Skip to content

Latest commit

 

History

History
690 lines (528 loc) · 18 KB

ExperimentConfig.md

File metadata and controls

690 lines (528 loc) · 18 KB

Experiment config reference

A config file is needed when create an experiment, the path of the config file is provide to nnictl. The config file is written in yaml format, and need to be written correctly. This document describes the rule to write config file, and will provide some examples and templates.

Template

  • light weight(without Annotation and Assessor)
authorName: 
experimentName: 
trialConcurrency: 
maxExecDuration: 
maxTrialNum: 
#choice: local, remote, pai, kubeflow
trainingServicePlatform: 
searchSpacePath: 
#choice: true, false
useAnnotation: 
tuner:
  #choice: TPE, Random, Anneal, Evolution
  builtinTunerName:
  classArgs:
    #choice: maximize, minimize
    optimize_mode:
  gpuNum: 
trial:
  command: 
  codeDir: 
  gpuNum: 
#machineList can be empty if the platform is local
machineList:
  - ip: 
    port: 
    username: 
    passwd:
  • Use Assessor
authorName: 
experimentName: 
trialConcurrency: 
maxExecDuration: 
maxTrialNum: 
#choice: local, remote, pai, kubeflow
trainingServicePlatform: 
searchSpacePath: 
#choice: true, false
useAnnotation: 
tuner:
  #choice: TPE, Random, Anneal, Evolution
  builtinTunerName:
  classArgs:
    #choice: maximize, minimize
    optimize_mode:
  gpuNum: 
assessor:
  #choice: Medianstop
  builtinAssessorName:
  classArgs:
    #choice: maximize, minimize
    optimize_mode:
  gpuNum: 
trial:
  command: 
  codeDir: 
  gpuNum: 
#machineList can be empty if the platform is local
machineList:
  - ip: 
    port: 
    username: 
    passwd:
  • Use Annotation
authorName: 
experimentName: 
trialConcurrency: 
maxExecDuration: 
maxTrialNum: 
#choice: local, remote, pai, kubeflow
trainingServicePlatform: 
#choice: true, false
useAnnotation: 
tuner:
  #choice: TPE, Random, Anneal, Evolution
  builtinTunerName:
  classArgs:
    #choice: maximize, minimize
    optimize_mode:
  gpuNum: 
assessor:
  #choice: Medianstop
  builtinAssessorName:
  classArgs:
    #choice: maximize, minimize
    optimize_mode:
  gpuNum: 
trial:
  command: 
  codeDir: 
  gpuNum: 
#machineList can be empty if the platform is local
machineList:
  - ip: 
    port: 
    username: 
    passwd:

Configuration

  • authorName

    • Description

      authorName is the name of the author who create the experiment. TBD: add default value

  • experimentName

    • Description

      experimentName is the name of the experiment created.
      TBD: add default value

  • trialConcurrency

    • Description

      trialConcurrency specifies the max num of trial jobs run simultaneously.

      Note: if trialGpuNum is bigger than the free gpu numbers, and the trial jobs running simultaneously can not reach trialConcurrency number, some trial jobs will be put into a queue to wait for gpu allocation.

  • maxExecDuration

    • Description

      maxExecDuration specifies the max duration time of an experiment.The unit of the time is {s, m, h, d}, which means {seconds, minutes, hours, days}.

  • maxTrialNum

    • Description

      maxTrialNum specifies the max number of trial jobs created by nni, including succeeded and failed jobs.

  • trainingServicePlatform

    • Description

      trainingServicePlatform specifies the platform to run the experiment, including {local, remote, pai, kubeflow}.

      • local run an experiment on local ubuntu machine.

      • remote submit trial jobs to remote ubuntu machines, and machineList field should be filed in order to set up SSH connection to remote machine.

      • pai submit trial jobs to OpenPai of Microsoft. For more details of pai configuration, please reference PAIMOdeDoc

      • kubeflow submit trial jobs to kubeflow, nni support kubeflow based on normal kubernetes and azure kubernetes.

  • searchSpacePath

    • Description

      searchSpacePath specifies the path of search space file, which should be a valid path in the local linux machine.

      Note: if set useAnnotation=True, the searchSpacePath field should be removed.

  • useAnnotation

    • Description

      useAnnotation use annotation to analysis trial code and generate search space.

