Tag: 微调
All the articles with the tag "微调".
- 6.0
Weight space Detection of Backdoors in LoRA Adapters
arXiv:2602.15195v3 Announce Type: replace-cross Abstract: LoRA adapters let users fine-tune large language models (LLMs) efficiently. However, LoRA...
- 6.0
Sim-CLIP: Unsupervised Siamese Adversarial Fine-Tuning for Robust and Semantically-Rich Vision-Language Models
arXiv:2407.14971v3 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) rely heavily on pretrained vision encoders to support downs...
- 7.0
UI-Voyager:通过失败经验自演化的移动端 GUI Agent
两阶段自演化框架,用拒绝微调和路径级信用分配持续改进 GUI 操作能力
- 7.7
Scaling DoRA:分解范数与融合内核让微调加速 2 倍、显存降低 7GB
通过数学重构和内核融合优化 DoRA 微调方法,在 8-32B VLM 上实现显著性能提升
- 7.6
NVIDIA 发布领域 Embedding 微调全流程:单 GPU 一天搞定
NVIDIA 官方博客公开端到端 Embedding 微调方案,覆盖合成数据生成、难负例挖掘、训练与部署全链路