Refactor nnictl and add config_pai.yml

docs/StartExperiment.md

@@ -0,0 +1,33 @@
+How to start an experiment
+===
+## 1.Introduce
+There are few steps to start an new experiment of nni, here are the  process.
+<img src="./img/experiment_process.jpg" width="50%" height="50%" />
+## 2.Details
+### 2.1 Check environment
+The first step to start an experiment is to check whether the environment is ready, nnictl will check if there is an old experiment running or the port of restfurl server is occupied.
+NNICTL will also validate the content of config yaml file, to ensure the experiment config is in correct format.
+
+### 2.2 Start restful server
+After check environment, nnictl will start an restful server process to manage nni experiment, the devault port is 51188.
+
+### 2.3 Check restful server
+Before next steps, nnictl will check whether restful server is successfully started, or the starting process will stop and show error message.
+
+### 2.4 Set experiment config
+NNICTL need to set experiment config before start an experiment, experiment config includes the config values in config yaml file.
+
+### 2.5 Check experiment cofig
+NNICTL will ensure the request to set config is successfully executed.
+
+### 2.6 Start Web UI
+NNICTL will start a Web UI process to show Web UI information,the default port of Web UI is 8080.
+
+### 2.7 Check Web UI
+If Web UI is not successfully started, nnictl will give a warning information, and will continue to start experiment.
+
+### 2.8 Start Experiment
+This is the most import step of starting an nni experiment, nnictl will call restful server process to setup an experiment.
+
+### 2.9 Check experiment
+After start experiment, nnictl will check whether the experiment is correctly created, and show more information of this experiment to users.

examples/trials/auto-gbdt/config.yml

@@ -3,7 +3,7 @@ experimentName: example_auto-gbdt
 trialConcurrency: 1
 maxExecDuration: 10h
 maxTrialNum: 10
-#choice: local, remote
+#choice: local, remote, pai
 trainingServicePlatform: local
 searchSpacePath: search_space.json
 #choice: true, false

examples/trials/auto-gbdt/config_pai.yml

@@ -0,0 +1,36 @@
+authorName: default
+experimentName: example_auto-gbdt
+trialConcurrency: 1
+maxExecDuration: 10h
+maxTrialNum: 10
+#choice: local, remote, pai
+trainingServicePlatform: pai
+searchSpacePath: search_space.json
+#choice: true, false
+useAnnotation: false
+tuner:
+  #choice: TPE, Random, Anneal, Evolution,
+  #SMAC (SMAC should be installed through nnictl)
+  builtinTunerName: TPE
+  classArgs:
+    #choice: maximize, minimize
+    optimize_mode: minimize
+trial:
+  command: python3 main.py
+  codeDir: .
+  gpuNum: 0
+  cpuNum: 1
+  memoryMB: 8196
+  #The docker image to run nni job on pai
+  image: openpai/pai.example.tensorflow
+  #The hdfs directory to store data on pai, format 'hdfs://host:port/directory'
+  hdfsDataDir: hdfs://10.10.10.10:9000/username/nni
+  #The hdfs directory to store output data generated by nni, format 'hdfs://host:port/directory'
+  hdfsOutputDir: hdfs://10.10.10.10:9000/username/nni
+paiConfig:
+  #The username to login pai
+  userName: username
+  #The password to login pai
+  passWord: password
+  #The host of restful server of pai
+  host: 10.10.10.10

examples/trials/ga_squad/config.yml

@@ -3,7 +3,7 @@ experimentName: example_ga_squad
 trialConcurrency: 1
 maxExecDuration: 1h
 maxTrialNum: 10
-#choice: local, remote
+#choice: local, remote, pai
 trainingServicePlatform: local
 #choice: true, false
 useAnnotation: false

examples/trials/ga_squad/config_pai.yml

@@ -0,0 +1,34 @@
+authorName: default
+experimentName: example_ga_squad
+trialConcurrency: 1
+maxExecDuration: 1h
+maxTrialNum: 10
+#choice: local, remote, pai
+trainingServicePlatform: pai
+#choice: true, false
+useAnnotation: false
+tuner:
+  codeDir: ../tuners/ga_customer_tuner
+  classFileName: customer_tuner.py
+  className: CustomerTuner
+  classArgs:
+    optimize_mode: maximize
+trial:
+  command: python3 trial.py
+  codeDir: .
+  gpuNum: 0
+  cpuNum: 1
+  memoryMB: 8196
+  #The docker image to run nni job on pai
+  image: openpai/pai.example.tensorflow
+  #The hdfs directory to store data on pai, format 'hdfs://host:port/directory'
+  hdfsDataDir: hdfs://10.10.10.10:9000/username/nni
+  #The hdfs directory to store output data generated by nni, format 'hdfs://host:port/directory'
+  hdfsOutputDir: hdfs://10.10.10.10:9000/username/nni
+paiConfig:
+  #The username to login pai
+  userName: username
+  #The password to login pai
+  passWord: password
+  #The host of restful server of pai
+  host: 10.10.10.10

