def learning_model_factory(weights=None):
return LearningModel(pcDataset,
trainingDataset,
batch=batchSize,
epocs=nEpocs,
vf=validation_fraction,
hidden=hiddenLayers,
activation=activation,
weights=weights)
def learning_model_factory(weights=None):
return LearningModel(inputDataset,
trainingDataset,
layers,
batch=batchSize,
lrate=learnRate,
stop_condition=stopCondition,
loss_function=loss_function,
momentum=momentum,
decay=decay,
nesterov=nesterov,
orthoWts=orthoWts,
epocs=nEpocs,
vf=validation_fraction,
weights=weights)
def direct_learning_model_factory_trainable():
layers3[0].setTrainable(True)
return LearningModel(reanalysisInputDataset,
trainingDataset,
layers3,
verbose=False,
eager=eagerExe,
batch=batchSize,
earlyTermIndex=earlyTermIndex,
lrate=learnRate,
stop_condition=stopCondition,
shuffle=shuffle,
loss_function=loss_function,
momentum=momentum,
decay=decay,
nesterov=nesterov,
epocs=nEpocs,
vf=validation_fraction,
weights=result.final_weights)
def pc_learning_model_factory(inputs=pcInputDataset,
target=trainingDataset,
weights=None):
return LearningModel(inputs,
target,
layers2,
verbose=False,
eager=eagerExe,
batch=batchSize,
earlyTermIndex=earlyTermIndex,
lrate=learnRate,
stop_condition=stopCondition,
shuffle=shuffle,
loss_function=loss_function,
momentum=momentum,
decay=decay,
nesterov=nesterov,
orthoWts=orthoWts,
epocs=nEpocs,
vf=validation_fraction,
weights=weights)
learning_range = CTimeRange.new( "1980-1-1", "2005-12-30" )
prediction_lag = CDuration.months(1)
nInterationsPerProc = 10
batchSize = 100
nEpocs = 200
validation_fraction = 0.15
hiddenLayers = [100]
activation = "relu"
plotPrediction = True
variables = [ Variable("ts"), Variable( "zg", 50000 ) ]
project = Project(outDir,projectName)
pcDataset = ClimateleDataset([Experiment(project, start_year, end_year, nModes, variable) for variable in variables], nts = nTS, smooth = smooth, timeRange = learning_range)
td = IITMDataSource( "AI", "monthly" )
trainingDataset = TrainingDataset.new( [td], pcDataset, prediction_lag, decycle=True )
#ref_time_range = ( "1980-1-1", "2014-12-1" )
#ref_ts = ProjectDataSource( "HadISST_1.cvdp_data.1980-2017", [ "nino34" ], ref_time_range )
def learning_model_factory( weights = None ):
return LearningModel( pcDataset, trainingDataset, batch=batchSize, epocs=nEpocs, vf=validation_fraction, hidden=hiddenLayers, activation=activation, weights=weights )
result = LearningModel.parallel_execute( learning_model_factory, nInterationsPerProc )
print "Got Best result, valuation loss = " + str( result.val_loss ) + " training loss = " + str( result.train_loss )
if plotPrediction:
plot_title = "Training data with Prediction ({0}->IITM-AI, lag {1}) {2}-{3} (loss: {4}, Epochs: {5})".format(pcDataset.getVariableIds(),prediction_lag,start_year,end_year,result.val_loss,result.nEpocs)
learningModel = learning_model_factory()
learningModel.plotPrediction( result, plot_title )
from cliMLe.trainingData import *
from cliMLe.learning import FitResult, LearningModel, ActivationTarget
import matplotlib.pyplot as plt
strong_monsoon = ActivationTarget(3.0, "Strong Monsoon")
weak_monsoon = ActivationTarget(-3.0, "Weak Monsoon")
instance = "ams-2"
activation_target = strong_monsoon
learningModel, model, result = LearningModel.getActivationBackProjection(
instance, activation_target.value)
inputs = learningModel.inputs
fig, ax = plt.subplots()
y_pos = np.arange(result.shape[0])
