Merge pull request #222 from microsoft/master

merge master

src/nni_manager/training_service/pai/paiJobRestServer.ts

@@ -19,19 +19,17 @@ export interface ParameterFileMeta {
  * PAI Training service Rest server, provides rest API to support pai job metrics update
  *
  */
-@component.Singleton
 export class PAIJobRestServer extends ClusterJobRestServer {
-    private parameterFileMetaList: ParameterFileMeta[] = [];
+    protected parameterFileMetaList: ParameterFileMeta[] = [];
 
-    @Inject
-    private readonly paiTrainingService: PAITrainingService;
+    protected readonly paiTrainingService: PAITrainingService;
 
     /**
      * constructor to provide NNIRestServer's own rest property, e.g. port
      */
-    constructor() {
+    constructor (paiTrainingService: PAITrainingService) {
         super();
-        this.paiTrainingService = component.get(PAITrainingService);
+        this.paiTrainingService = paiTrainingService;
     }
 
     protected handleTrialMetrics(jobId: string, metrics: any[]): void {

src/nni_manager/training_service/pai/paiK8S/paiK8SConfig.ts

+/**
+ * Copyright (c) Microsoft Corporation
+ * All rights reserved.
+ *
+ * MIT License
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated
+ * documentation files (the "Software"), to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and
+ * to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+ * The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED *AS IS*, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING
+ * BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+'use strict';
+import {TrialConfig} from '../../common/trialConfig';
+
+/**
+ * PAI trial configuration
+ */
+export class NNIPAIK8STrialConfig extends TrialConfig {
+    public readonly cpuNum: number;
+    public readonly memoryMB: number;
+    public readonly image: string;
+    public virtualCluster?: string;
+    public readonly nniManagerNFSMountPath: string;
+    public readonly containerNFSMountPath: string;
+    public readonly paiStoragePlugin: string;
+
+    constructor(command: string, codeDir: string, gpuNum: number, cpuNum: number, memoryMB: number,
+                image: string, nniManagerNFSMountPath: string, containerNFSMountPath: string,
+                paiStoragePlugin: string, virtualCluster?: string) {
+        super(command, codeDir, gpuNum);
+        this.cpuNum = cpuNum;
+        this.memoryMB = memoryMB;
+        this.image = image;
+        this.virtualCluster = virtualCluster;
+        this.nniManagerNFSMountPath = nniManagerNFSMountPath;
+        this.containerNFSMountPath = containerNFSMountPath;
+        this.paiStoragePlugin = paiStoragePlugin;
+    }
+}

src/nni_manager/training_service/pai/paiK8S/paiK8SData.ts

+/**
+ * Copyright (c) Microsoft Corporation
+ * All rights reserved.
+ *
+ * MIT License
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated
+ * documentation files (the "Software"), to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and
+ * to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+ * The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED *AS IS*, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING
+ * BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+'use strict';
+
+export const PAI_INSTALL_NNI_SHELL_FORMAT: string =
+`#!/bin/bash
+if python3 -c 'import nni' > /dev/null 2>&1; then
+  # nni module is already installed, skip
+  return
+else
+  # Install nni
+  python3 -m pip install --user nni
+fi`;
+
+export const PAI_K8S_TRIAL_COMMAND_FORMAT: string =
+`export NNI_PLATFORM=pai NNI_SYS_DIR={0} NNI_OUTPUT_DIR={1} NNI_TRIAL_JOB_ID={2} NNI_EXP_ID={3} NNI_TRIAL_SEQ_ID={4} MULTI_PHASE={5} \
+&& ls $NNI_SYS_DIR \
+&& cd $NNI_SYS_DIR && sh install_nni.sh \
+&& python3 -m nni_trial_tool.trial_keeper --trial_command '{6}' --nnimanager_ip '{7}' --nnimanager_port '{8}' \
+--nni_manager_version '{9}' --log_collection '{10}'`;