examples/trials/mnist-annotation/config.yml

@@ -3,7 +3,7 @@ experimentName: example_mnist
 trialConcurrency: 1
 maxExecDuration: 1h
 maxTrialNum: 10
-#choice: local, remote
+#choice: local, remote, pai
 trainingServicePlatform: local
 #choice: true, false
 useAnnotation: true

examples/trials/mnist-annotation/config_pai.yml

@@ -0,0 +1,35 @@
+authorName: default
+experimentName: example_mnist
+trialConcurrency: 1
+maxExecDuration: 1h
+maxTrialNum: 10
+#choice: local, remote, pai
+trainingServicePlatform: pai
+#choice: true, false
+useAnnotation: true
+tuner:
+  #choice: TPE, Random, Anneal, Evolution,
+  #SMAC (SMAC should be installed through nnictl)
+  builtinTunerName: TPE
+  classArgs:
+    #choice: maximize, minimize
+    optimize_mode: maximize
+trial:
+  command: python3 mnist.py
+  codeDir: .
+  gpuNum: 0
+  cpuNum: 1
+  memoryMB: 8196
+  #The docker image to run nni job on pai
+  image: openpai/pai.example.tensorflow
+  #The hdfs directory to store data on pai, format 'hdfs://host:port/directory'
+  hdfsDataDir: hdfs://10.10.10.10:9000/username/nni
+  #The hdfs directory to store output data generated by nni, format 'hdfs://host:port/directory'
+  hdfsOutputDir: hdfs://10.10.10.10:9000/username/nni
+paiConfig:
+  #The username to login pai
+  userName: username
+  #The password to login pai
+  passWord: password
+  #The host of restful server of pai
+  host: 10.10.10.10

examples/trials/mnist-batch-tune-keras/config.yml

@@ -3,7 +3,7 @@ experimentName: example_mnist-keras
 trialConcurrency: 1
 maxExecDuration: 1h
 maxTrialNum: 10
-#choice: local, remote
+#choice: local, remote, pai
 trainingServicePlatform: local
 searchSpacePath: search_space.json
 #choice: true, false

examples/trials/mnist-batch-tune-keras/config_pai.yml

@@ -0,0 +1,36 @@
+authorName: default
+experimentName: example_mnist-keras
+trialConcurrency: 1
+maxExecDuration: 1h
+maxTrialNum: 10
+#choice: local, remote, pai
+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: BatchTuner
+  classArgs:
+    #choice: maximize, minimize
+    optimize_mode: maximize
+trial:
+  command: python3 mnist-keras.py
+  codeDir: .
+  gpuNum: 0
+  cpuNum: 1
+  memoryMB: 8196
+  #The docker image to run nni job on pai
+  image: openpai/pai.example.tensorflow
+  #The hdfs directory to store data on pai, format 'hdfs://host:port/directory'
+  hdfsDataDir: hdfs://10.10.10.10:9000/username/nni
+  #The hdfs directory to store output data generated by nni, format 'hdfs://host:port/directory'
+  hdfsOutputDir: hdfs://10.10.10.10:9000/username/nni
+paiConfig:
+  #The username to login pai
+  userName: username
+  #The password to login pai
+  passWord: password
+  #The host of restful server of pai
+  host: 10.10.10.10

examples/trials/mnist-keras/config.yml

@@ -3,7 +3,7 @@ experimentName: example_mnist-keras
 trialConcurrency: 1
 maxExecDuration: 1h
 maxTrialNum: 10
-#choice: local, remote
+#choice: local, remote, pai
 trainingServicePlatform: local
 searchSpacePath: search_space.json
 #choice: true, false

