Skip to content

Latest commit

 

History

History
765 lines (650 loc) · 23.4 KB

dsl_conf_v2_setting_guide.md

File metadata and controls

765 lines (650 loc) · 23.4 KB

DSL & Task Submit Runtime Conf Setting V2

[中文]

To make the modeling task more flexible, currently, FATE uses its own domain-specific language(DSL) to describe modeling task. With usage of this DSL, modeling components such as data-transform, feature-engineering and classification/regression module etc. can be combined as a Directed Acyclic Graph(DAG). Therefore, user can take and combine the algorithm components flexibly according to their needs.

In addition, parameters of each component need to be configured. Also, the configuration may vary from party to party. For convenience, FATE configure all parameters for all parties and all components in one file. This guide will show you how to create such a configure file.

Starting at FATE-1.5.0, V2 of dsl and submit conf is recommended.

DSL Configure File

1. Overview

We use json file which is actually a dictionary as a dsl config file.

2. Components

  • definition: components in your modeling task, always the first level of dsl dict.

  • example:

    {
      "components" : {
        ...
      }
    }
  • explanation:

    Then each component should be defined on the second level. Here is an example of setting a component:

    "data_transform_0": {
          "module": "DataTransform",
          "input": {
              "data": {
                  "data": [
                      "reader_0.train_data"
                  ]
              }
          },
          "output": {
              "data": ["train"],
              "model": ["model"]
          }
      }

    As the example shows, user define the component name as key of this module.

    Please note that in DSL V2, all modeling task config should contain a Reader component to reader data from storage service, this component has "output" field only, like the following:

    "reader_0": {
          "module": "Reader",
          "output": {
              "data": ["train"]
          }
    }

3. Module

  • definition: Specify which component to use.

  • explanation: This field should strictly match the ComponentMeta define in python file under the fold

  • example:

    "hetero_feature_binning_1": {
        "module": "HeteroFeatureBinning",
        ...
    }

4. Input

  • definition: There are two types of input, data and model.

4.1 Data Input

  • definition: Data input from previous modules; there are four possible data_input type:
    1. data: typically used in data_transform, feature_engineering modules and evaluation.
    2. train_data: uses in training components like HeteroLR、HeteroSBT and so on. If this field is provided, the task will be parse as a fit task
    3. validate_data: If train_data is provided, this field is optional. In this case, this data will be used as validation set.
    4. test_data: specify the data used to predict, if this field is set up, the model also needs.

4.2 Model Input

  • definition: Model input from previous modules; there are two possible model-input types:

    1. model: This is a model input by the same type of component. For example, hetero_binning_0 run as a fit component, and hetero_binning_1 takes model output of hetero_binning_0 as input so that can be used to transform or predict. Here's an example showing this logic:

      "hetero_feature_binning_1": {
          "module": "HeteroFeatureBinning",
          "input": {
              "data": {
                  "data": [
                      "data_transform_1.validate_data"
                  ]
              },
              "model": [
                  "hetero_feature_binning_0.fit_model"
              ]
          },
          "output": {
              "data": ["validate_data"],
              "model": ["eval_model"]
          }
      }
    2. isometric_model: This is used to specify the model input from upstream components. For example, feature selection will take feature binning as upstream model, since it will use information value as feature importance. Here's an example of feature selection component:

      "hetero_feature_selection_0": {
          "module": "HeteroFeatureSelection",
          "input": {
              "data": {
                  "data": [
                      "hetero_feature_binning_0.train"
                  ]
              },
              "isometric_model": [
                  "hetero_feature_binning_0.output_model"
              ]
          },
          "output": {
              "data": ["train"],
              "model": ["output_model"]
          }
      }

4.3 Model Output

  • definition: Same as input, two types of output may occur: which are data and model.