src/nni_manager/training_service/pai/paiK8S/paiK8STrainingService.ts

+/**
+ * Copyright (c) Microsoft Corporation
+ * All rights reserved.
+ *
+ * MIT License
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated
+ * documentation files (the "Software"), to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and
+ * to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+ * The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED *AS IS*, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING
+ * BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+'use strict';
+
+import * as cpp from 'child-process-promise';
+import * as fs from 'fs';
+import * as path from 'path';
+// tslint:disable-next-line:no-implicit-dependencies
+import * as request from 'request';
+import * as component from '../../../common/component';
+
+import { Deferred } from 'ts-deferred';
+import { String } from 'typescript-string-operations';
+import {
+    HyperParameters, NNIManagerIpConfig, TrainingService,
+    TrialJobApplicationForm, TrialJobDetail, TrialJobMetric
+} from '../../../common/trainingService';
+import { delay, generateParamFileName,
+    getExperimentRootDir, getIPV4Address, getVersion, uniqueString, unixPathJoin } from '../../../common/utils';
+import { CONTAINER_INSTALL_NNI_SHELL_FORMAT } from '../../common/containerJobData';
+import { TrialConfigMetadataKey } from '../../common/trialConfigMetadataKey';
+import { execMkdir, validateCodeDir, execCopydir } from '../../common/util';
+import { PAI_K8S_TRIAL_COMMAND_FORMAT } from './paiK8SData';
+import { NNIPAIK8STrialConfig } from './paiK8SConfig';
+import { PAITrainingService } from '../paiTrainingService';
+import { PAIClusterConfig, PAITrialJobDetail } from '../paiConfig';
+import { PAIJobRestServer } from '../paiJobRestServer';
+
+const yaml = require('js-yaml');
+
+/**
+ * Training Service implementation for OpenPAI (Open Platform for AI)
+ * Refer https://github.com/Microsoft/pai for more info about OpenPAI
+ */
+@component.Singleton
+class PAIK8STrainingService extends PAITrainingService {
+    protected paiTrialConfig: NNIPAIK8STrialConfig | undefined;
+
+    constructor() {
+        super();
+    }
+
+    public async setClusterMetadata(key: string, value: string): Promise<void> {
+        switch (key) {
+            case TrialConfigMetadataKey.PAI_CLUSTER_CONFIG:
+                this.paiJobRestServer = new PAIJobRestServer(component.get(PAIK8STrainingService));
+                this.paiClusterConfig = <PAIClusterConfig>JSON.parse(value);
+                if(this.paiClusterConfig.passWord) {
+                    // Get PAI authentication token
+                    await this.updatePaiToken();
+                } else if(this.paiClusterConfig.token) {
+                    this.paiToken = this.paiClusterConfig.token;
+                }
+                break;
+
+            case TrialConfigMetadataKey.TRIAL_CONFIG:
+                if (this.paiClusterConfig === undefined) {
+                    this.log.error('pai cluster config is not initialized');
+                    break;
+                }
+                this.paiTrialConfig = <NNIPAIK8STrialConfig>JSON.parse(value);
+                // Validate to make sure codeDir doesn't have too many files
+                await validateCodeDir(this.paiTrialConfig.codeDir);
+                break;
+            case TrialConfigMetadataKey.VERSION_CHECK:
+                this.versionCheck = (value === 'true' || value === 'True');
+                break;
+            case TrialConfigMetadataKey.LOG_COLLECTION:
+                this.logCollection = value;
+                break;
+            case TrialConfigMetadataKey.MULTI_PHASE:
+                this.isMultiPhase = (value === 'true' || value === 'True');
+                break;
+            default:
+                //Reject for unknown keys
+                this.log.error(`Uknown key: ${key}`);
+        }
+    }
+    
+    //TODO: update trial parameters
+    public async updateTrialJob(trialJobId: string, form: TrialJobApplicationForm): Promise<TrialJobDetail> {
+        const trialJobDetail: undefined | TrialJobDetail = this.trialJobsMap.get(trialJobId);
+        if (trialJobDetail === undefined) {
+            throw new Error(`updateTrialJob failed: ${trialJobId} not found`);
+        }
+        return trialJobDetail;
+    }
+
+    public async submitTrialJob(form: TrialJobApplicationForm): Promise<TrialJobDetail> {
+        if (this.paiClusterConfig === undefined) {
+            throw new Error(`paiClusterConfig not initialized!`);
+        }
+        if (this.paiTrialConfig === undefined) {
+            throw new Error(`paiTrialConfig not initialized!`);
+        }
+
+        this.log.info(`submitTrialJob: form: ${JSON.stringify(form)}`);
+
+        const trialJobId: string = uniqueString(5);
+        //TODO: use HDFS working folder instead
+        const trialWorkingFolder: string = path.join(this.expRootDir, 'trials', trialJobId);
+        const paiJobName: string = `nni_exp_${this.experimentId}_trial_${trialJobId}`;
+        const logPath: string = path.join(this.paiTrialConfig.nniManagerNFSMountPath, this.experimentId, trialJobId);
+        const trialJobDetail: PAITrialJobDetail = new PAITrialJobDetail(
+            trialJobId,
+            'WAITING',
+            paiJobName,
+            Date.now(),
+            trialWorkingFolder,
+            form,
+            logPath);
+
+        this.trialJobsMap.set(trialJobId, trialJobDetail);
+        this.jobQueue.push(trialJobId);
+
+        return trialJobDetail;
+    }
+
+    public generateJobConfigInYamlFormat(trialJobId: string, command: string) {
+        if (this.paiTrialConfig === undefined) {
+            throw new Error('trial config is not initialized');
+        }
+        const jobName = `nni_exp_${this.experimentId}_trial_${trialJobId}`
+        const paiJobConfig: any = {
+            protocolVersion: 2, 
+            name: jobName,
+            type: 'job',
+            jobRetryCount: 0,
+            prerequisites: [
+              {
+                type: 'dockerimage',
+                uri: this.paiTrialConfig.image,
+                name: 'docker_image_0'
+              }
+            ],
+            taskRoles: {
+                taskrole: {
+                    instances: 1,
+                    completion: {
+                       minFailedInstances: 1,
+                       minSucceededInstances: -1
+                    },
+                    taskRetryCount: 0,
+                    dockerImage: 'docker_image_0',
+                    resourcePerInstance: {
+                        gpu: this.paiTrialConfig.gpuNum,
+                        cpu: this.paiTrialConfig.cpuNum,
+                        memoryMB: this.paiTrialConfig.memoryMB
+                    },
+                    commands: [
+                        command
+                    ]
+                }
+            },
+            extras: {
+                'com.microsoft.pai.runtimeplugin': [
+                    {
+                        plugin: this.paiTrialConfig.paiStoragePlugin
+                    }
+                ],
+                submitFrom: 'submit-job-v2'
+            }
+        }
+        if (this.paiTrialConfig.virtualCluster) {
+            paiJobConfig.defaults=  {
+                virtualCluster: this.paiTrialConfig.virtualCluster
+            }
+        }
+
+        return yaml.safeDump(paiJobConfig);
+      }
+
+    protected async submitTrialJobToPAI(trialJobId: string): Promise<boolean> {
+        const deferred: Deferred<boolean> = new Deferred<boolean>();
+        const trialJobDetail: PAITrialJobDetail | undefined = this.trialJobsMap.get(trialJobId);
+
+        if (trialJobDetail === undefined) {
+            throw new Error(`Failed to find PAITrialJobDetail for job ${trialJobId}`);
+        }
+
+        if (this.paiClusterConfig === undefined) {
+            throw new Error('PAI Cluster config is not initialized');
+        }
+        if (this.paiTrialConfig === undefined) {
+            throw new Error('trial config is not initialized');
+        }
+        if (this.paiToken === undefined) {
+            throw new Error('PAI token is not initialized');
+        }
+
+        if (this.paiJobRestServer === undefined) {
+            throw new Error('paiJobRestServer is not initialized');
+        }
+
+        this.paiRestServerPort = this.paiJobRestServer.clusterRestServerPort;
+
+        // Step 1. Prepare PAI job configuration
+        const trialLocalFolder: string = path.join(this.paiTrialConfig.nniManagerNFSMountPath, this.experimentId, trialJobId);
+        //create trial local working folder locally.
+        await execMkdir(trialLocalFolder);
+
+        const runScriptContent: string = CONTAINER_INSTALL_NNI_SHELL_FORMAT;
+        // Write NNI installation file to local files
+        await fs.promises.writeFile(path.join(trialLocalFolder, 'install_nni.sh'), runScriptContent, { encoding: 'utf8' });
+
+        // Write file content ( parameter.cfg ) to local working folders
+        if (trialJobDetail.form !== undefined) {
+            await fs.promises.writeFile(
+                path.join(trialLocalFolder, generateParamFileName(trialJobDetail.form.hyperParameters)),
+                trialJobDetail.form.hyperParameters.value, { encoding: 'utf8' }
+            );
+        }
+
+        //Copy codeDir files to local working folder
+        await execCopydir(this.paiTrialConfig.codeDir, trialLocalFolder);
+        
+        const nniManagerIp: string = this.nniManagerIpConfig ? this.nniManagerIpConfig.nniManagerIp : getIPV4Address();
+        const version: string = this.versionCheck ? await getVersion() : '';
+        const containerWorkingDir: string = `${this.paiTrialConfig.containerNFSMountPath}/${this.experimentId}/${trialJobId}`;
+        const nniPaiTrialCommand: string = String.Format(
+            PAI_K8S_TRIAL_COMMAND_FORMAT,
+            `${containerWorkingDir}`,
+            `${containerWorkingDir}/nnioutput`,
+            trialJobId,
+            this.experimentId,
+            trialJobDetail.form.sequenceId,
+            this.isMultiPhase,
+            this.paiTrialConfig.command,
+            nniManagerIp,
+            this.paiRestServerPort,
+            version,
+            this.logCollection
+        )
+        .replace(/\r\n|\n|\r/gm, '');
+
+        this.log.info(`nniPAItrial command is ${nniPaiTrialCommand.trim()}`);
+        
+        const paiJobConfig = this.generateJobConfigInYamlFormat(trialJobId, nniPaiTrialCommand);
+
+        // Step 3. Submit PAI job via Rest call
+        // Refer https://github.com/Microsoft/pai/blob/master/docs/rest-server/API.md for more detail about PAI Rest API
+        const submitJobRequest: request.Options = {
+            uri: `http://${this.paiClusterConfig.host}/rest-server/api/v2/jobs`,
+            method: 'POST',
+            body: paiJobConfig,
+            headers: {
+                'Content-Type': 'text/yaml',
+                Authorization: `Bearer ${this.paiToken}`
+            }
+        };
+        request(submitJobRequest, (error: Error, response: request.Response, body: any) => {
+            if ((error !== undefined && error !== null) || response.statusCode >= 400) {
+                const errorMessage: string = (error !== undefined && error !== null) ? error.message :
+                    `Submit trial ${trialJobId} failed, http code:${response.statusCode}, http body: ${body}`;
+
+                this.log.error(errorMessage);
+                trialJobDetail.status = 'FAILED';
+            } else {
+                trialJobDetail.submitTime = Date.now();
+            }
+            deferred.resolve(true);
+        });
+
+        return deferred.promise;
+    }
+}
+
+export { PAIK8STrainingService };

src/nni_manager/training_service/pai/hdfsClientUtility.ts → src/nni_manager/training_service/pai/paiYarn/hdfsClientUtility.ts

@@ -4,9 +4,9 @@
 import * as fs from 'fs';
 import * as path from 'path';
 import { Deferred } from 'ts-deferred';
-import { getExperimentId } from '../../common/experimentStartupInfo';
-import { getLogger } from '../../common/log';
-import { unixPathJoin } from '../../common/utils';
+import { getExperimentId } from '../../../common/experimentStartupInfo';
+import { getLogger } from '../../../common/log';
+import { unixPathJoin } from '../../../common/utils';
 
