How Far Are Video Models from True Multimodal Reasoning?
Evaluating video models for true multimodal reasoning.
x.-zhang
W. Xu
•
•
1 min read
ROSE: A Reward-Oriented Data Selection Framework for LLM Task-Specific Instruction Tuning
Reward-oriented data selection for task-specific LLM instruction tuning.
y.-wu
KVDirect: Distributed Disaggregated LLM Inference
Distributed disaggregated inference for efficient LLM serving.
s.-chen