5.1 Data Output

  • definition: data output, there are four types:

    1. data: normal data output
    2. train_data: only for Data Split
    3. validate_data: only for Data Split
    4. test_data:only for Data Split

5.2 Model Output

  • definition: model output, only use model

JOB RUNTIME CONFIG Guide (for version 1.5.x and above)

1. Overview

Job Runtime Conf configures job and module settings for all participants. Configurable values include:

2. DSL version

  • definition: conf version, default 1, 2 must be set if fate's version >= 1.7

  • example:

    "dsl_version": 2

3. Job Participants

3.1 Initiator

  • definition: role and party_id of job initiator

  • example:

    "initiator": {
        "role": "guest",
        "party_id": 9999
    }

3.2 Role

  • definition: Information on all participants

  • explanation: each key-value pair in role represents a role type and corresponding party ids; party_id should be specified as list since multiple parties may take the same role in a job

  • examples

    "role": {
        "guest": [9999],
        "host": [10000],
        "arbiter": [10000]
    }

4. System Runtime Parameters

  • definition: main system configuration when running jobs

4.1 Configuration Applicable Range Policy

  • common: applies to all participants

  • role: applies only to specific participant; specify participant in (role:)party_index format; note that role configuration takes priority over common

    "common": {
    }
    
    "role": {
      "guest": {
        "0": {
        }
      }
    }

In the example above, configuration insidecommon applies to all participants; configuration inside role-guest-0 only applies to participant guest_0

Note: current version does not perform strict checking on role-specific runtime parameters; common is suggested for setting runtime configuration

4.2 Configurable Job Parameters

Parameter Name Default Value Acceptable Values Information
inheritance_info job_id, component_list jobid and component list of job that need to be inherited
job_type train train, predict job type
task_cores 4 positive integer total cpu cores requested
task_parallelism 1 positive int maximum number of tasks allowed to run in parallel
computing_partitions same as task_cores positive integer partition number for table computing
eggroll_run processors_per_node configuration specific for EGGROLL computing engine; generally set automatically based on task_cores; if specified, task_cores value ineffective
spark_run num-executors, executor-cores configuration specific for SPARK computing engine; generally set automatically based on task_cores; if specified, task_cores value ineffective
rabbitmq_run queue, exchange etc. parameters for rabbitmq to set up queue, exchange, etc.; generally takes system default
federated_status_collect_type PUSH PUSH, PULL way to collect federated job status; PUSH: participants report to initiator, PULL: initiator regularly queries from all participants
timeout 259200 (3 days) positive int time elapse (in second) for a job to timeout
model_id - - id of model, needed for prediction task
model_version - - version of model, needed for prediction task

Configurable Job Parameters

!!! Note

1. Some types of `computing_engine`, `storage_engine` are only compatible with each other. 
2. Developer may implement other types of engines and set new engine combinations in runtime conf.

4.3 Non-Configurable Job Parameters

Parameter Name Default Value Acceptable Values Information
computing_engine Configure in service_conf.yaml EGGROLL, SPARK, STANDALONE engine for computation
storage_engine Configure in service_conf.yaml EGGROLL, HDFS, STANDALONE engine for storage
federation_engine Configure in service_conf.yaml EGGROLL, RABBITMQ, STANDALONE, PULSAR engine for communication among parties
federated_mode set automatically based on federation_engine SINGLE, MULTIPLE federation mode

Non-configurable Job Parameters

4.4 Example Job Parameter Configuration

  1. EGGROLL conf example with default CPU settings:

    "job_parameters": {
      "common": {
          "task_cores": 4
      }
    }
  2. EGGROLL conf example with manually specified CPU settings:

    "job_parameters": {
      "common": {
          "job_type": "train",
          "eggroll_run": {
            "eggroll.session.processors.per.node": 2
          },
          "task_parallelism": 2,
          "computing_partitions": 8,
          "timeout": 36000,
      }
    }
  3. SPARK With RabbitMQ conf example with manually specified CPU settings:

    "job_parameters": {
      "common": {
          "job_type": "train",
          "spark_run": {
              "num-executors": 1,
              "executor-cores": 2
          },
          "task_parallelism": 2,
          "computing_partitions": 8,
          "timeout": 36000,
          "rabbitmq_run": {
              "queue": {
                  "durable": true
              },
              "connection": {
                  "heartbeat": 10000
              }
          }
      }
    }
  4. SPARK With Pulsar conf example with default setting :

    "job_parameters": {
      "common": {
          "job_type": "train",
          "spark_run": {
              "num-executors": 1,
              "executor-cores": 2
          }
      }
    }

4.5 Resource Management

Starting at version 1.5.0, FATE-Flow implements improved, more fine-grained resource management policy on cpu cores, lifting restrictions on number of parallel tasks in previous versions.