plt.bar(y_pos, result, align='center', alpha=0.5)
plt.ylabel('Activation')
plt.title(activation_target.title + ' Activtion Plot')
plt.show()
from cliMLe.trainingData import *
from cliMLe.learning import FitResult, LearningModel
instance = "good_1"
(learningModel, result) = LearningModel.loadInstance(instance)
varIds = learningModel.inputs.getVariableIds()
trainingId = learningModel.outputs.id()
plot_title = "Training data with Prediction ({0}->{1}) (loss: {2}, Epochs: {3})".format(
varIds, trainingId, result.val_loss, result.nEpocs)
learningModel.plotPrediction(
result, "All-India Monsoon Rainfall Prediction, 1 Year Lag Time")
from cliMLe.learning import LearningModel
from cliMLe.analytics import Visualizer
instances = ["ams-zg-16-16-" + str(index) for index in range(1, 11)]
target_values = [3.0, -3.0]
results = {}
for iT, target_value in enumerate(target_values):
for iI, instance in enumerate(instances):
learningModel, model, weights = LearningModel.getActivationBackProjection(
instance, target_value)
patterns = Visualizer.getWeightedPatterns(learningModel, weights)
if iI == 0:
results[iT] = patterns[0]
else:
results[iT] = results[iT] + patterns[0]
results[2] = results[0] - results[1]
Visualizer.mpl_spaceplot(results.values(), [
"500 mbar Height - Strong Monsoon", "500 mbar Height - Weak Monsoon",
"Difference"
])
earlyTermIndex=earlyTermIndex,
lrate=learnRate,
stop_condition=stopCondition,
shuffle=shuffle,
loss_function=loss_function,
momentum=momentum,
decay=decay,
nesterov=nesterov,
orthoWts=orthoWts,
epocs=nEpocs,
vf=validation_fraction,
weights=weights)
learning_model_factory = direct_learning_model_factory
result = LearningModel.fit(learning_model_factory, nInterationsPerProc,
nShuffles, parallelExe)
def direct_learning_model_factory_trainable():
layers3[0].setTrainable(True)
return LearningModel(reanalysisInputDataset,
trainingDataset,
layers3,
verbose=False,
eager=eagerExe,
batch=batchSize,
earlyTermIndex=earlyTermIndex,
lrate=learnRate,
stop_condition=stopCondition,
shuffle=shuffle,
loss_function=loss_function,
end_year = 2015
nModes = 32
nTS = 1
smooth = 0
learning_range = CTimeRange.new( "1980-1-1", "2014-12-30" )
prediction_lag = CDuration.months(6)
nInterations = 10
batchSize = 100
nEpocs = 200
validation_fraction = 0.15
hiddenLayers = [100]
activation = "relu"
plotPrediction = True
variables = [ Variable("ts") ] # , Variable( "zg", 50000 ) ]
project = Project(outDir,projectName)
pcDataset = ClimateleDataset([Experiment(project, start_year, end_year, nModes, variable) for variable in variables], nts = nTS, smooth = smooth, timeRange = learning_range)
td = ProjectDataSource( "HadISST_1.cvdp_data.1980-2017", [ "nino34" ] ) # , "pdo_timeseries_mon", "indian_ocean_dipole", "nino34"
trainingDataset = TrainingDataset.new( [td], pcDataset, prediction_lag )
def learning_model_factory():
return LearningModel( pcDataset, trainingDataset, batch=batchSize, epocs=nEpocs, vf=validation_fraction, hidden=hiddenLayers, activation=activation )
result = LearningModel.serial_execute( learning_model_factory, nInterations )
print "Got Best result, val_loss = " + str( result.val_loss )
if plotPrediction:
plot_title = "Training data with Prediction ({0}->nino34, lag {1}) {2}-{3} ({4} Epochs)".format(pcDataset.getVariableIds(),prediction_lag,start_year,end_year,nEpocs)
learningModel = learning_model_factory( )
learningModel.plotPrediction( result, plot_title )