FLoRA: Fused forward-backward adapters for parameter efficient fine-tuning and reducing inference-time latencies of LLMs
"参数高效" 相关文章
Hierarchical LoRA MoE for Efficient CTR Model Scaling
VER: Vision Expert Transformer for Robot Learning via Foundation Distillation and Dynamic Routing
QuIC: Quantum-Inspired Compound Adapters for Parameter Efficient Fine-Tuning
LLM工业级自进化：北邮与腾讯AI Lab提出MoE-CL架构，解决大模型持续学习核心痛点
Stable Forgetting: Bounded Parameter-Efficient Unlearning in LLMs
Exploring Data and Parameter Efficient Strategies for Arabic Dialect Identifications
Cross-Modal Adapter for Vision-Language Retrieval
MoE也有Scaling Law，「百万专家」利用率近100%！DeepMind华人挑战MoE极限
Samsung Researchers Introduce LoRA-Guard: A Parameter-Efficient Guardrail Adaptation Method that Relies on Knowledge Sharing between LLMs and Guardrail Models
