Different Prompts, Different Ranks: Prompt-aware Dynamic Rank Selection for SVD-based LLM Compression

Documenti analoghi

• LLM-NAS: LLM-driven Hardware-Aware Neural Architecture Search
  di: Zhu, Hengyi, et al.
  Pubblicazione: (2025)
• IO-SVD: Input-Output Whitened SVD for Adaptive-Rank LLM Compression
  di: Abbasi, Ali, et al.
  Pubblicazione: (2026)
• Layer-wise dynamic rank for compressing large language models
  di: Mi, Zhendong, et al.
  Pubblicazione: (2025)
• Towards Fast LLM Fine-tuning through Zeroth-Order Optimization with Projected Gradient-Aligned Perturbations
  di: Mi, Zhendong, et al.
  Pubblicazione: (2025)
• Zero Sum SVD: Balancing Loss Sensitivity for Low Rank LLM Compression
  di: Abbasi, Ali, et al.
  Pubblicazione: (2026)