1 回答
- 最新
- 投票最多
- 评论最多
0
If you are referring to using local mode via the SM PySDK, then pipeline deployment is not supported.
As an alternative, given your three inference containers, you could manually run the services locally and then implement a kind of facade function that invokes the three services in pipeline and manages input/output accordingly.
相关内容
- AWS 官方已更新 3 年前