Papers

Filtered to gpu memory optimization · clear filter

Browse by term

continual learning 64reinforcement learning 34large language models 17benchmarking 12vision-language models 10generative models 8language models 8video generation 7multimodal models 6natural language processing 6robotics 6world models 6benchmarks 5diffusion models 5on-policy distillation 5policy optimization 5scalability 5self-distillation 5vision-language-action models 5autoregressive models 4computer vision 4diffusion transformers 4LLMs 4multimodal large language models 4verifiable rewards 4attention mechanisms 3embodied intelligence 3image editing 3long-term memory 3multimodal learning 3

Matching papers

OmniDelta: Skill-Driven Budget Allocation for Token Compression in OmniLLMs

8 upvotes · 28 JUL 2026 · Haoyang Huang, Wenjie Huang, Tianqi Xu et al.

This paper proposes a method to compress large language models by allocating a fixed budget to select the most important tokens, allowing for efficient inference and reduced memory usage. Practitioners may care about this research if they work on optimizing the performance of large language models for real-world applications.

continual learningtoken compressionomnilarge language modelsinference efficiencygpu memory optimization