Unverified50% confidenceSolutionExact time
领域适应的标准路径是微调，使用 LoRA 或全量微调对预训练权重进行领域适应，以 WER（词错误率）指标评估效果
1
Sources
50%
Confidence
Long-term
Relevance
7/22/2026
First Seen
Sources
Related Claims
Unverified领域自适应预训练（DAPT）指在通用基础模型上用领域特定语料继续训练，能以较低成本显著提升特定任务性能72% similarUnverified领域自适应作为迁移学习子领域，常用方法包括对抗训练（DANN）、最大均值差异最小化（MMD）以及基于元学习框架（如MAML）的少样本适应72% similarVerifiedLoRA (Low-Rank Adaptation) was proposed by Microsoft in 2021, with the core idea of freezing the weights of the pretrained model and only training injected low-rank decomposition matrices.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/589387API
curl https://kongchang.com/api/v1/knowledge/claims/589387MCP
get_claim(id=589387)