FateWithShellmiaoComment/Fate/examples/dsl/v2/hetero_sshe_lr

Hetero SSHE LR Logistic Regression Configuration Usage Guide.

This section introduces the dsl and conf for usage of different types of tasks.

Example Task

1. Train_task: dsl: hetero_lr_normal_dsl.json runtime_config : hetero_lr_normal_conf.json

2. LR Compute Loss: dsl: hetero_lr_compute_loss_dsl.json runtime_config: hetero_lr_compute_loss_conf.json

3. Cross Validation Task(with fold history data output of predict score): dsl: hetero_lr_cv_dsl.json runtime_config: hetero_lr_cv_conf.json

4. One vs Rest(OVR) Task: dsl: hetero_lr_ovr_dsl.json conf: hetero_lr_ovr_conf.json

5. LR with validation: dsl: hetero_lr_with_validate_dsl.json conf: hetero_lr_with_validate_conf.json

6. LR with Warm start task: dsl: hetero_lr_warm_start_dsl.json conf: hetero_lr_warm_start_conf.json

7. LR with Encrypted Reveal in Host task: dsl: hetero_lr_encrypted_reveal_in_host_dsl.json conf: hetero_lr_encrypted_reveal_in_host_conf.json

8. LR L1 penalty task: dsl: hetero_lr_l1_dsl.json conf: hetero_lr_l1_conf.json

9. OVR LR with Encrypted Reveal in Host task: dsl: hetero_lr_ovr_encrypted_reveal_in_host_dsl.json conf: hetero_lr_ovr_encrypted_reveal_in_host_conf.json

10. LR OVR None-penalty task: dsl: hetero_lr_ovr_none_penalty_dsl.json conf: hetero_lr_ovr_none_penalty_conf.json

11. LR OVR L1 penalty task: dsl: hetero_lr_ovr_l1_dsl.json conf: hetero_lr_ovr_l1_conf.json

12. LR with Large Init Weight: dsl: hetero_lr_large_init_w_compute_loss_dsl.json conf: hetero_lr_large_init_w_compute_loss_conf.json

13. LR without intercept: dsl: hetero_lr_normal_not_fit_intercept_dsl.json conf: hetero_lr_normal_not_fit_intercept_conf.json

14. LR Compute Loss without reveal: dsl: hetero_lr_compute_loss_not_reveal_dsl.json conf: hetero_lr_compute_loss_not_reveal_conf.json

15. LR Normal Predict: dsl: hetero_lr_normal_predict_dsl.json conf: hetero_lr_normal_predict_conf.json

16. LR OVR Predict: dsl: hetero_lr_ovr_predict_dsl.json conf: hetero_lr_ovr_predict_conf.json

Users can use following commands to running the task.

flow job submit -c ${runtime_config} -d ${dsl}

After having finished a successful training task, you can use it to predict, you can use the obtained model to perform prediction. You need to add the corresponding model id and model version to the configuration file