 /**
  * HDFS client utility, including copy file/directory

src/nni_manager/training_service/pai/paiYarn/paiYarnConfig.ts

+// Copyright (c) Microsoft Corporation.
+// Licensed under the MIT license.
+
+'use strict';
+
+import {TrialConfig} from '../../common/trialConfig';
+
+/**
+ * Task role for PAI
+ */
+export class PAITaskRole {
+    // Name for the task role
+    public readonly name: string;
+    // Number of tasks for the task role, no less than 1
+    public readonly taskNumber: number;
+    // CPU number for one task in the task role, no less than 1
+    public readonly cpuNumber: number;
+    // Memory for one task in the task role, no less than 100
+    public readonly memoryMB: number;
+    // GPU number for one task in the task role, no less than 0
+    public readonly gpuNumber: number;
+    // Executable command for tasks in the task role, can not be empty
+    public readonly command: string;
+    //Shared memory for one task in the task role
+    public readonly shmMB?: number;
+    //portList to specify the port used in container
+    public portList?: PortListMetaData[];
+
+    /**
+     * Constructor
+     * @param name Name for the task role
+     * @param taskNumber Number of tasks for the task role, no less than 1
+     * @param cpuNumber CPU number for one task in the task role, no less than 1
+     * @param memoryMB Memory for one task in the task role, no less than 100
+     * @param gpuNumber GPU number for one task in the task role, no less than 0
+     * @param command Executable command for tasks in the task role, can not be empty
+     */
+    constructor(name: string, taskNumber: number, cpuNumber: number, memoryMB: number, gpuNumber: number,
+                command: string, shmMB?: number, portList?: PortListMetaData[]) {
+        this.name = name;
+        this.taskNumber = taskNumber;
+        this.cpuNumber = cpuNumber;
+        this.memoryMB = memoryMB;
+        this.gpuNumber = gpuNumber;
+        this.command = command;
+        this.shmMB = shmMB;
+        this.portList = portList;
+    }
+}
+
+/**
+ * Trial job configuration submitted to PAI
+ */
+export class PAIJobConfig {
+    // Name for the job, need to be unique
+    public readonly jobName: string;
+    // URL pointing to the Docker image for all tasks in the job
+    public readonly image: string;
+    // Code directory on HDFS
+    public readonly codeDir: string;
+    //authentication file used for private Docker registry 
+    public readonly authFile?: string;
+
+    // List of taskRole, one task role at least
+    public taskRoles: PAITaskRole[];
+
+    //The virtual cluster job runs on.
+    public readonly virtualCluster: string;
+
+    /**
+     * Constructor
+     * @param jobName Name for the job, need to be unique
+     * @param image URL pointing to the Docker image for all tasks in the job
+     * @param dataDir Data directory existing on HDFS
+     * @param outputDir Output directory on HDFS
+     * @param taskRoles List of taskRole, one task role at least
+     */
+    constructor(jobName: string, image: string, codeDir: string,
+                taskRoles: PAITaskRole[], virtualCluster: string, authFile?: string) {
+        this.jobName = jobName;
+        this.image = image;
+        this.codeDir = codeDir;
+        this.taskRoles = taskRoles;
+        this.virtualCluster = virtualCluster;
+        this.authFile = authFile;
+    }
+}
+
+/**
+ * portList data structure used in PAI taskRole
+ */
+export class PortListMetaData {
+    public readonly label: string = '';
+    public readonly beginAt: number = 0;
+    public readonly portNumber: number = 0;
+}
+  
+
+/**
+ * PAI trial configuration
+ */
+export class NNIPAITrialConfig extends TrialConfig {
+    public readonly cpuNum: number;
+    public readonly memoryMB: number;
+    public readonly image: string;
+
+    //The virtual cluster job runs on. If omitted, the job will run on default virtual cluster
+    public virtualCluster?: string;
+    //Shared memory for one task in the task role
+    public shmMB?: number;
+    //authentication file used for private Docker registry 
+    public authFile?: string;
+    //portList to specify the port used in container
+    public portList?: PortListMetaData[];
+
+    constructor(command: string, codeDir: string, gpuNum: number, cpuNum: number, memoryMB: number,
+                image: string, virtualCluster?: string, shmMB?: number, authFile?: string, portList?: PortListMetaData[]) {
+        super(command, codeDir, gpuNum);
+        this.cpuNum = cpuNum;
+        this.memoryMB = memoryMB;
+        this.image = image;
+        this.virtualCluster = virtualCluster;
+        this.shmMB = shmMB;
+        this.authFile = authFile;
+        this.portList = portList;
+    }
+}

src/nni_manager/training_service/pai/paiData.ts → src/nni_manager/training_service/pai/paiYarn/paiYarnData.ts

@@ -3,37 +3,7 @@
 
 'use strict';
 
-import { TrialJobApplicationForm, TrialJobDetail, TrialJobStatus  } from '../../common/trainingService';
-
-/**
- * PAI trial job detail
- */
-export class PAITrialJobDetail implements TrialJobDetail {
-    public id: string;
-    public status: TrialJobStatus;
-    public paiJobName: string;
-    public submitTime: number;
-    public startTime?: number;
-    public endTime?: number;
-    public tags?: string[];
-    public url?: string;
-    public workingDirectory: string;
-    public form: TrialJobApplicationForm;
-    public hdfsLogPath: string;
-    public isEarlyStopped?: boolean;
-
-    constructor(id: string, status: TrialJobStatus, paiJobName: string,
-                submitTime: number, workingDirectory: string, form: TrialJobApplicationForm, hdfsLogPath: string) {
-        this.id = id;
-        this.status = status;
-        this.paiJobName = paiJobName;
-        this.submitTime = submitTime;
-        this.workingDirectory = workingDirectory;
-        this.form = form;
-        this.tags = [];
-        this.hdfsLogPath = hdfsLogPath;
-    }
-}
+import { TrialJobApplicationForm, TrialJobDetail, TrialJobStatus  } from '../../../common/trainingService';
 
 export const PAI_INSTALL_NNI_SHELL_FORMAT: string =
 `#!/bin/bash
@@ -46,7 +16,7 @@ else
 fi`;
 
 export const PAI_TRIAL_COMMAND_FORMAT: string =
-`export NNI_PLATFORM=pai NNI_SYS_DIR={0} NNI_OUTPUT_DIR={1} NNI_TRIAL_JOB_ID={2} NNI_EXP_ID={3} NNI_TRIAL_SEQ_ID={4} MULTI_PHASE={5} \
+`export NNI_PLATFORM=paiYarn NNI_SYS_DIR={0} NNI_OUTPUT_DIR={1} NNI_TRIAL_JOB_ID={2} NNI_EXP_ID={3} NNI_TRIAL_SEQ_ID={4} MULTI_PHASE={5} \
 && cd $NNI_SYS_DIR && sh install_nni.sh \
 && python3 -m nni_trial_tool.trial_keeper --trial_command '{6}' --nnimanager_ip '{7}' --nnimanager_port '{8}' \
 --pai_hdfs_output_dir '{9}' --pai_hdfs_host '{10}' --pai_user_name {11} --nni_hdfs_exp_dir '{12}' --webhdfs_path '/webhdfs/api/v1' \