4.5.1 Total Resource Setting
  • resource comes from underlying engines; since current version does automatically obtain resource information from engines, FATE-Flow server obtains and register engine information to t_engine_registry from user-specified conf file $PROJECT_BASE/conf/service_conf.yaml
  • fate_on_eggroll:total_cores=cores_per_node*nodes
  • fate_on_spark:total_cores=cores_per_node*nodes
  • fate_on_standalone:total_cores=cores_per_node*nodes
  • separate computing resources for different engines
  • above settings effective after restarting FATE-Flow server
4.5.2 Calculate Computing Resource

Calculate actual task_run_cores each task requests at computing engine, may not equal to the amount applied by resource manager

  1. only set task_cores in job conf:

    • task_run_cores(guest, host):max(task_cores / total_nodes, 1) * total_nodes
    • task_run_cores(arbiter):max(1 / total_nodes, 1) * total_nodes
    • FATE-Flow will automatically convert task_cores value into engine-specific configuration: eggroll.session.processors.per.node for EGGROLL, and executor-cores & num-executors for SPARK
  2. set eggroll_run in job conf:

    • task_run_cores(guest, host, arbiter):eggroll.session.processors.per.node * total_nodes
  3. set spark_run in job conf:

    • task_run_cores(guest, host, arbiter):executor-cores * num-executors
4.5.3 Resource Manager
  1. Apply Resource for Jobs
    • Computing Engine set to EGGROLL, STANDALONE
      • apply_cores(guest, host): task_run_cores * task_parallelism
      • apply_cores(arbiter): 0, because actual resource cost is minimal and EGGROLL currently sets the same cores for all nodes, set to 0 to avoid unnecessary job queueing due to resource need from arbiter
      • note: on EGGROLL cluster, each node always assigns arbiter task_run_cores/nodes cores
    • Computing Engine set to SPARK
      • SPARK supports executor-cores * num-executors; not strongly correlated with number of cluster nodes due to SPARK own resource manager; if the calculated resource different from the one actually applied, jobs may keep waiting on SPARK engine
      • apply_cores(guest, host, arbiter): task_run_cores * task_parallelism
  2. Job Management Policy
    • Enqueue by job submission time
    • Currently only support FIFO policy: manager only applies resources for the first job, deque the first job if success, wait for the next round if failure
  3. Resource Application Policy
    • Manager selects job following the above guidelines and distribute federated resource application request to all participants
    • If all participants successfully secure resource, i.e.: (total_cores - apply_cores > 0), then the job succeeds in resource application
    • If not all participants succeeds, then send rollback request to succeeded participants, and the job fails in resource application

5. Component Parameter Configuration

5.1 Configuration Applicable Range Policy

  • common: applies to all participants
  • role: applies only to specific participant; specify participant in $role:$party_index format; note that role configuration takes priority over common
"commom": {
}

"role": {
  "guest": {
    "0": {
    }
  }
}

In the example above, configuration insidecommon applies to all participants; configuration inside role-guest-0 only applies to participant guest_0

!!!Note current version now supports checking on both fields of specification.

5.2 Example Component Parameter Configuration

  • Configuration of modules intersection_0& hetero_lr_0are put inside common, thus applies to all participants
  • Configuration of modules reader_0& data_transform_0are specified for each participant
  • Names of the above modules are specified in dsl file
"component_parameters": {
  "common": {
    "intersection_0": {
      "intersect_method": "raw",
      "sync_intersect_ids": true,
      "only_output_key": false
    },
    "hetero_lr_0": {
      "penalty": "L2",
      "optimizer": "rmsprop",
      "alpha": 0.01,
      "max_iter": 3,
      "batch_size": 320,
      "learning_rate": 0.15,
      "init_param": {
        "init_method": "random_uniform"
      }
    }
  },
  "role": {
    "guest": {
      "0": {
        "reader_0": {
          "table": {"name": "breast_hetero_guest", "namespace": "experiment"}
        },
        "data_transform_0":{
          "with_label": true,
          "label_name": "y",
          "label_type": "int",
          "output_format": "dense"
        }
      }
    },
    "host": {
      "0": {
        "reader_0": {
          "table": {"name": "breast_hetero_host", "namespace": "experiment"}
        },
        "data_transform_0":{
          "with_label": false,
          "output_format": "dense"
        }
      }
    }
  }
}

5.3 Multi-host configuration

For multi-host modeling case, all the host's party ids should be list in the role field.

