* Added test directory and core file to gitignore.

Closes #601.

* Fixed pipeline fit parameters.

Closes #600.

* Reverted changes to gitignore.

* support latest xgboost version

* Update test_classification.py

* Update 

Exists problems when installing xgb1.6.1 in py3.6

* cleanup

* xgboost version

* remove time_budget_s in test

* remove redundancy

* stop support of python 3.6

Co-authored-by: zsk <shaokunzhang529@gmail.com>
Co-authored-by: Qingyun Wu <qingyun.wu@psu.edu>

* update the signature of set_search_properties

* Update Installation.md

* trying to fix the indexerror for ner

* fix resource limit issue

* add python 3.10 in CI

* reinstall libomp in macos

* fixing a bug for ner

* backup & recover global vars for nested tune.run

* ensure recovering global vars before return

* test reproducibility

* cost_attr, _eps and penalty as instance variables

* enable ensemble when using ray

* sanitize config

* init value type match

* bump version to 1.0.6

* add a note about flaml version in notebook

* add note about mismatched ITER_HP

* catch SSLError when accessing OpenML data

* catch errors in autovw test

Co-authored-by: Qingyun Wu <qingyun.wu@psu.edu>

* Update website/docs/FAQ.md

* set holiday version <0.14 for prophet

* bump version to 1.0.5

* add zeroshot notebook

* Update Task-Oriented-AutoML.md

* fix checkpoint naming + trial id for non-ray mode, fix the bug in running test mode, delete all the checkpoints in non-ray mode

* finished testing for checkpoint naming, delete checkpoint, ray, max iter = 1

* update doc about scheduler exception

* remove assert

* Update website/docs/FAQ.md

* simplify warmstart in blendsearch

* bump version to 1.0.3

* use zeroshot when no budget is given; custom_hp

* update Getting-Started

* protobuf version

* X_val

Added one more use case of tune.

* add bs restore test

* use default metric when not provided

* update documentation

* remove print

* period

* remove bs restore test

* Update website/docs/Use-Cases/Task-Oriented-AutoML.md

Issue I encountered:

#542 run test_restore.py and got _pickle.UnpicklingError: state is not a dictionary

I observed:

1. numpy version
  i. When numpy==1.16*, np.random.RandomState.__getstate__() returns a tuple, not a dict.
    _pickle.UnpicklingError occurs
  ii. When numpy>1.17.0rc1, it returns a dict;
    _pickle.UnpicklingError does not occur
  iii. When numpy>1.17.0rc1, flaml uses np_random_generator = np.random.Generator,
    _pickle.UnpicklingError does not occur

2. class _BackwardsCompatibleNumpyRng
  When I remove func _BackwardsCompatibleNumpyRng.__getattr__() , _pickle.UnpicklingError doesn't occur (regardless of numpy version == 1.16* or 1.17*)

To sum up,
I think making modifications to class _BackwardsCompatibleNumpyRng is not a good choice (_BackwardsCompatibleNumpyRng came from ray）and we still need to learn more about the operation mechanism of pickle.

So I upgraded the numpy version that flaml requires：
  setup.py："NumPy>=1.17.0rc1"

* align indent and add missing quotation

* fix indentation in flaml.py

* fix doc string; enable label transform in automl.score

* set n_jobs in ensemble dict

* catch the ensemble error

* choose n_jobs for stacker

* clarify

* fixing roberta add_prefix_space bug

* fix conll bug

* update DataCollatorForAuto

* adding label_list comments

* handle non-flaml scheduler in flaml.tune

* revise time budget

* Update website/docs/Use-Cases/Tune-User-Defined-Function.md

Co-authored-by: Chi Wang <wang.chi@microsoft.com>

* Update website/docs/Use-Cases/Tune-User-Defined-Function.md

Co-authored-by: Chi Wang <wang.chi@microsoft.com>

* Update flaml/tune/tune.py

Co-authored-by: Chi Wang <wang.chi@microsoft.com>

* add docstr

* remove random seed

* StopIteration

* StopIteration format

* format

* Update flaml/tune/tune.py

Co-authored-by: Chi Wang <wang.chi@microsoft.com>

* revise docstr

Co-authored-by: Chi Wang <wang.chi@microsoft.com>

* fixing use_ray

* Bump cross-fetch from 3.1.4 to 3.1.5 in /website

* Bump async from 2.6.3 to 2.6.4 in /website

* cleanup

* refactoring TransformersEstimator to support default and custom_hp

* handling starting_points not in search space

* addressing starting point more than max_iter

* fixing upper < lower bug

* allow evaluated_rewards shorter than points_to_evaluate

* docstr update

close #520

* handling fractional gpu_per_trial

* Bump minimist from 1.2.5 to 1.2.6 in /website

* check best_config in test