Position: Pause Recycling LoRAs and Prioritize Mechanisms to Uncover Limits and Effectiveness

Documenti analoghi

• Mixture-of-LoRAs: An Efficient Multitask Tuning for Large Language Models
  di: Feng, Wenfeng, et al.
  Pubblicazione: (2024)
• Rethinking Parameter Sharing for LLM Fine-Tuning with Multiple LoRAs
  di: Ban, Hao, et al.
  Pubblicazione: (2025)
• Merging LoRAs like Playing LEGO: Pushing the Modularity of LoRA to Extremes Through Rank-Wise Clustering
  di: Zhao, Ziyu, et al.
  Pubblicazione: (2024)
• MoSLD: An Extremely Parameter-Efficient Mixture-of-Shared LoRAs for Multi-Task Learning
  di: Zhao, Lulu, et al.
  Pubblicazione: (2024)
• Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs
  di: Microsoft, et al.
  Pubblicazione: (2025)