      Note: if set useAnnotation=True, the searchSpacePath field should be removed.

  • nniManagerIp

    • Description

      nniManagerIp set the IP address of the machine on which nni manager process runs. This field is optional, and if it's not set, eth0 device IP will be used instead.

      Note: run ifconfig on NNI manager's machine to check if eth0 device exists. If not, we recommend to set nnimanagerIp explicitly.

  • tuner

    • Description

      tuner specifies the tuner algorithm in the experiment, there are two kinds of ways to set tuner. One way is to use tuner provided by nni sdk, need to set builtinTunerName and classArgs. Another way is to use users' own tuner file, and need to set codeDirectory, classFileName, className and classArgs.

    • builtinTunerName and classArgs

      • builtinTunerName

        builtinTunerName specifies the name of system tuner, nni sdk provides four kinds of tuner, including {TPE, Random, Anneal, Evolution, BatchTuner, GridSearch}

      • classArgs

        classArgs specifies the arguments of tuner algorithm. If the builtinTunerName is in {TPE, Random, Anneal, Evolution}, user should set optimize_mode.

    • codeDir, classFileName, className and classArgs

      • codeDir

        codeDir specifies the directory of tuner code.

      • classFileName

        classFileName specifies the name of tuner file.

      • className

        className specifies the name of tuner class.

      • classArgs

        classArgs specifies the arguments of tuner algorithm.

    • gpuNum

      gpuNum specifies the gpu number to run the tuner process. The value of this field should be a positive number.

      Note: users could only specify one way to set tuner, for example, set {tunerName, optimizationMode} or {tunerCommand, tunerCwd}, and could not set them both.

  • assessor

    • Description

      assessor specifies the assessor algorithm to run an experiment, there are two kinds of ways to set assessor. One way is to use assessor provided by nni sdk, users need to set builtinAssessorName and classArgs. Another way is to use users' own assessor file, and need to set codeDirectory, classFileName, className and classArgs.

    • builtinAssessorName and classArgs

      • builtinAssessorName

        builtinAssessorName specifies the name of system assessor, nni sdk provides one kind of assessor {Medianstop}

      • classArgs

        classArgs specifies the arguments of assessor algorithm

    • codeDir, classFileName, className and classArgs

      • codeDir

        codeDir specifies the directory of assessor code.

      • classFileName

        classFileName specifies the name of assessor file.

      • className

        className specifies the name of assessor class.

      • classArgs

        classArgs specifies the arguments of assessor algorithm.

    • gpuNum

      gpuNum specifies the gpu number to run the assessor process. The value of this field should be a positive number.

      Note: users' could only specify one way to set assessor, for example,set {assessorName, optimizationMode} or {assessorCommand, assessorCwd}, and users could not set them both.If users do not want to use assessor, assessor fileld should leave to empty.

  • trial(local, remote)

    • command

      command specifies the command to run trial process.

    • codeDir

      codeDir specifies the directory of your own trial file.

    • gpuNum

      gpuNum specifies the num of gpu to run the trial process. Default value is 0.

  • trial(pai)

    • command

      command specifies the command to run trial process.

    • codeDir

      codeDir specifies the directory of the own trial file.

    • gpuNum

      gpuNum specifies the num of gpu to run the trial process. Default value is 0.

    • cpuNum

      cpuNum is the cpu number of cpu to be used in pai container.

    • memoryMB

      memoryMB set the momory size to be used in pai's container.

    • image

      image set the image to be used in pai.

    • dataDir

      dataDir is the data directory in hdfs to be used.

    • outputDir

      outputDir is the output directory in hdfs to be used in pai, the stdout and stderr files are stored in the directory after job finished.

  • trial(kubeflow)

    • codeDir

      codeDir is the local directory where the code files in.

    • ps(optional)

      ps is the configuration for kubeflow's tensorflow-operator.

      • replicas

        replicas is the replica number of ps role.

      • command

        command is the run script in ps's container.

      • gpuNum

        gpuNum set the gpu number to be used in ps container.

      • cpuNum

        cpuNum set the cpu number to be used in ps container.

      • memoryMB

        memoryMB set the memory size of the container.

      • image

        image set the image to be used in ps.

    • worker

      worker is the configuration for kubeflow's tensorflow-operator.

      • replicas

        replicas is the replica number of worker role.

