window sklearn regression example trial keeps waiting

[KersyTom]

Describe the issue: The example of sklearn is executed under the windows platform, and the trial status on the web-ui is always displayed as waiting

Environment:

• NNI version:2.3
• Training service (local|remote|pai|aml|etc):local
• Client OS:windows7
• Server OS (for remote mode only):
• Python version:3.6.3
• PyTorch/TensorFlow version:None
• Is conda/virtualenv/venv used?:virtualenv
• Is running in Docker?:no

Configuration:

• Experiment config (remember to remove secrets!):
• Search space: { "model_name":{"_type":"choice","_value":["LinearRegression", "Lars", "Ridge", "ARDRegression"]}, "normalize": {"_type":"choice","_value":["true", "false"]} }

Log message:normal

• nnimanager.log:
• dispatcher.log:
  • state:
• nnictl stdout and stderr:nothing
  • main.py

    
    import nni
    from nni.experiment import ExperimentConfig
    from sklearn.datasets import load_boston
    from sklearn.model_selection import train_test_split
    from sklearn import linear_model
    import logging
    import numpy as np
    from sklearn.metrics import r2_score
    from sklearn.preprocessing import StandardScaler
    from sklearn.linear_model import LinearRegression
    from sklearn.linear_model import Ridge
    from sklearn.linear_model import Lars
    from sklearn.linear_model import ARDRegression

LOG = logging.getLogger('sklearn_regression')

def load_data(): '''Load dataset, use boston dataset''' boston = load_boston() X_train, X_test, y_train, y_test = train_test_split( boston.data, boston.target, random_state=99, test_size=0.25)

normalize data

ss_X = StandardScaler()
ss_y = StandardScaler()

X_train = ss_X.fit_transform(X_train)
X_test = ss_X.transform(X_test)
y_train = ss_y.fit_transform(y_train[:, None])[:, 0]
y_test = ss_y.transform(y_test[:, None])[:, 0]

return X_train, X_test, y_train, y_test

def get_default_parameters(): '''get default parameters''' params = {'model_name': 'LinearRegression'} return params

def get_model(PARAMS): '''Get model according to parameters''' model_dict = { 'LinearRegression': LinearRegression(), 'Ridge': Ridge(), 'Lars': Lars(), 'ARDRegression': ARDRegression()

}
if not model_dict.get(PARAMS['model_name']):
    LOG.exception('Not supported model!')
    exit(1)

model = model_dict[PARAMS['model_name']]
# model.normalize = bool(PARAMS['normalize'])

return model

def run(X_train, X_test, y_train, y_test, model): '''Train model and predict result''' model.fit(X_train, y_train) predict_y = model.predict(X_test) score = r2_score(y_test, predict_y) LOG.debug('r2 score: %s', score) nni.report_final_result(score)

if name == 'main': X_train, X_test, y_train, y_test = load_data()

try:
    # get parameters from tuner
    RECEIVED_PARAMS = nni.get_next_parameter()
    LOG.debug(RECEIVED_PARAMS)
    PARAMS = get_default_parameters()
    PARAMS.update(RECEIVED_PARAMS)
    LOG.debug(PARAMS)
    model = get_model(PARAMS)
    run(X_train, X_test, y_train, y_test, model)
except Exception as exception:
    LOG.exception(exception)
    raise

- conf.yaml
```yaml
searchSpaceFile: search_space.json
trialCommand: python main.py
trialConcurrency: 1
maxTrialNumber: 30
maxExperimentDuration: 1h
tuner:
  name: TPE
  classArgs:
    optimize_mode: maximize
trainingService:  # For other platforms, check mnist-pytorch example
  platform: local
  use_active_gpu: false

How to reproduce it?:

[QuanluZhang]

@KersyTom ，could you try to set use_active_gpu true, or remove this field from conf.yaml, to check whether it works?

[KersyTom]

@QuanluZhang,hi quanlu,thanks for your reply,I have changed these configurations, and my computer does not have a GPU device, but the problem still exists, the following is my screenshot

[QuanluZhang]

@KersyTom , I cannot reproduce your problem using your provided code and configuration. my environment is

NNI version:2.3 Training service (local|remote|pai|aml|etc):local Client OS:windows10 Server OS (for remote mode only): Python version:3.8.8 PyTorch/TensorFlow version:None Is conda/virtualenv/venv used?:conda Is running in Docker?:no

[QuanluZhang]

@KersyTom , if you are still encountering this problem, please provide the complete log file. From the log you have provided above, nnimanager has received the first trial config, it is strange that this trial is waiting.

[KersyTom]

@QuanluZhang, my CLient OS is win7, I just used my colleague’s computer, and it runs successfully, i'm not sure my experiment runs failed caused by my win7 os，i will try again if i can find another win7 computer