src/nni_manager/training_service/pai/paiYarn/paiYarnTrainingService.ts

+// Copyright (c) Microsoft Corporation.
+// Licensed under the MIT license.
+
+'use strict';
+
+import * as fs from 'fs';
+import * as path from 'path';
+import * as request from 'request';
+import * as component from '../../../common/component';
+
+import { EventEmitter } from 'events';
+import { Deferred } from 'ts-deferred';
+import { String } from 'typescript-string-operations';
+import { getExperimentId } from '../../../common/experimentStartupInfo';
+import { getLogger, Logger } from '../../../common/log';
+import {
+    HyperParameters, NNIManagerIpConfig, TrainingService,
+    TrialJobApplicationForm, TrialJobDetail, TrialJobMetric
+} from '../../../common/trainingService';
+import { delay, generateParamFileName,
+    getExperimentRootDir, getIPV4Address, getVersion, uniqueString, unixPathJoin } from '../../../common/utils';
+import { CONTAINER_INSTALL_NNI_SHELL_FORMAT } from '../../common/containerJobData';
+import { TrialConfigMetadataKey } from '../../common/trialConfigMetadataKey';
+import { execMkdir, validateCodeDir } from '../../common/util';
+import { HDFSClientUtility } from './hdfsClientUtility';
+import { NNIPAITrialConfig, PAIJobConfig, PAITaskRole } from './paiYarnConfig';
+import { PAI_LOG_PATH_FORMAT, PAI_TRIAL_COMMAND_FORMAT } from './paiYarnData';
+import { PAIJobInfoCollector } from '../paiJobInfoCollector';
+import { PAITrainingService } from '../paiTrainingService';
+import { PAIClusterConfig, PAITrialJobDetail } from '../paiConfig';
+
+import * as WebHDFS from 'webhdfs';
+import { PAIJobRestServer, ParameterFileMeta } from '../paiJobRestServer';
+
+/**
+ * Training Service implementation for OpenPAI (Open Platform for AI)
+ * Refer https://github.com/Microsoft/pai for more info about OpenPAI
+ */
+@component.Singleton
+class PAIYarnTrainingService extends PAITrainingService {
+    private hdfsClient: any;
+    private copyExpCodeDirPromise?: Promise<void>;
+    private copyAuthFilePromise?: Promise<void>;
+    private paiTrialConfig?: NNIPAITrialConfig;
+
+    constructor() {
+        super();
+    }
+
+    public async submitTrialJob(form: TrialJobApplicationForm): Promise<TrialJobDetail> {
+        if (this.paiClusterConfig === undefined) {
+            throw new Error(`paiBaseClusterConfig not initialized!`);
+        }
+
+        this.log.info(`submitTrialJob: form: ${JSON.stringify(form)}`);
+
+        const trialJobId: string = uniqueString(5);
+        //TODO: use HDFS working folder instead
+        const trialWorkingFolder: string = path.join(this.expRootDir, 'trials', trialJobId);
+        const paiJobName: string = `nni_exp_${this.experimentId}_trial_${trialJobId}`;
+        const hdfsCodeDir: string = HDFSClientUtility.getHdfsTrialWorkDir(this.paiClusterConfig.userName, trialJobId);
+        const hdfsOutputDir: string = unixPathJoin(hdfsCodeDir, 'nnioutput');
+
+        const hdfsLogPath: string = String.Format(
+            PAI_LOG_PATH_FORMAT,
+            this.paiClusterConfig.host,
+            hdfsOutputDir
+            );
+
+        const trialJobDetail: PAITrialJobDetail = new PAITrialJobDetail(
+            trialJobId,
+            'WAITING',
+            paiJobName,
+            Date.now(),
+            trialWorkingFolder,
+            form,
+            hdfsLogPath);
+
+        this.trialJobsMap.set(trialJobId, trialJobDetail);
+        this.jobQueue.push(trialJobId);
+
+        return trialJobDetail;
+    }
+
+    public async setClusterMetadata(key: string, value: string): Promise<void> {
+        switch (key) {
+            case TrialConfigMetadataKey.NNI_MANAGER_IP:
+                this.nniManagerIpConfig = <NNIManagerIpConfig>JSON.parse(value);
+                break;
+
+            case TrialConfigMetadataKey.PAI_YARN_CLUSTER_CONFIG:
+                this.paiJobRestServer = new PAIJobRestServer(component.get(PAIYarnTrainingService));
+                this.paiClusterConfig = <PAIClusterConfig>JSON.parse(value);
+
+                this.hdfsClient = WebHDFS.createClient({
+                    user: this.paiClusterConfig.userName,
+                    // Refer PAI document for Pylon mapping https://github.com/Microsoft/pai/tree/master/docs/pylon
+                    port: 80,
+                    path: '/webhdfs/api/v1',
+                    host: this.paiClusterConfig.host
+                });
+                if(this.paiClusterConfig.passWord) {
+                    // Get PAI authentication token
+                    await this.updatePaiToken();
+                } else if(this.paiClusterConfig.token) {
+                    this.paiToken = this.paiClusterConfig.token;
+                } else {
+                    throw new Error('pai cluster config format error, please set password or token!');
+                }
+
+                break;
+
+            case TrialConfigMetadataKey.TRIAL_CONFIG:
+                if (this.paiClusterConfig === undefined) {
+                    this.log.error('pai cluster config is not initialized');
+                    break;
+                }
+                this.paiTrialConfig = <NNIPAITrialConfig>JSON.parse(value);
+
+                // Validate to make sure codeDir doesn't have too many files
+                await validateCodeDir(this.paiTrialConfig.codeDir);
+           
+                // Copy experiment files from local folder to HDFS
+                this.copyExpCodeDirPromise = HDFSClientUtility.copyDirectoryToHdfs(
+                    this.paiTrialConfig.codeDir,
+                    HDFSClientUtility.getHdfsExpCodeDir(this.paiClusterConfig.userName),
+                    this.hdfsClient
+                );
+                
+                // Upload authFile to hdfs
+                if (this.paiTrialConfig.authFile) {
+                    this.authFileHdfsPath = unixPathJoin(HDFSClientUtility.hdfsExpRootDir(this.paiClusterConfig.userName), 'authFile');
+                    this.copyAuthFilePromise = HDFSClientUtility.copyFileToHdfs(this.paiTrialConfig.authFile, this.authFileHdfsPath, this.hdfsClient);
+                }
+
+                break;
+            case TrialConfigMetadataKey.VERSION_CHECK:
+                this.versionCheck = (value === 'true' || value === 'True');
+                break;
+            case TrialConfigMetadataKey.LOG_COLLECTION:
+                this.logCollection = value;
+                break;
+            case TrialConfigMetadataKey.MULTI_PHASE:
+                this.isMultiPhase = (value === 'true' || value === 'True');
+                break;
+            default:
+                //Reject for unknown keys
+                throw new Error(`Uknown key: ${key}`);
+        }
+    }
+
+    protected async submitTrialJobToPAI(trialJobId: string): Promise<boolean> {
+        const deferred: Deferred<boolean> = new Deferred<boolean>();
+        const trialJobDetail: PAITrialJobDetail | undefined = this.trialJobsMap.get(trialJobId);
+
+        if (trialJobDetail === undefined) {
+            throw new Error(`Failed to find PAITrialJobDetail for job ${trialJobId}`);
+        }
+
+        if (this.paiClusterConfig === undefined) {
+            throw new Error('PAI Cluster config is not initialized');
+        }
+        if (this.paiTrialConfig === undefined) {
+            throw new Error('trial config is not initialized');
+        }
+        if (this.paiToken === undefined) {
+            throw new Error('PAI token is not initialized');
+        }
+
+        if (this.paiJobRestServer === undefined) {
+            throw new Error('paiJobRestServer is not initialized');
+        }
+
+        this.paiRestServerPort = this.paiJobRestServer.clusterRestServerPort;
+
+        // Make sure experiment code files is copied from local to HDFS
+        if (this.copyExpCodeDirPromise !== undefined) {
+            await this.copyExpCodeDirPromise;
+        }
+
+        //Make sure authFile is copied from local to HDFS
+        if (this.paiTrialConfig.authFile) {
+            await this.copyAuthFilePromise;
+        }
+        // Step 1. Prepare PAI job configuration
+
+        const trialLocalTempFolder: string = path.join(getExperimentRootDir(), 'trials-local', trialJobId);
+        //create tmp trial working folder locally.
+        await execMkdir(trialLocalTempFolder);
+
+        const runScriptContent: string = CONTAINER_INSTALL_NNI_SHELL_FORMAT;
+        // Write NNI installation file to local tmp files
+        await fs.promises.writeFile(path.join(trialLocalTempFolder, 'install_nni.sh'), runScriptContent, { encoding: 'utf8' });
+
+        // Write file content ( parameter.cfg ) to local tmp folders
+        if (trialJobDetail.form !== undefined) {
+            await fs.promises.writeFile(
+                path.join(trialLocalTempFolder, generateParamFileName(trialJobDetail.form.hyperParameters)),
+                trialJobDetail.form.hyperParameters.value, { encoding: 'utf8' }
+            );
+        }
+        const hdfsCodeDir: string = HDFSClientUtility.getHdfsTrialWorkDir(this.paiClusterConfig.userName, trialJobId);
+        const hdfsOutputDir: string = unixPathJoin(hdfsCodeDir, 'nnioutput');
+        const nniManagerIp: string = this.nniManagerIpConfig ? this.nniManagerIpConfig.nniManagerIp : getIPV4Address();
+        const version: string = this.versionCheck ? await getVersion() : '';
+        const nniPaiTrialCommand: string = String.Format(
+            PAI_TRIAL_COMMAND_FORMAT,
+            // PAI will copy job's codeDir into /root directory
+            `$PWD/${trialJobId}`,
+            `$PWD/${trialJobId}/nnioutput`,
+            trialJobId,
+            this.experimentId,
+            trialJobDetail.form.sequenceId,
+            this.isMultiPhase,
+            this.paiTrialConfig.command,
+            nniManagerIp,
+            this.paiRestServerPort,
+            hdfsOutputDir,
+            this.paiClusterConfig.host,
+            this.paiClusterConfig.userName,
+            HDFSClientUtility.getHdfsExpCodeDir(this.paiClusterConfig.userName),
+            version,
+            this.logCollection
+        )
+        .replace(/\r\n|\n|\r/gm, '');
+
+        this.log.info(`nniPAItrial command is ${nniPaiTrialCommand.trim()}`);
+        const paiTaskRoles: PAITaskRole[] = [
+            new PAITaskRole(
+                `nni_trail_${trialJobId}`,
+                // Task role number
+                1,
+                // Task CPU number
+                this.paiTrialConfig.cpuNum,
+                // Task memory
+                this.paiTrialConfig.memoryMB,
+                // Task GPU number
+                this.paiTrialConfig.gpuNum,
+                // Task command
+                nniPaiTrialCommand,
+                // Task shared memory
+                this.paiTrialConfig.shmMB,
+                // Task portList
+                this.paiTrialConfig.portList
+            )
+        ];
+
+        const paiJobConfig: PAIJobConfig = new PAIJobConfig(
+            // Job name
+            trialJobDetail.paiJobName,
+            // Docker image
+            this.paiTrialConfig.image,
+            // codeDir
+            `$PAI_DEFAULT_FS_URI${hdfsCodeDir}`,
+            // PAI Task roles
+            paiTaskRoles,
+            // Add Virutal Cluster
+            this.paiTrialConfig.virtualCluster === undefined ? 'default' : this.paiTrialConfig.virtualCluster.toString(),
+            //Task auth File
+            this.authFileHdfsPath
+        );
+
+        // Step 2. Upload code files in codeDir onto HDFS
+        try {
+            await HDFSClientUtility.copyDirectoryToHdfs(trialLocalTempFolder, hdfsCodeDir, this.hdfsClient);
+        } catch (error) {
+            this.log.error(`PAI Training service: copy ${this.paiTrialConfig.codeDir} to HDFS ${hdfsCodeDir} failed, error is ${error}`);
+            trialJobDetail.status = 'FAILED'; // eslint-disable-line require-atomic-updates
+            return true;
+        }
+
+        // Step 3. Submit PAI job via Rest call
+        // Refer https://github.com/Microsoft/pai/blob/master/docs/rest-server/API.md for more detail about PAI Rest API
+        const submitJobRequest: request.Options = {
+            uri: `http://${this.paiClusterConfig.host}/rest-server/api/v1/user/${this.paiClusterConfig.userName}/jobs`,
+            method: 'POST',
+            json: true,
+            body: paiJobConfig,
+            headers: {
+                'Content-Type': 'application/json',
+                Authorization: `Bearer ${this.paiToken}`
+            }
+        };
+        request(submitJobRequest, (error: Error, response: request.Response, body: any) => {
+            if ((error !== undefined && error !== null) || response.statusCode >= 400) {
+                const errorMessage: string = (error !== undefined && error !== null) ? error.message :
+                    `Submit trial ${trialJobId} failed, http code:${response.statusCode}, http body: ${response.body.message}`;
+                trialJobDetail.status = 'FAILED';
+                deferred.resolve(true);
+            } else {
+                trialJobDetail.submitTime = Date.now();
+                deferred.resolve(true);
+            }
+        });
+
+        return deferred.promise;
+    }
+
+    public async updateTrialJob(trialJobId: string, form: TrialJobApplicationForm): Promise<TrialJobDetail> {
+        const trialJobDetail: undefined | TrialJobDetail = this.trialJobsMap.get(trialJobId);
+        if (trialJobDetail === undefined) {
+            throw new Error(`updateTrialJob failed: ${trialJobId} not found`);
+        }
+        await this.writeParameterFile(trialJobId, form.hyperParameters);
+
+        return trialJobDetail;
+    }
+
+    protected async writeParameterFile(trialJobId: string, hyperParameters: HyperParameters): Promise<void> {
+        if (this.paiClusterConfig === undefined) {
+            throw new Error('PAI Cluster config is not initialized');
+        }
+        if (this.paiTrialConfig === undefined) {
+            throw new Error('PAI trial config is not initialized');
+        }
+
+        const trialLocalTempFolder: string = path.join(getExperimentRootDir(), 'trials-local', trialJobId);
+        const hpFileName: string = generateParamFileName(hyperParameters);
+        const localFilepath: string = path.join(trialLocalTempFolder, hpFileName);
+        await fs.promises.writeFile(localFilepath, hyperParameters.value, { encoding: 'utf8' });
+        const hdfsCodeDir: string = HDFSClientUtility.getHdfsTrialWorkDir(this.paiClusterConfig.userName, trialJobId);
+        const hdfsHpFilePath: string = path.join(hdfsCodeDir, hpFileName);
+
+        await HDFSClientUtility.copyFileToHdfs(localFilepath, hdfsHpFilePath, this.hdfsClient);
+
+        await this.postParameterFileMeta({
+            experimentId: this.experimentId,
+            trialId: trialJobId,
+            filePath: hdfsHpFilePath
+        });
+    }
+
+    protected postParameterFileMeta(parameterFileMeta: ParameterFileMeta): Promise<void> {
+        const deferred: Deferred<void> = new Deferred<void>();
+        if (this.paiJobRestServer === undefined) {
+            throw new Error('paiJobRestServer not implemented!');
+        }
+        const req: request.Options = {
+            uri: `${this.paiJobRestServer.endPoint}${this.paiJobRestServer.apiRootUrl}/parameter-file-meta`,
+            method: 'POST',
+            json: true,
+            body: parameterFileMeta
+        };
+        request(req, (err: Error, res: request.Response) => {
+            if (err) {
+                deferred.reject(err);
+            } else {
+                deferred.resolve();
+            }
+        });
+
+        return deferred.promise;
+    }
+}
+
+export { PAIYarnTrainingService };

