KAgBMan issues

Results 9 issues of


                                            KAgBMan

模型发布存在问题

当训练神经网络模型后，通过在guest方deploy-publish-bind操作后，在线推理返回结果无score {"retcode":0,"retmsg":"","data":{"modelId":"guest#9999#arbiter-10000#guest-9999#host-10000#model","modelVersion":"202303150203349569250","timestamp":1678845832116},"flag":0} pipeline只有transform ![image](https://user-images.githubusercontent.com/22344179/225189955-5e0ab19d-5be2-4be0-b621-41078983a0ae.png)

kubefate1.9新增自定义算法

新增的自定义算法相关py文件如何加入到fate_flow容器中？目前将文件添加进去之后，然后重启容器之后，系统会自动再重启并重置容器，新增的文件也会消失

The kubefate update job succeeded, but did not take effect

I want to add a partner by 'kubefate cluster update -f xxx.yaml', but it did not take effect, because the router_table.json has not changed . What can i do now?

How to dynamically expand multiple parties

Now I use the K8S to successfully deploy two parties, but now I need to add a participant for third-party computing, How can I add a participant without changing the...

upload data issues

使用docker部署fate后，使用fateclient调用server上传数据时，可以上传本地文件吗？如果可以，json文件应该怎么编写？

某些模型训练后使用在线推理没有score输出

横向联邦模型或某些纵向联邦模型在完成训练后经过模型的deploy、load、bind后，使用在线推理访问接口获得的反馈中没有score，如下 `{"retcode":0,"retmsg":"","data":{"modelId":"guest#9999#arbiter-10000#guest-9999#host-10000#model","modelVersion":"202303150203349569250","timestamp":1678845832116},"flag":0}` 是否因为在fate-serving-federatedml.jar包中没有开发对应的功能，如下图为包中现有文件 ![image](https://user-images.githubusercontent.com/22344179/225229540-af98a53a-0e50-42f0-87bc-348879ad428d.png)

KAgBMan