知识卡片
定制ML模型也是套餐路径
传统机器学习也可用共享基础模型服务普通租户,为高级租户训练独立模型或提供专用推理。模型训练、部署和消费是另一类可定价服务能力;选择生成式AI还是定制模型,应该由任务与数据需求决定,而非只跟随技术热度。
这张卡片看不懂?
这张卡片的反馈已满 10 条,暂时无法继续提交。
已有反馈 · 0 / 10 条
还没有人提交反馈。
在整书笔记中的位置
原书出处 · 1 处
- GenAI and Multi-Tenancy · 查看原书依据
throughout this book. I can, for example, have a single, pooled ML model that offers a baseline experience to my basic tier tenants. These tenants all share some pre-packaged model with no support for tenant specialization. Then, for my premium tier, I could offer a more custom experience where they would upload their training data and create a tenant-specific ML
参考来源
- 位置:《Building Multi-Tenant SaaS Architectures》第16章《GenAI and Multi-Tenancy》(来源项目 a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Building Multi-Tenant SaaS Architectures/_epub-src/OEBPS/ch16.html)
- 结论依据:原文具体说明传统机器学习也可用共享基础模型服务普通租户,为高级租户训练独立模型或提供专用推理;本卡据此归纳机制及适用边界。
- 原始内容:throughout this book. I can, for example, have a single, pooled ML model that offers a baseline experience to my basic tier tenants. These tenants all share some pre-packaged model with no support for tenant specialization. Then, for my premium tier, I could offer a more custom experience where they would upload their training data and create a tenant-specific ML