src/nni_manager/training_service/pai/paiTrialConfig.ts → src/nni_manager/training_service/pai/paiYarn/paiYarnTrialConfig.ts

@@ -3,12 +3,12 @@
 
 'use strict';
 
-import {TrialConfig} from '../common/trialConfig';
+import {TrialConfig} from '../../common/trialConfig';
 
 /**
  * PAI configuration to run trials
  */
-export class PAITrialConfig extends TrialConfig {
+export class PAIYarnTrialConfig extends TrialConfig {
     public readonly cpuNum: number;
     public readonly memoryMB: number;
     public readonly image: string;

src/nni_manager/training_service/test/hdfsClientUtility.test.ts

@@ -9,7 +9,7 @@ import * as os from 'os';
 import * as path from 'path';
 import * as tmp from 'tmp';
 import { cleanupUnitTest, prepareUnitTest, uniqueString } from '../../common/utils';
-import { HDFSClientUtility } from '../pai/hdfsClientUtility';
+import { HDFSClientUtility } from '../pai/paiYarn/hdfsClientUtility';
 
 var WebHDFS = require('webhdfs');
 var rmdir = require('rmdir');

src/nni_manager/training_service/test/paiTrainingService.test.ts → src/nni_manager/training_service/test/paiYarnTrainingService.test.ts