      • command

        command is the run script in worker's container.

      • gpuNum

        gpuNum set the gpu number to be used in worker container.

      • cpuNum

        cpuNum set the cpu number to be used in worker container.

      • memoryMB

        memoryMB set the memory size of the container.

      • image

        image set the image to be used in worker.

  • machineList

    machineList should be set if users set trainingServicePlatform=remote, or it could be empty.

    • ip

      ip is the ip address of remote machine.

    • port

      port is the ssh port to be used to connect machine.

      Note: if users set port empty, the default value will be 22.

    • username

      username is the account of remote machine.

    • passwd

      passwd specifies the password of the account.

    • sshKeyPath

      If users use ssh key to login remote machine, could set sshKeyPath in config file. sshKeyPath is the path of ssh key file, which should be valid.

      Note: if users set passwd and sshKeyPath simultaneously, nni will try passwd.

    • passphrase

      passphrase is used to protect ssh key, which could be empty if users don't have passphrase.

  • kubeflowConfig:

    • operator

      operator specify the kubeflow's operator to be used, nni support tf-operator in current version.

    • storage

      storage specify the storage type of kubeflow, including {nfs, azureStorage}. This field is optional, and the default value is nfs. If the config use azureStorage, this field must be completed.

    • nfs

      server is the host of nfs server

      path is the mounted path of nfs

    • keyVault

      If users want to use azure kubernetes service, they should set keyVault to storage the private key of your azure storage account. Refer: https://docs.microsoft.com/en-us/azure/key-vault/key-vault-manage-with-cli2

      • vaultName

        vaultName is the value of --vault-name used in az command.

      • name

        name is the value of --name used in az command.

    • azureStorage

      If users use azure kubernetes service, they should set azure storage account to store code files.

      • accountName

        accountName is the name of azure storage account.

      • azureShare

        azureShare is the share of the azure file storage.

  • paiConfig

    • userName

      userName is the user name of your pai account.

    • password

      password is the password of the pai account.

    • host

      host is the host of pai.

Examples

  • local mode

    If users want to run trial jobs in local machine, and use annotation to generate search space, could use the following config:

authorName: test
experimentName: test_experiment
trialConcurrency: 3
maxExecDuration: 1h
maxTrialNum: 10
#choice: local, remote, pai, kubeflow
trainingServicePlatform: local
#choice: true, false
useAnnotation: true
tuner:
  #choice: TPE, Random, Anneal, Evolution
  builtinTunerName: TPE
  classArgs:
    #choice: maximize, minimize
    optimize_mode: maximize
  gpuNum: 0
trial:
  command: python3 mnist.py
  codeDir: /nni/mnist
  gpuNum: 0

Could add assessor configuration in config file if set assessor.

authorName: test
experimentName: test_experiment
trialConcurrency: 3
maxExecDuration: 1h
maxTrialNum: 10
#choice: local, remote, pai, kubeflow
trainingServicePlatform: local
searchSpacePath: /nni/search_space.json
#choice: true, false
useAnnotation: false
tuner:
  #choice: TPE, Random, Anneal, Evolution
  builtinTunerName: TPE
  classArgs:
    #choice: maximize, minimize
    optimize_mode: maximize
  gpuNum: 0
assessor:
  #choice: Medianstop
  builtinAssessorName: Medianstop
  classArgs:
    #choice: maximize, minimize
    optimize_mode: maximize
  gpuNum: 0
trial:
  command: python3 mnist.py
  codeDir: /nni/mnist
  gpuNum: 0

Or you could specify your own tuner and assessor file as following:

authorName: test
experimentName: test_experiment
trialConcurrency: 3
maxExecDuration: 1h
maxTrialNum: 10
#choice: local, remote, pai, kubeflow
trainingServicePlatform: local
searchSpacePath: /nni/search_space.json
#choice: true, false
useAnnotation: false
tuner:
  codeDir: /nni/tuner
  classFileName: mytuner.py
  className: MyTuner
  classArgs:
    #choice: maximize, minimize
    optimize_mode: maximize
  gpuNum: 0
assessor:
  codeDir: /nni/assessor
  classFileName: myassessor.py
  className: MyAssessor
  classArgs:
    #choice: maximize, minimize
    optimize_mode: maximize
  gpuNum: 0
trial:
  command: python3 mnist.py
  codeDir: /nni/mnist
  gpuNum: 0
  • remote mode

