FATE

FATE copied to clipboard

**Describe the bug** Fate v1.7.0 PSI processing(dh, raw), host:200 millions ids, guest:25w ids, both host and guest store data with eggroll 16 partitions ==== detail start, at 20220707.123237.648 ==== Traceback...

JasonBian

Signed-off-by: Xinye Fan [email protected] Fixes ISSUE #3930 Changes: 1. Adding semi-supervised learning component

Answerfxy

failed to connect to all addresses

2

两方 运行 flow model load -c fateflow/examples/model/publish_load_model.json 报错 failed to connect to all addresses retmsg": ""

houjibofa2050

Can we use torch backend in HomoNN?

2

I have noticed the you have implemented the torch module for HomoNN, but the config_type in Component module is fixed as keras. Can we use the torch backend if we...

patricxu

**Describe the bug** When the user wants to upload/bind data from the MySQL database, he can use [table bind](https://federatedai.github.io/FATE-Flow/latest/fate_flow_client/#bind) API. However, the user **must** specify the id name and the...

gxcuit

[FATE-Builder](https://github.com/FederatedAI/FATE-Builder) Goal: - [x] moving code files(https://github.com/FederatedAI/FATE-Builder/pull/1) - [x] Make sure it's available after moving - [ ] remove legacy code

owlet42

关于Fate 中Table这个object的问题

1

目前在看secureboost部分的源码，遇到了一个问题，就是grad_and_hess这个参数的结构是Dtable类型的，Dtable这个类型的操作在哪看，还有就是我在调试的时候提示，Table不支持indexing，那么为什么这个地方可以这么用，不胜感激 

sy0507

1.7.0 issue during debug：`RuntimeError: table not exist:` after adding breakpoints in IDE

6

**Describe the bug** For convenience, I will describe this issue in Chinese. 用IDE（pycharm or vscode）debug的时候，之前（1.6.0）debug某一组件（例如hetero_lr）的方法都是，通过schedule log 里查询相关信息，运行`task_executor.py`，并且指定`task_id、party` 等等。 在1.6.0 版本时是works fine的。 但是在1.7.0版本时，这样做通常也是可以的。通过IDE的run，或者不加断点的debug，都不会有问题，最终的task 都会success。 **但是：如果加上一个断点，这一轮运行也没有问题。但第二次运行`task_executor.py` 时**，就会出现`RuntimeError: table not exist:!` 这个bug。 **To...

gxcuit

**Describe the bug** Hello, an error occurred when using high dimensional data for HeteroPearson algorithm, The dimension of data is 300000 lines and 100 columns. Task is succeed when using...

timber-vine

In some scenarios, it has lots of unlabeled data. If we can use these data for model training, it would be great. Adding semi-supervised learning component has two advantages: 1....

Answerfxy