Zero Shot Transfer Image Classification On 1

评估指标

Accuracy (Private)

评测结果

各个模型在此基准测试上的表现结果

		Paper Title	Repository
M2-Encoder	88.5	M2-Encoder: Advancing Bilingual Image-Text Understanding by Large-scale Efficient Pretraining
BASIC (Lion)	88.3	-	-
CoCa	86.3	CoCa: Contrastive Captioners are Image-Text Foundation Models
LiT-22B	85.9	Scaling Vision Transformers to 22 Billion Parameters
BASIC	85.7	Combined Scaling for Zero-shot Transfer Learning	-
LiT ViT-e	85.4	PaLI: A Jointly-Scaled Multilingual Language-Image Model
LiT-tuning	84.5	LiT: Zero-Shot Transfer with Locked-image text Tuning
IMP-MoE-L	83.9	Alternating Gradient Descent and Mixture-of-Experts for Integrated Multimodal Perception	-
EVA-CLIP-18B	83.8	EVA-CLIP-18B: Scaling CLIP to 18 Billion Parameters
InternVL-C	83.2	InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks
MAWS (ViT-2B)	82.1	The effectiveness of MAE pre-pretraining for billion-scale pretraining
EVA-CLIP-E/14+	82	EVA-CLIP: Improved Training Techniques for CLIP at Scale
CLIPA (ViT-H/14-336px)	81.8	-	-
MAWS (ViT-H)	81.1	The effectiveness of MAE pre-pretraining for billion-scale pretraining
REACT	78.5	Learning Customized Visual Models with Retrieval-Augmented Knowledge
ALIGN	76.4	Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision
CLIP（ViT-L/14-336px）	76.2	Learning Transferable Visual Models From Natural Language Supervision
AltCLIP	74.5	AltCLIP: Altering the Language Encoder in CLIP for Extended Language Capabilities
PaLI	72.11	PaLI: A Jointly-Scaled Multilingual Language-Image Model
Diffusion Classifier (zero-shot)	61.4	Your Diffusion Model is Secretly a Zero-Shot Classifier

0 of 23 row(s) selected.