examples/trials/mnist-keras/config_pai.yml

@@ -0,0 +1,36 @@
+authorName: default
+experimentName: example_mnist-keras
+trialConcurrency: 1
+maxExecDuration: 1h
+maxTrialNum: 10
+#choice: local, remote, pai
+trainingServicePlatform: pai
+searchSpacePath: search_space.json
+#choice: true, false
+useAnnotation: false
+tuner:
+  #choice: TPE, Random, Anneal, Evolution,
+  #SMAC (SMAC should be installed through nnictl)
+  builtinTunerName: TPE
+  classArgs:
+    #choice: maximize, minimize
+    optimize_mode: maximize
+trial:
+  command: python3 mnist-keras.py
+  codeDir: .
+  gpuNum: 0
+  cpuNum: 1
+  memoryMB: 8196
+  #The docker image to run nni job on pai
+  image: openpai/pai.example.tensorflow
+  #The hdfs directory to store data on pai, format 'hdfs://host:port/directory'
+  hdfsDataDir: hdfs://10.10.10.10:9000/username/nni
+  #The hdfs directory to store output data generated by nni, format 'hdfs://host:port/directory'
+  hdfsOutputDir: hdfs://10.10.10.10:9000/username/nni
+paiConfig:
+  #The username to login pai
+  userName: username
+  #The password to login pai
+  passWord: password
+  #The host of restful server of pai
+  host: 10.10.10.10

examples/trials/mnist-smartparam/config.yml

@@ -3,7 +3,7 @@ experimentName: example_mnist-smartparam
 trialConcurrency: 1
 maxExecDuration: 1h
 maxTrialNum: 10
-#choice: local, remote
+#choice: local, remote, pai
 trainingServicePlatform: local
 #choice: true, false
 useAnnotation: true

examples/trials/mnist-smartparam/config_pai.yml

@@ -0,0 +1,35 @@
+authorName: default
+experimentName: example_mnist-smartparam
+trialConcurrency: 1
+maxExecDuration: 1h
+maxTrialNum: 10
+#choice: local, remote, pai
+trainingServicePlatform: pai
+#choice: true, false
+useAnnotation: true
+tuner:
+  #choice: TPE, Random, Anneal, Evolution,
+  #SMAC (SMAC should be installed through nnictl)
+  builtinTunerName: TPE
+  classArgs:
+    #choice: maximize, minimize
+    optimize_mode: maximize
+trial:
+  command: python3 mnist.py
+  codeDir: .
+  gpuNum: 0
+  cpuNum: 1
+  memoryMB: 8196
+  #The docker image to run nni job on pai
+  image: openpai/pai.example.tensorflow
+  #The hdfs directory to store data on pai, format 'hdfs://host:port/directory'
+  hdfsDataDir: hdfs://10.10.10.10:9000/username/nni
+  #The hdfs directory to store output data generated by nni, format 'hdfs://host:port/directory'
+  hdfsOutputDir: hdfs://10.10.10.10:9000/username/nni
+paiConfig:
+  #The username to login pai
+  userName: username
+  #The password to login pai
+  passWord: password
+  #The host of restful server of pai
+  host: 10.10.10.10

examples/trials/mnist/config.yml

@@ -3,7 +3,7 @@ experimentName: example_mnist
 trialConcurrency: 1
 maxExecDuration: 1h
 maxTrialNum: 10
-#choice: local, remote
+#choice: local, remote, pai
 trainingServicePlatform: local
 searchSpacePath: search_space.json
 #choice: true, false

examples/trials/mnist/config_pai.yml

@@ -0,0 +1,36 @@
+authorName: default
+experimentName: example_mnist
+trialConcurrency: 1
+maxExecDuration: 1h
+maxTrialNum: 10
+#choice: local, remote, pai
+trainingServicePlatform: pai
+searchSpacePath: search_space.json
+#choice: true, false
+useAnnotation: false
+tuner:
+  #choice: TPE, Random, Anneal, Evolution,
+  #SMAC (SMAC should be installed through nnictl)
+  builtinTunerName: TPE
+  classArgs:
+    #choice: maximize, minimize
+    optimize_mode: maximize
+trial:
+  command: python3 mnist.py
+  codeDir: .
+  gpuNum: 0
+  cpuNum: 1
+  memoryMB: 8196
+  #The docker image to run nni job on pai
+  image: openpai/pai.example.tensorflow
+  #The hdfs directory to store data on pai, format 'hdfs://host:port/directory'
+  hdfsDataDir: hdfs://10.10.10.10:9000/username/nni
+  #The hdfs directory to store output data generated by nni, format 'hdfs://host:port/directory'
+  hdfsOutputDir: hdfs://10.10.10.10:9000/username/nni
+paiConfig:
+  #The username to login pai
+  userName: username
+  #The password to login pai
+  passWord: password
+  #The host of restful server of pai
+  host: 10.10.10.10

examples/trials/pytorch_cifar10/config.yml

@@ -3,7 +3,7 @@ experimentName: example_pytorch_cifar10
 trialConcurrency: 1
 maxExecDuration: 100h
 maxTrialNum: 10
-#choice: local, remote
+#choice: local, remote, pai
 trainingServicePlatform: local
 searchSpacePath: search_space.json
 #choice: true, false