@@ -11,14 +11,14 @@ import * as component from '../../common/component';
 import { TrialJobApplicationForm } from '../../common/trainingService';
 import { cleanupUnitTest, prepareUnitTest } from '../../common/utils';
 import { TrialConfigMetadataKey } from '../common/trialConfigMetadataKey';
-import { PAITrainingService } from '../pai/paiTrainingService';
+import { PAIYarnTrainingService } from '../pai/paiYarn/paiYarnTrainingService';
 
 // TODO: copy mockedTrail.py to local folder
 const localCodeDir: string = tmp.dirSync().name
 const mockedTrialPath: string = './training_service/test/mockedTrial.py'
 fs.copyFileSync(mockedTrialPath, localCodeDir + '/mockedTrial.py')
 
-describe('Unit Test for PAITrainingService', () => {
+describe('Unit Test for PAIYarnTrainingService', () => {
     let skip: boolean = false;
     let testPaiClusterInfo: any;
     let paiCluster: any;
@@ -33,7 +33,7 @@ describe('Unit Test for PAITrainingService', () => {
         skip = true;
     }
 
-    let paiTrainingService: PAITrainingService;
+    let paiYarnTrainingService: PAIYarnTrainingService;
 
     console.log(tmp.dirSync().name);
 
@@ -51,15 +51,15 @@ describe('Unit Test for PAITrainingService', () => {
         if (skip) {
             return;
         }
-        paiTrainingService = component.get(PAITrainingService);
-        paiTrainingService.run();
+        paiYarnTrainingService = component.get(PAIYarnTrainingService);
+        paiYarnTrainingService.run();
     });
 
     afterEach(() => {
         if (skip) {
             return;
         }
-        paiTrainingService.cleanUp();
+        paiYarnTrainingService.cleanUp();
     });
 
     it('Get PAI token', async () => {
@@ -67,14 +67,14 @@ describe('Unit Test for PAITrainingService', () => {
             return;
         }
         console.log(`paiCluster is ${paiCluster}`)
-        await paiTrainingService.setClusterMetadata(TrialConfigMetadataKey.PAI_CLUSTER_CONFIG, paiCluster);
-        await paiTrainingService.setClusterMetadata(TrialConfigMetadataKey.TRIAL_CONFIG, paiTrialConfig);
+        await paiYarnTrainingService.setClusterMetadata(TrialConfigMetadataKey.PAI_YARN_CLUSTER_CONFIG, paiCluster);
+        await paiYarnTrainingService.setClusterMetadata(TrialConfigMetadataKey.TRIAL_CONFIG, paiTrialConfig);
         const form: TrialJobApplicationForm = {
             sequenceId: 0,
             hyperParameters: { value: '', index: 0 }
         };
         try {
-            const trialDetail = await paiTrainingService.submitTrialJob(form);
+            const trialDetail = await paiYarnTrainingService.submitTrialJob(form);
             chai.expect(trialDetail.status).to.be.equals('WAITING');
         } catch(error) {
             console.log('Submit job failed:' + error);

src/sdk/pynni/nni/__init__.py

 # Copyright (c) Microsoft Corporation.
 # Licensed under the MIT license.
 
-from .trial import *
-from .smartparam import *
-from .nas_utils import training_update
+from .env_vars import dispatcher_env_vars
+
+if dispatcher_env_vars.SDK_PROCESS != 'dispatcher':
+    from .trial import *
+    from .smartparam import *
+    from .nas_utils import training_update
 
 class NoMoreTrialError(Exception):
     def __init__(self, ErrorInfo):

src/sdk/pynni/nni/__main__.py

 # Copyright (c) Microsoft Corporation.
 # Licensed under the MIT license.
 
-'''
-__main__.py
-'''
 import os
 import sys
 import argparse
 import logging
 import json
 import importlib
+import base64
 
 from .common import enable_multi_thread, enable_multi_phase
 from .constants import ModuleName, ClassName, ClassArgs, AdvisorModuleName, AdvisorClassName
@@ -29,99 +27,67 @@ def augment_classargs(input_class_args, classname):
                 input_class_args[key] = value
     return input_class_args
 
-def create_builtin_class_instance(classname, jsonstr_args, is_advisor=False):
-    if is_advisor:
-        if classname not in AdvisorModuleName or \
-            importlib.util.find_spec(AdvisorModuleName[classname]) is None:
-            raise RuntimeError('Advisor module is not found: {}'.format(classname))
-        class_module = importlib.import_module(AdvisorModuleName[classname])
-        class_constructor = getattr(class_module, AdvisorClassName[classname])
-    else:
-        if classname not in ModuleName or \
-            importlib.util.find_spec(ModuleName[classname]) is None:
-            raise RuntimeError('Tuner module is not found: {}'.format(classname))
-        class_module = importlib.import_module(ModuleName[classname])
-        class_constructor = getattr(class_module, ClassName[classname])
-    if jsonstr_args:
-        class_args = augment_classargs(json.loads(jsonstr_args), classname)
-    else:
-        class_args = augment_classargs({}, classname)
-    if class_args:
+
+def create_builtin_class_instance(class_name, class_args, builtin_module_dict, builtin_class_dict):
+    if class_name not in builtin_module_dict or \
+        importlib.util.find_spec(builtin_module_dict[class_name]) is None:
+        raise RuntimeError('Builtin module is not found: {}'.format(class_name))
+    class_module = importlib.import_module(builtin_module_dict[class_name])
+    class_constructor = getattr(class_module, builtin_class_dict[class_name])
+
+    if class_args is None:
+        class_args = {}
+    class_args = augment_classargs(class_args, class_name)
     instance = class_constructor(**class_args)
-    else:
-        instance = class_constructor()
+
     return instance
 
-def create_customized_class_instance(class_dir, class_filename, classname, jsonstr_args):
-    if not os.path.isfile(os.path.join(class_dir, class_filename)):
+
+def create_customized_class_instance(class_params):
+    code_dir = class_params.get('codeDir')
+    class_filename = class_params.get('classFileName')
+    class_name = class_params.get('className')
+    class_args = class_params.get('classArgs')
+
+    if not os.path.isfile(os.path.join(code_dir, class_filename)):
         raise ValueError('Class file not found: {}'.format(
-            os.path.join(class_dir, class_filename)))
-    sys.path.append(class_dir)
+            os.path.join(code_dir, class_filename)))
+    sys.path.append(code_dir)
     module_name = os.path.splitext(class_filename)[0]
     class_module = importlib.import_module(module_name)
-    class_constructor = getattr(class_module, classname)
-    if jsonstr_args:
-        class_args = json.loads(jsonstr_args)
+    class_constructor = getattr(class_module, class_name)
+
+    if class_args is None:
+        class_args = {}
     instance = class_constructor(**class_args)
-    else:
-        instance = class_constructor()
+
     return instance
 
