Remove outputDir and dataDir in config file #1361
Conversation
merge master
merge master
Delete the reference and fix the broken links (microsoft#1320)
merge master
| @@ -32,8 +32,6 @@ trial: | |||
| cpuNum: 1 | |||
| memoryMB: 8196 | |||
| image: msranni/nni:latest | |||
There was a problem hiding this comment.
so I want to output my models, where is the output path?
There was a problem hiding this comment.
PAI does not have dataDir and outputDir in their doc anymore, so we remove these two path accordingly. In other training modes, users use OUTPUTDIR in trial code to output the data they want, @SparkSnail could you check the content of OUTPUTDIR in PAI mode?
There was a problem hiding this comment.
Users could use PAI_OUTPUT_DIR to get the output directory of PAI, the defaulet value is $PAI_DEFAULT_FS_URI/Output/$jobName, refer https://github.com/microsoft/pai/blob/b2324866d0280a2d22958717ea6025740f71b9f0/docs/job_tutorial.md.
In NNI, we use a variable NNI_OUTPUT_DIR to store the log data of trial_keeper, and the data will be uploaded to hdfs after trial is finished, so users may use os.environ['NNI_OUTPUT_DIR'] in their code to store model data, the data will be uploaded to hdfs finally. refer https://github.com/microsoft/nni/blob/master/docs/en_US/TrainingService/PaiMode.md#run-an-experiment
| // dataDir | ||
| this.paiTrialConfig.dataDir, | ||
| // outputDir | ||
| this.paiTrialConfig.outputDir, |
There was a problem hiding this comment.
so in the previous version, outputDir in nni config is also outputDir in pai job config?
tools/nni_cmd/launcher_utils.py
Outdated
| @@ -253,6 +253,12 @@ def validate_pai_trial_conifg(experiment_config): | |||
| experiment_config['trial']['shmMB'] > experiment_config['trial']['memoryMB']: | |||
| print_error('shmMB should be no more than memoryMB!') | |||
| exit(1) | |||
| #backward compatibility | |||
| warning_information = '{0} is not supported in NNI anymore, please remove the field in config file!' | |||
There was a problem hiding this comment.
could add a little more information, for example, "please refer to the link for the practices of how to get data and output model in trial code".
| @@ -80,9 +74,10 @@ And you will be redirected to HDFS web portal to browse the output files of that | |||
|
|
|||
| You can see there're three fils in output folder: stderr, stdout, and trial.log | |||
|
|
|||
| If you also want to save trial's other output into HDFS, like model files, you can use environment variable `NNI_OUTPUT_DIR` in your trial code to save your own output files, and NNI SDK will copy all the files in `NNI_OUTPUT_DIR` from trial's container to HDFS. | |||
| ## data management | |||
| If your training data is not to large, the could be put into codeDir, and nni will upload the data to hdfs, or you could build your own docker image with the data. If you have large dataset, it's not appropriate to put the data in codeDir, and you could follow the [guidance](https://github.com/microsoft/pai/blob/master/docs/user/storage.md) to mount the data folder in container. | |||
|  | ||
|
|
||
| Any problems when using NNI in pai mode, please create issues on [NNI github repo](https://github.com/Microsoft/nni). |
There was a problem hiding this comment.
Suggest to remove this sentence, the process to create issue is identical for all nni component, no difference.
#1342