If run trial jobs in remote machine, users could specify the remote mahcine information as fllowing format:

authorName: test
experimentName: test_experiment
trialConcurrency: 3
maxExecDuration: 1h
maxTrialNum: 10
#choice: local, remote, pai, kubeflow
trainingServicePlatform: remote
searchSpacePath: /nni/search_space.json
#choice: true, false
useAnnotation: false
tuner:
  #choice: TPE, Random, Anneal, Evolution
  builtinTunerName: TPE
  classArgs:
    #choice: maximize, minimize
    optimize_mode: maximize
  gpuNum: 0
trial:
  command: python3 mnist.py
  codeDir: /nni/mnist
  gpuNum: 0
#machineList can be empty if the platform is local
machineList:
  - ip: 10.10.10.10
    port: 22
    username: test
    passwd: test
  - ip: 10.10.10.11
    port: 22
    username: test
    passwd: test
  - ip: 10.10.10.12
    port: 22
    username: test
    sshKeyPath: /nni/sshkey
    passphrase: qwert
  • pai mode
authorName: test
experimentName: nni_test1
trialConcurrency: 1
maxExecDuration:500h
maxTrialNum: 1
#choice: local, remote, pai, kubeflow
trainingServicePlatform: pai
searchSpacePath: search_space.json
#choice: true, false
useAnnotation: false
tuner:
  #choice: TPE, Random, Anneal, Evolution, BatchTuner
  #SMAC (SMAC should be installed through nnictl)
  builtinTunerName: TPE
  classArgs:
    #choice: maximize, minimize
    optimize_mode: maximize
trial:
  command: python3 main.py 
  codeDir: .
  gpuNum: 4
  cpuNum: 2
  memoryMB: 10000
  #The docker image to run nni job on pai
  image: msranni/nni:latest
  #The hdfs directory to store data on pai, format 'hdfs://host:port/directory'
  dataDir: hdfs://10.11.12.13:9000/test
  #The hdfs directory to store output data generated by nni, format 'hdfs://host:port/directory'
  outputDir: hdfs://10.11.12.13:9000/test
paiConfig:
  #The username to login pai
  userName: test
  #The password to login pai
  passWord: test
  #The host of restful server of pai
  host: 10.10.10.10
  • kubeflow mode

kubeflow use nfs as storage.

authorName: default
experimentName: example_mni
trialConcurrency: 1
maxExecDuration: 1h
maxTrialNum: 1
#choice: local, remote, pai, kubeflow
trainingServicePlatform: kubeflow
searchSpacePath: search_space.json
#choice: true, false
useAnnotation: false
tuner:
  #choice: TPE, Random, Anneal, Evolution
  builtinTunerName: TPE
  classArgs:
    #choice: maximize, minimize
    optimize_mode: maximize
trial:
  codeDir: .
  worker:
    replicas: 1
    command: python3 mnist.py
    gpuNum: 0
    cpuNum: 1
    memoryMB: 8192
    image: msranni/nni:latest
kubeflowConfig:
  operator: tf-operator
  nfs:
    server: 10.10.10.10
    path: /var/nfs/general

kubeflow use azure storage

authorName: default
experimentName: example_mni
trialConcurrency: 1
maxExecDuration: 1h
maxTrialNum: 1
#choice: local, remote, pai, kubeflow
trainingServicePlatform: kubeflow
searchSpacePath: search_space.json
#choice: true, false
useAnnotation: false
#nniManagerIp: 10.10.10.10
tuner:
  #choice: TPE, Random, Anneal, Evolution
  builtinTunerName: TPE
  classArgs:
    #choice: maximize, minimize
    optimize_mode: maximize
assessor:
  builtinAssessorName: Medianstop
  classArgs:
    optimize_mode: maximize
  gpuNum: 0
trial:
  codeDir: .
  worker:
    replicas: 1
    command: python3 mnist.py
    gpuNum: 0
    cpuNum: 1
    memoryMB: 4096
    image: msranni/nni:latest
kubeflowConfig:
  operator: tf-operator
  keyVault:
    vaultName: Contoso-Vault
    name: AzureStorageAccountKey
  azureStorage:
    accountName: storage
    azureShare: share01