"role": {
   "guest": [
     10000
   ],
   "host": [
     10000, 10001, 10002
   ],
   "arbiter": [
     10000
   ]
}

Each parameter set for host should also be config The number of elements should match the number of hosts.

"component_parameters": {
   "role": {
      "host": {
         "0": {
            "reader_0": {
               "table":
                {
                  "name": "hetero_breast_host_0",
                  "namespace": "hetero_breast_host"
                }
            }
         },
         "1": {
            "reader_0": {
               "table":
               {
                  "name": "hetero_breast_host_1",
                  "namespace": "hetero_breast_host"
               }
            }
         },
         "2": {
            "reader_0": {
               "table":
               {
                  "name": "hetero_breast_host_2",
                  "namespace": "hetero_breast_host"
               }
            }
         }
      }
   }
}

The parameters set in common parameters need not be copied into host role parameters. Common parameters will be copied for every party.

5.4 Prediction configuration

5.4.1 Overview

Please note that in dsl v2,predict dsl is not automatically generated after training. User should first deploy needed components with Flow Client. Please refer to FATE-Flow document for details on using deploy command:

$ flow model deploy --model-id $model_id --model-version $model_version --cpn-list ...

Optionally, user can add additional component(s) to predict dsl, like Evaluation:

5.4.2 Example

training dsl:

"components": {
    "reader_0": {
        "module": "Reader",
        "output": {
            "data": [
                "data"
            ]
        }
    },
    "data_transform_0": {
        "module": "DataTransform",
        "input": {
            "data": {
                "data": [
                    "reader_0.data"
                ]
            }
        },
        "output": {
            "data": [
                "data"
            ],
            "model": [
                "model"
            ]
        }
    },
    "intersection_0": {
        "module": "Intersection",
        "input": {
            "data": {
                "data": [
                    "data_transform_0.data"
                ]
            }
        },
        "output": {
            "data":[
                "data"
            ]
        }
    },
    "hetero_nn_0": {
        "module": "HeteroNN",
        "input": {
            "data": {
                "train_data": [
                    "intersection_0.data"
                ]
            }
        },
        "output": {
            "data": [
                "data"
            ],
            "model": [
                "model"
            ]
        }
    }
}

predict dsl:

"components": {
    "reader_0": {
        "module": "Reader",
        "output": {
            "data": [
                "data"
            ]
        }
    },
    "data_transform_0": {
        "module": "DataTransform",
        "input": {
            "data": {
                "data": [
                    "reader_0.data"
                ]
            }
        },
        "output": {
            "data": [
                "data"
            ],
            "model": [
                "model"
            ]
        }
    },
    "intersection_0": {
        "module": "Intersection",
        "input": {
            "data": {
                "data": [
                    "data_transform_0.data"
                ]
            }
        },
        "output": {
            "data":[
                "data"
            ]
        }
    },
    "hetero_nn_0": {
        "module": "HeteroNN",
        "input": {
            "data": {
                "train_data": [
                    "intersection_0.data"
                ]
            }
        },
        "output": {
            "data": [
                "data"
            ],
            "model": [
                "model"
            ]
        }
    },
    "evaluation_0": {
        "module": "Evaluation",
        "input": {
            "data": {
                "data": [
                    "hetero_nn_0.data"
                ]
            }
         },
         "output": {
             "data": [
                 "data"
             ]
          }
    }

6. Basic Workflow

  1. After job submission, FATE-Flow obtains job dsl and job config and store them inside job folder under corresponding directory $PROJECT_BASE/fateflow/jobs/$jobid/
  2. Parse job dsl & job config, generate fine-grained configuration according to provided settings and fill in default parameter values
  3. Distribute and store common configuration to each party, generate and store party-specific job_runtime_on_party_confunder jobs directory
  4. Each party execute job following job_runtime_on_party_conf