DynamiCS dynamically scales semantic clusters per training epoch to reduce VLM pre-training compute while improving accuracy on long-tail concepts compared to static or flattening baselines.
OpenVision: A Fully-Open, Cost-Effective Family of Advanced Vision Encoders for Multimodal Learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Dynamic Cluster Data Sampling for Efficient and Long-Tail-Aware Vision-Language Pre-training
DynamiCS dynamically scales semantic clusters per training epoch to reduce VLM pre-training compute while improving accuracy on long-tail concepts compared to static or flattening baselines.