fix minimize mode err

examples/trials/auto-gbdt/config.yml

 authorName: default
 experimentName: example_auto-gbdt
 trialConcurrency: 1
-maxExecDuration: 48h
-maxTrialNum: 1000
+maxExecDuration: 10h
+maxTrialNum: 10
 #choice: local, remote, pai
 trainingServicePlatform: local
 searchSpacePath: search_space.json
@@ -11,11 +11,10 @@ useAnnotation: false
 tuner:
   #choice: TPE, Random, Anneal, Evolution, BatchTuner, MetisTuner
   #SMAC (SMAC should be installed through nnictl)
-  builtinTunerName: GPTuner
+  builtinTunerName: TPE
   classArgs:
     #choice: maximize, minimize
     optimize_mode: minimize
-    cold_start_num : 1
 trial:
   command: python3 main.py
   codeDir: .

examples/trials/mnist/config.yml

@@ -11,11 +11,10 @@ useAnnotation: false
 tuner:
   #choice: TPE, Random, Anneal, Evolution, BatchTuner, MetisTuner, GPTuner
   #SMAC (SMAC should be installed through nnictl)
-  builtinTunerName: GPTuner
+  builtinTunerName: TPE
   classArgs:
     #choice: maximize, minimize
     optimize_mode: maximize
-    cold_start_num : 1
 trial:
   command: python3 mnist.py
   codeDir: .

pylintrc

@@ -15,7 +15,8 @@ max-attributes=15
 const-naming-style=any
 
 disable=duplicate-code,
-        super-init-not-called
+        super-init-not-called,
+        cell-var-from-loop
 
 # List of members which are set dynamically and missed by pylint inference
 generated-members=numpy.*,torch.*

src/sdk/pynni/nni/gp_tuner/gp_tuner.py

@@ -29,7 +29,7 @@ from sklearn.gaussian_process.kernels import Matern
 from sklearn.gaussian_process import GaussianProcessRegressor
 
 from nni.tuner import Tuner
-from nni.utils import OptimizeMode, extract_scalar_reward
+from nni.utils import extract_scalar_reward
 
 from .target_space import TargetSpace
 from .util import UtilityFunction, acq_max
@@ -42,12 +42,12 @@ class GPTuner(Tuner):
     GPTuner
     '''
 
-    def __init__(self, optimize_mode="maximize", utility_kind='ei', kappa=5, xi=0, nu=2.5, alpha=1e-6, cold_start_num=10,
-                 selection_num_warm_up=1e5, selection_num_starting_points=250):
+    def __init__(self, optimize_mode="maximize", utility='ei', kappa=5, xi=0, nu=2.5, alpha=1e-6, cold_start_num=10,
+                 selection_num_warm_up=100000, selection_num_starting_points=250):
         self.optimize_mode = optimize_mode
 
         # utility function related
-        self.utility_kind = utility_kind
+        self.utility = utility
         self.kappa = kappa
         self.xi = xi
 
@@ -107,7 +107,7 @@ class GPTuner(Tuner):
                 self._gp.fit(self._space.params, self._space.target)
 
             util = UtilityFunction(
-                kind=self.utility_kind, kappa=self.kappa, xi=self.xi)
+                kind=self.utility, kappa=self.kappa, xi=self.xi)
 
             results = acq_max(
                 f_acq=util.utility,
@@ -134,7 +134,7 @@ class GPTuner(Tuner):
             if value is dict, it should have "default" key.
         """
         value = extract_scalar_reward(value)
-        if self.optimize_mode == OptimizeMode.Minimize:
+        if self.optimize_mode == 'minimize':
             value = -value
 
         logger.info("Received trial result.")

src/sdk/pynni/nni/gp_tuner/util.py

@@ -38,7 +38,7 @@ def _match_val_type(vals, bounds):
         if _type == "choice":
             # Find the closest integer in the array, vals_bounds
             vals_new.append(
-                min(bounds[i]['_value'], key=lambda x: abs(x - vals[i])))
+                min(bound['_value'], key=lambda x: abs(x - vals[i])))
         elif _type in ['quniform', 'randint']:
             vals_new.append(np.around(vals[i]))
         else:
@@ -82,11 +82,12 @@ def acq_max(f_acq, gp, y_max, bounds, space, num_warmup, num_starting_points):
 
     # Warm up with random points
     x_tries = [space.random_sample()
-               for _ in range(int(num_warmup))]
+               for _ in range(num_warmup)]
     ys = f_acq(x_tries, gp=gp, y_max=y_max)
     x_max = x_tries[ys.argmax()]
     max_acq = ys.max()
 
+
     # Explore the parameter space more throughly
     x_seeds = [space.random_sample() for _ in range(num_starting_points)]
 
@@ -156,7 +157,7 @@ class UtilityFunction():
     def _ei(x, gp, y_max, xi):
         with warnings.catch_warnings():
             warnings.simplefilter("ignore")
-            mean, std = gp.predict(x, return_std=True)
+            mean, std = gp.predict(x, return_std=True)  # TODO: sample_y ??
 
         z = (mean - y_max - xi)/std
         return (mean - y_max - xi) * norm.cdf(z) + std * norm.pdf(z)