-def parse_args():
-    parser = argparse.ArgumentParser(description='parse command line parameters.')
-    parser.add_argument('--advisor_class_name', type=str, required=False,
-                        help='Advisor class name, the class must be a subclass of nni.MsgDispatcherBase')
-    parser.add_argument('--advisor_class_filename', type=str, required=False,
-                        help='Advisor class file path')
-    parser.add_argument('--advisor_args', type=str, required=False,
-                        help='Parameters pass to advisor __init__ constructor')
-    parser.add_argument('--advisor_directory', type=str, required=False,
-                        help='Advisor directory')
-
-    parser.add_argument('--tuner_class_name', type=str, required=False,
-                        help='Tuner class name, the class must be a subclass of nni.Tuner')
-    parser.add_argument('--tuner_class_filename', type=str, required=False,
-                        help='Tuner class file path')
-    parser.add_argument('--tuner_args', type=str, required=False,
-                        help='Parameters pass to tuner __init__ constructor')
-    parser.add_argument('--tuner_directory', type=str, required=False,
-                        help='Tuner directory')
-
-    parser.add_argument('--assessor_class_name', type=str, required=False,
-                        help='Assessor class name, the class must be a subclass of nni.Assessor')
-    parser.add_argument('--assessor_args', type=str, required=False,
-                        help='Parameters pass to assessor __init__ constructor')
-    parser.add_argument('--assessor_directory', type=str, required=False,
-                        help='Assessor directory')
-    parser.add_argument('--assessor_class_filename', type=str, required=False,
-                        help='Assessor class file path')
-
-    parser.add_argument('--multi_phase', action='store_true')
-    parser.add_argument('--multi_thread', action='store_true')
-
-    flags, _ = parser.parse_known_args()
-    return flags
 
 def main():
-    '''
-    main function.
-    '''
+    parser = argparse.ArgumentParser(description='Dispatcher command line parser')
+    parser.add_argument('--exp_params', type=str, required=True)
+    args, _ = parser.parse_known_args()
+
+    exp_params_decode = base64.b64decode(args.exp_params).decode('utf-8')
+    logger.debug('decoded exp_params: [%s]', exp_params_decode)
+    exp_params = json.loads(exp_params_decode)
+    logger.debug('exp_params json obj: [%s]', json.dumps(exp_params, indent=4))
 
-    args = parse_args()
-    if args.multi_thread:
+    if exp_params.get('multiThread'):
         enable_multi_thread()
-    if args.multi_phase:
+    if exp_params.get('multiPhase'):
         enable_multi_phase()
 
-    if args.advisor_class_name:
+    if exp_params.get('advisor') is not None:
         # advisor is enabled and starts to run
-        _run_advisor(args)
-
+        _run_advisor(exp_params)
     else:
         # tuner (and assessor) is enabled and starts to run
-        tuner = _create_tuner(args)
-        if args.assessor_class_name:
-            assessor = _create_assessor(args)
+        assert exp_params.get('tuner') is not None
+        tuner = _create_tuner(exp_params)
+        if exp_params.get('assessor') is not None:
+            assessor = _create_assessor(exp_params)
         else:
             assessor = None
         dispatcher = MsgDispatcher(tuner, assessor)
@@ -139,17 +105,14 @@ def main():
             raise
 
 
-def _run_advisor(args):
-    if args.advisor_class_name in AdvisorModuleName:
+def _run_advisor(exp_params):
+    if exp_params.get('advisor').get('builtinAdvisorName') in AdvisorModuleName:
         dispatcher = create_builtin_class_instance(
-            args.advisor_class_name,
-            args.advisor_args, True)
+            exp_params.get('advisor').get('builtinAdvisorName'),
+            exp_params.get('advisor').get('classArgs'),
+            AdvisorModuleName, AdvisorClassName)
     else:
-        dispatcher = create_customized_class_instance(
-            args.advisor_directory,
-            args.advisor_class_filename,
-            args.advisor_class_name,
-            args.advisor_args)
+        dispatcher = create_customized_class_instance(exp_params.get('advisor'))
     if dispatcher is None:
         raise AssertionError('Failed to create Advisor instance')
     try:
@@ -159,33 +122,27 @@ def _run_advisor(args):
         raise
 
 
-def _create_tuner(args):
-    if args.tuner_class_name in ModuleName:
+def _create_tuner(exp_params):
+    if exp_params.get('tuner').get('builtinTunerName') in ModuleName:
         tuner = create_builtin_class_instance(
-            args.tuner_class_name,
-            args.tuner_args)
+            exp_params.get('tuner').get('builtinTunerName'),
+            exp_params.get('tuner').get('classArgs'),
+            ModuleName, ClassName)
     else:
-        tuner = create_customized_class_instance(
-            args.tuner_directory,
-            args.tuner_class_filename,
-            args.tuner_class_name,
-            args.tuner_args)
+        tuner = create_customized_class_instance(exp_params.get('tuner'))
     if tuner is None:
         raise AssertionError('Failed to create Tuner instance')
     return tuner
 
 
-def _create_assessor(args):
-    if args.assessor_class_name in ModuleName:
+def _create_assessor(exp_params):
+    if exp_params.get('assessor').get('builtinAssessorName') in ModuleName:
         assessor = create_builtin_class_instance(
-            args.assessor_class_name,
-            args.assessor_args)
+            exp_params.get('assessor').get('builtinAssessorName'),
+            exp_params.get('assessor').get('classArgs'),
+            ModuleName, ClassName)
     else:
-        assessor = create_customized_class_instance(
-            args.assessor_directory,
-            args.assessor_class_filename,
-            args.assessor_class_name,
-            args.assessor_args)
+        assessor = create_customized_class_instance(exp_params.get('assessor'))
     if assessor is None:
         raise AssertionError('Failed to create Assessor instance')
     return assessor

src/sdk/pynni/nni/compression/torch/builtin_quantizers.py

@@ -100,7 +100,7 @@ def get_bits_length(config, quant_type):
 
 
 class QAT_Quantizer(Quantizer):
-    """Quantizer using the DoReFa scheme, as defined in:
+    """Quantizer defined in:
     Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference
     http://openaccess.thecvf.com/content_cvpr_2018/papers/Jacob_Quantization_and_Training_CVPR_2018_paper.pdf
     """
@@ -227,16 +227,13 @@ class DoReFaQuantizer(Quantizer):
     (https://arxiv.org/abs/1606.06160)
     """
     def __init__(self, model, config_list):
-        """
-        config_list: supported keys:
-            - q_bits
-        """
         super().__init__(model, config_list)
 
     def quantize_weight(self, weight, config, **kwargs):
+        weight_bits = get_bits_length(config, 'weight')
         out = weight.tanh()
         out = out / (2 * out.abs().max()) + 0.5
-        out = self.quantize(out, config['q_bits'])
+        out = self.quantize(out, weight_bits)
         out = 2 * out -1
         return out
 

src/sdk/pynni/nni/compression/torch/compressor.py

@@ -250,6 +250,10 @@ class Quantizer(Compressor):
     Base quantizer for pytorch quantizer
     """
 
+    def __init__(self, model, config_list):
+        super().__init__(model, config_list)
+        self.quant_grad = QuantGrad
+
     def quantize_weight(self, weight, config, op, op_type, op_name):
         """
         quantize should overload this method to quantize weight.
@@ -262,7 +266,7 @@ class Quantizer(Compressor):
         config : dict
             the configuration for weight quantization
         """
-        raise NotImplementedError("Quantizer must overload quantize_weight()")
+        raise NotImplementedError('Quantizer must overload quantize_weight()')
 
     def quantize_output(self, output, config, op, op_type, op_name):
         """
@@ -276,7 +280,7 @@ class Quantizer(Compressor):
         config : dict
             the configuration for output quantization
         """
-        raise NotImplementedError("Quantizer must overload quantize_output()")
+        raise NotImplementedError('Quantizer must overload quantize_output()')
 
     def quantize_input(self, *inputs, config, op, op_type, op_name):
         """
@@ -290,7 +294,7 @@ class Quantizer(Compressor):
         config : dict
             the configuration for inputs quantization
         """
-        raise NotImplementedError("Quantizer must overload quantize_input()")
+        raise NotImplementedError('Quantizer must overload quantize_input()')
 
 
     def _instrument_layer(self, layer, config):
@@ -305,62 +309,93 @@ class Quantizer(Compressor):
             the configuration for quantization
         """
         assert layer._forward is None, 'Each model can only be compressed once'
-        assert "quant_types" in config, 'must provide quant_types in config'
-        assert isinstance(config["quant_types"], list), 'quant_types must be list type'
-        assert "quant_bits" in config, 'must provide quant_bits in config'
-        assert isinstance(config["quant_bits"], int) or isinstance(config["quant_bits"], dict), 'quant_bits must be dict type or int type'
+        assert 'quant_types' in config, 'must provide quant_types in config'
+        assert isinstance(config['quant_types'], list), 'quant_types must be list type'
+        assert 'quant_bits' in config, 'must provide quant_bits in config'
+        assert isinstance(config['quant_bits'], int) or isinstance(config['quant_bits'], dict), 'quant_bits must be dict type or int type'
 
