UME-R1 is a framework designed to endow multimodal embedding models with the flexibility to switch between discriminative and generative embeddings
Zhibin Lan
zhibinlan
AI & ML interests
None yet
Recent Activity
updated
a model
2 days ago
zhibinlan/UME-R1-7B
updated
a model
2 days ago
zhibinlan/UME-R1-2B
updated
a dataset
2 days ago
zhibinlan/UME-sft-train
Organizations
None yet
LLaVE is a series of large language and vision embedding models trained on a variety of multimodal embedding datasets
-
zhibinlan/LLaVE-0.5B
Image-Text-to-Text • 0.9B • Updated • 36 • 7 -
zhibinlan/LLaVE-2B
Image-Text-to-Text • 2B • Updated • 20 • 45 -
zhibinlan/LLaVE-7B
Image-Text-to-Text • 8B • Updated • 254 • 5 -
LLaVE: Large Language and Vision Embedding Models with Hardness-Weighted Contrastive Learning
Paper • 2503.04812 • Published • 15
UME-R1 is a framework designed to endow multimodal embedding models with the flexibility to switch between discriminative and generative embeddings
LLaVE is a series of large language and vision embedding models trained on a variety of multimodal embedding datasets
-
zhibinlan/LLaVE-0.5B
Image-Text-to-Text • 0.9B • Updated • 36 • 7 -
zhibinlan/LLaVE-2B
Image-Text-to-Text • 2B • Updated • 20 • 45 -
zhibinlan/LLaVE-7B
Image-Text-to-Text • 8B • Updated • 254 • 5 -
LLaVE: Large Language and Vision Embedding Models with Hardness-Weighted Contrastive Learning
Paper • 2503.04812 • Published • 15
models
7
zhibinlan/UME-R1-7B
Image-Text-to-Text
•
8B
•
Updated
•
80
•
3
zhibinlan/UME-R1-2B
Image-Text-to-Text
•
2B
•
Updated
•
107
•
3
zhibinlan/LLaVE-7B
Image-Text-to-Text
•
8B
•
Updated
•
254
•
5
zhibinlan/LLaVE-0.5B
Image-Text-to-Text
•
0.9B
•
Updated
•
36
•
7
zhibinlan/LLaVE-2B
Image-Text-to-Text
•
2B
•
Updated
•
20
•
45
zhibinlan/AVG-LLaVA
Image-Text-to-Text
•
7B
•
Updated
•
2
zhibinlan/AVG-LLaVA-Stage3
Image-Text-to-Text
•
7B
•
Updated