniVL:One Foundation Model for Image-Language and Video-Language Tasks	-
VinVL	96.1	5.7M	13.8	VinVL: Revisiting Visual Representations in Vision-Language Models
BLIP-2 ViT-G FlanT5 XL (zero-shot)	120.2	1.1B	15.9	BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models
Enc-Dec	88.3	-	12.1	Conceptual 12M: Pushing Web-Scale Image-Text Pre-Training To Recognize Long-Tail Visual Concepts
BLIP_ViT-L	112.1	129M	14.9	BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation
LEMON_large	113.3	200M	15.1	Scaling Up Vision-Language Pre-training for Image Captioning	-
SimVLM	110.9	1.8B	-	SimVLM: Simple Visual Language Model Pretraining with Weak Supervision
BLIP_CapFilt-L	108.6	129M	14.8	BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation
BLIP-2 ViT-G OPT 2.7B (zero-shot)	117.8	1.1B	15.4	BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models

0 of 10 row(s) selected.