-        if isinstance(config["quant_bits"], dict):
-            for quant_type in config["quant_types"]:
-                assert quant_type in config["quant_bits"], 'bits length for %s must be specified in quant_bits dict' % quant_type
+        if isinstance(config['quant_bits'], dict):
+            for quant_type in config['quant_types']:
+                assert quant_type in config['quant_bits'], 'bits length for %s must be specified in quant_bits dict' % quant_type
 
-        if 'weight' in config["quant_types"]:
+        if 'weight' in config['quant_types']:
             if not _check_weight(layer.module):
                 _logger.warning('Module %s does not have parameter "weight"', layer.name)
+            else:
+                # old_weight is used to store origin weight and weight is used to store quantized weight
+                # the reason why weight is buffer instead of parameter is because in pytorch parameter is used as leaf
+                # if weight is leaf , then old_weight can not be updated.
+                layer.module.register_parameter('old_weight', torch.nn.Parameter(layer.module.weight))
+                delattr(layer.module, 'weight')
+                layer.module.register_buffer('weight', layer.module.old_weight)
+
         layer._forward = layer.module.forward
 
         def new_forward(*inputs):
-            if 'input' in config["quant_types"]:
-                inputs = straight_through_quantize_input.apply(inputs, self, config, layer)
+            if 'input' in config['quant_types']:
+                inputs = self.quant_grad.apply(inputs, QuantType.QUANT_INPUT, self.quantize_input, config, layer)
 
-            if 'weight' in config["quant_types"] and _check_weight(layer.module):
-                weight = layer.module.weight.data
-                new_weight = self.quantize_weight(weight, config, op=layer.module, op_type=layer.type, op_name=layer.name)
-                layer.module.weight.data = new_weight
+            if 'weight' in config['quant_types'] and _check_weight(layer.module):
+                new_weight = self.quant_grad.apply(layer.module.old_weight, QuantType.QUANT_WEIGHT, self.quantize_weight, config, layer)
+                layer.module.weight = new_weight
                 result = layer._forward(*inputs)
-                layer.module.weight.data = weight
             else:
                 result = layer._forward(*inputs)
 
-            if 'output' in config["quant_types"]:
-                result = straight_through_quantize_output.apply(result, self, config, layer)
+            if 'output' in config['quant_types']:
+                result = self.quant_grad.apply(result, QuantType.QUANT_OUTPUT, self.quantize_output, config, layer)
             return result
 
         layer.module.forward = new_forward
 
+class QuantType:
+    """
+    Enum class for quantization type.
+    """
+    QUANT_INPUT = 0
+    QUANT_WEIGHT = 1
+    QUANT_OUTPUT = 2
 
-class straight_through_quantize_output(torch.autograd.Function):
+class QuantGrad(torch.autograd.Function):
+    """
+    Base class for overriding backward function of quantization operation.
+    """
     @staticmethod
-    def forward(ctx, output, quantizer, config, layer):
-        return quantizer.quantize_output(output, config, op=layer.module, op_type=layer.type, op_name=layer.name)
+    def quant_backward(tensor, grad_output, quant_type):
+        """
+        This method should be overrided by subclass to provide customized backward function,
+        default implementation is Straight-Through Estimator
 
-    @staticmethod
-    def backward(ctx, grad_output):
-        # Straight-through estimator
-        return grad_output, None, None, None
+        Parameters
+        ----------
+        tensor : Tensor
+            input of quantization operation
+        grad_output : Tensor
+            gradient of the output of quantization operation
+        quant_type : QuantType
+            the type of quantization, it can be `QuantType.QUANT_INPUT`, `QuantType.QUANT_WEIGHT`, `QuantType.QUANT_OUTPUT`,
+            you can define different behavior for different types.
 
-class straight_through_quantize_input(torch.autograd.Function):
-    @staticmethod
-    def forward(ctx, inputs, quantizer, config, layer):
-        return quantizer.quantize_input(inputs, config, op=layer.module, op_type=layer.type, op_name=layer.name)
+        Returns
+        -------
+        tensor
+            gradient of the input of quantization operation
+        """
+        return grad_output
 
     @staticmethod
-    def backward(ctx, grad_output):
-        # Straight-through estimator
-        return grad_output, None, None, None
+    def forward(ctx, tensor, quant_type, quant_func, config, layer):
+        ctx.save_for_backward(tensor, torch.Tensor([quant_type]))
+        return quant_func(tensor, config, op=layer.module, op_type=layer.type, op_name=layer.name)
+
+    @classmethod
+    def backward(cls, ctx, grad_output):
+        tensor, quant_type = ctx.saved_variables
+        output = cls.quant_backward(tensor, grad_output, quant_type)
+        return output, None, None, None, None
 
 def _check_weight(module):
     try:
-        return isinstance(module.weight, torch.nn.Parameter) and isinstance(module.weight.data, torch.Tensor)
+        return isinstance(module.weight.data, torch.Tensor)
     except AttributeError:
         return False

src/sdk/pynni/nni/env_vars.py

@@ -16,6 +16,7 @@ _trial_env_var_names = [
 ]
 
 _dispatcher_env_var_names = [
+    'SDK_PROCESS',
     'NNI_MODE',
     'NNI_CHECKPOINT_DIRECTORY',
     'NNI_LOG_DIRECTORY',

src/sdk/pynni/nni/hyperopt_tuner/hyperopt_tuner.py

@@ -11,7 +11,6 @@ import logging
 import hyperopt as hp
 import numpy as np
 from nni.tuner import Tuner
-from nni.nas_utils import rewrite_nas_space
 from nni.utils import NodeType, OptimizeMode, extract_scalar_reward, split_index
 
 logger = logging.getLogger('hyperopt_AutoML')
@@ -226,7 +225,6 @@ class HyperoptTuner(Tuner):
             return hp.anneal.suggest
         raise RuntimeError('Not support tuner algorithm in hyperopt.')
 
-    @rewrite_nas_space
     def update_search_space(self, search_space):
         """
         Update search space definition in tuner by search_space in parameters.

src/sdk/pynni/nni/nas/pytorch/callbacks.py

@@ -4,6 +4,9 @@
 import logging
 import os
 
+import torch
+import torch.nn as nn
+
 _logger = logging.getLogger(__name__)
 
 
@@ -44,11 +47,28 @@ class LRSchedulerCallback(Callback):
 
 
 class ArchitectureCheckpoint(Callback):
-    def __init__(self, checkpoint_dir, every="epoch"):
+    def __init__(self, checkpoint_dir):
+        super().__init__()
+        self.checkpoint_dir = checkpoint_dir
+        os.makedirs(self.checkpoint_dir, exist_ok=True)
+
+    def on_epoch_end(self, epoch):
+        dest_path = os.path.join(self.checkpoint_dir, "epoch_{}.json".format(epoch))
+        _logger.info("Saving architecture to %s", dest_path)
+        self.trainer.export(dest_path)
+
+
+class ModelCheckpoint(Callback):
+    def __init__(self, checkpoint_dir):
         super().__init__()
-        assert every == "epoch"
         self.checkpoint_dir = checkpoint_dir
         os.makedirs(self.checkpoint_dir, exist_ok=True)
 
     def on_epoch_end(self, epoch):
-        self.trainer.export(os.path.join(self.checkpoint_dir, "epoch_{}.json".format(epoch)))
+        if isinstance(self.model, nn.DataParallel):
+            state_dict = self.model.module.state_dict()
+        else:
+            state_dict = self.model.state_dict()
+        dest_path = os.path.join(self.checkpoint_dir, "epoch_{}.pth.tar".format(epoch))
+        _logger.info("Saving model to %s", dest_path)
+        torch.save(state_dict, dest_path)

src/sdk/pynni/nni/nas/pytorch/classic_nas/__init__.py

+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT license.
+
 from .mutator import get_and_apply_next_architecture