GitLab

Designed new interface to support reusable training service, currently only applies to OpenPAI, and default disabled.

Replace trial_keeper.py to trial_runner.py, trial_runner holds an environment, and receives commands from nni manager to run or stop an trial, and return events to nni manager.
Add trial dispatcher, which inherits from original trianing service interface. It uses to share as many as possible code of all training service, and isolate with training services.
Add EnvironmentService interface to manage environment, including start/stop an environment, refresh status of environments.
Add command channel on both nni manager and trial runner parts, it supports different ways to pass messages between them. Current supported channels are file, web sockets. and supported commands from nni manager are start, kill trial, send new parameters; from runner are initialized(support some channel doesn't know which runner connected), trial end, stdout ((new type), including metric like before), version check (new type), gpu info (new type).
Add storage service to wrapper a storage to standard file operations, like NFS, azure storage and so on.
Partial support run multiple trials in parallel on runner side, but not supported by trial dispatcher side.
Other minor changes,

Add log_level to TS UT, so that UT can show debug level log.
Expose platform to start info.
Add RouterTrainingService to keep origianl OpenPAI training service, and support dynamic IOC binding.
Add more GPU info for future usage, including GPU mem total/free/used, gpu type.
Make some license information consistence.
Fix async/await problems on Array.forEach, this method doesn't support async actually.
Fix IT errors on download data, which causes by my #2484 .
Accelerate some run loop pattern by reducing sleep seconds.

Co-authored-by: liuzhe <zhe.liu@microsoft.com>

* skeleton of dlts training service (#1844)

* Hello, DLTS!

* Revert version

* Remove fs-extra

* Add some default cluster config

* schema

* fix

* Optional cluster (default to `.default`)

Depends on DLWorkspace#837

* fix

* fix

* optimize gpu type

* No more copy

* Format

* Code clean up

* Issue fix

* Add optional fields in config

* Issue fix

* Lint

* Lint

* Validate email, password and team

* Doc

* Doc fix

* Set TMPDIR

* Use metadata instead of gpu_capacity

* Cancel paused DLTS job

* workaround lint rules

* pylint

* doc
Co-authored-by: QuanluZhang <z.quanluzhang@gmail.com>

* update eslint rules
* auto fix eslint
* manually fix eslint (#1833)

* Refactor web UI to support incremental metric loading

* refactor

* Remove host job

* Move sequence ID to NNI manager

* implement incremental loading

* support specifying gpu for tuner and advisor

* PPO tuner for NAS, supports NNI's NAS interface (#1380)

* dev oneshot and darts mode, add mnist example

* fix link err in docs

* add spaces

* re-organise links for detailed descriptions of the tuners and accessors; fix link err in HpoComparision.md

* add in-page link by change .md to .html

* delete #section from cross-file links to make links work in both readthedocs and github docs

* gp_tuner init from fmfn's repo

* fix params bug by adding float>int transition

* add optimal choices; support randint&quniform type; add doc

* refine doc and code

* change mnist yml comments

* typo fix

* fix val err

* fix minimize mode err

* add config test and Hpo result

* support quniform type; update doc; update test config

* update doc

* un-commit changed in yarn.lock

* fix optimize mode bug

* optimize mode

* optimize mode

* reset pylint, gitignore

* revert .gitignore yarn.lock

* update

* dev-tf-master

* fix bugs

* fix bugs

* remove unnecessary lines

* dev oneshot and darts

* dev nas

* dev enas and oneshot

* dev enas and oneshot

* dev enas and oneshot

* dev oneshot and enas

* dev oneshot

* add ut

* add docstring

* add docstring

* fix

* resolve comments by changing docstring

* resolve comments

* Refactoring local training service
* Designated GPU for local training service
* RemoteMachine designated GPU configuration

add BOHB Advisor

1. Route tuner and assessor commands to 2 seperate queues  issue #841
2. Allow tuner to leverage intermediate result when trial is early stopped.  issue #843

If user set remoteloggingType in config file, log content will not be transmitted from trialkeeper

check nni version in trialkeeper, to make sure the version of trialkeeper is consistent with trainingService
add a debug mode in config file

* Pull code (#22)

* Support distributed job for frameworkcontroller (#612)

support distributed job for frameworkcontroller

* Multiphase doc (#519)

* multiPhase doc

* updates

* updates

* Add time parser for 'nnictl update duration' (#632)

Current nnictl update duration only support seconds unit, add a parser for this command to support {s, m, h, d}

* fix experiment state bug (#629)

* update top README.md (#622)

* Update README.md

* update (#634)

* Integration tests refactoring (#625)

* Integration test refactoring (#21) (#616)

* Integration test refactoring (#21)

* Refactoring integration tests

* test metrics

* update azure pipeline

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* updates

* update trigger

* Integration test refactoring (#618)

* updates

* updates

* update pipeline (#619)

* update pipeline

* updates

* updates

* updates

* updates

* updates

* test pipeline (#623)

* test pipeline

* updates

* updates

* updates

* Update integration test (#624)

* Update integration test

* updates

* updates

* updates

* updates

* updates

* updates

* Revert "Pull code (#22)"

This reverts commit 62fc165ad7b2ba724eead3b99f010aa34491e2c7.

* Update nnimanager logs

* updates

* Update README.md

* Revert "Update README.md"

This reverts commit bc67061160e5d57305a6e7fb63d491d12d0e9002.

* updates

* updates

* add serviceAccountName

* add serviceAccountName for frameworkcontroller

* update readme in ga_squad

* update readme

* fix typo

* Update README.md

* Update README.md

* Update README.md

* update readme

* update

* fix path

* update reference

* fix bug in config file

* update nni_arch_overview.png

* update

* update

* update

* add metis tuner code

* 1. fix bug about import 2.update other sdk file

* add auto-gbdt-example and remove unused code

* add metis_tuner into README

* update the README

* update README | remove unused variable

* fix typo

* add sklearn into requirments

* Update src/sdk/pynni/nni/metis_tuner/metis_tuner.py

add default value in __init__
Co-Authored-By: xuehui1991 <xuehui@microsoft.com>

* Update docs/HowToChooseTuner.md
Co-Authored-By: xuehui1991 <xuehui@microsoft.com>

* Update docs/HowToChooseTuner.md
Co-Authored-By: xuehui1991 <xuehui@microsoft.com>

* fix typo | add more comments

Remote TrainingService, passwd is not required when users set sshKeyPath.

Add frameworkcontroller training service based on kubeflow training service.
Refactor code structure, add kubernetes training service as father class, and set kubeflow training service and frameworkcontroller training service as child class.