InfLLM: Training-Free Long-Context Extrapolation for LLMs with an Efficient Context Memory

Documenti analoghi

• InfLLM-V2: Dense-Sparse Switchable Attention for Seamless Short-to-Long Adaptation
  di: Zhao, Weilin, et al.
  Pubblicazione: (2025)
• Variator: Accelerating Pre-trained Models with Plug-and-Play Compression Modules
  di: Xiao, Chaojun, et al.
  Pubblicazione: (2023)
• Locret: Enhancing Eviction in Long-Context LLM Inference with Trained Retaining Heads on Consumer-Grade Devices
  di: Huang, Yuxiang, et al.
  Pubblicazione: (2024)
• Exploring the Benefit of Activation Sparsity in Pre-training
  di: Zhang, Zhengyan, et al.
  Pubblicazione: (2024)
• APB: Accelerating Distributed Long-Context Inference by Passing Compressed Context Blocks across GPUs
  di: Huang, Yuxiang, et al.
  Pubblicazione: (2025)