pylearn: learner.py annotate

annotate learner.py @ 143:b7ca3545186b

added missing raise

author	Frederic Bastien <bastienf@iro.umontreal.ca>
date	Mon, 12 May 2008 15:08:09 -0400
parents	3f4e5c9bdc5e
children	ceae4de18981

rev	line source
1 2cd82666b9a7 Added statscollector and started writing dataset and learner. bengioy@esprit.iro.umontreal.ca parents: 0 diff changeset	1
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	2 from dataset import AttributesHolder,AbstractFunction,ApplyFunctionDataSet,DataSet,CachedDataSet
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	3 import theano
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	4 from theano import compile
129 4c2280edcaf5 Fixed typos in learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 128 diff changeset	5 from theano import tensor as t
1 2cd82666b9a7 Added statscollector and started writing dataset and learner. bengioy@esprit.iro.umontreal.ca parents: 0 diff changeset	6
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	7 class Learner(AttributesHolder):
132 f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	8 """
f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	9 Base class for learning algorithms, provides an interface
1 2cd82666b9a7 Added statscollector and started writing dataset and learner. bengioy@esprit.iro.umontreal.ca parents: 0 diff changeset	10 that allows various algorithms to be applicable to generic learning
2cd82666b9a7 Added statscollector and started writing dataset and learner. bengioy@esprit.iro.umontreal.ca parents: 0 diff changeset	11 algorithms.
2cd82666b9a7 Added statscollector and started writing dataset and learner. bengioy@esprit.iro.umontreal.ca parents: 0 diff changeset	12
132 f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	13 A L{Learner} can be seen as a learning algorithm, a function that when
10 80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	14 applied to training data returns a learned function, an object that
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	15 can be applied to other data and return some output data.
1 2cd82666b9a7 Added statscollector and started writing dataset and learner. bengioy@esprit.iro.umontreal.ca parents: 0 diff changeset	16 """
2cd82666b9a7 Added statscollector and started writing dataset and learner. bengioy@esprit.iro.umontreal.ca parents: 0 diff changeset	17
10 80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	18 def __init__(self):
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	19 pass
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	20
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	21 def forget(self):
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	22 """
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	23 Reset the state of the learner to a blank slate, before seeing
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	24 training data. The operation may be non-deterministic if the
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	25 learner has a random number generator that is set to use a
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	26 different seed each time it forget() is called.
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	27 """
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	28 raise NotImplementedError
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	29
14 5ede27026e05 Working on gradient_based_learner bengioy@bengiomac.local parents: 13 diff changeset	30 def update(self,training_set,train_stats_collector=None):
10 80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	31 """
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	32 Continue training a learner, with the evidence provided by the given training set.
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	33 Hence update can be called multiple times. This is particularly useful in the
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	34 on-line setting or the sequential (Bayesian or not) settings.
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	35 The result is a function that can be applied on data, with the same
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	36 semantics of the Learner.use method.
75 90e4c0784d6e Added draft of LinearRegression learner bengioy@bengiomac.local parents: 20 diff changeset	37
132 f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	38 The user may optionally provide a training L{StatsCollector} that is used to record
75 90e4c0784d6e Added draft of LinearRegression learner bengioy@bengiomac.local parents: 20 diff changeset	39 some statistics of the outputs computed during training. It is update(d) during
90e4c0784d6e Added draft of LinearRegression learner bengioy@bengiomac.local parents: 20 diff changeset	40 training.
10 80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	41 """
13 633453635d51 Starting to work on gradient_based_learner.py bengioy@bengiomac.local parents: 10 diff changeset	42 return self.use # default behavior is 'non-adaptive', i.e. update does not do anything
10 80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	43
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	44
20 266c68cb6136 Minor editions, plus adding untested ApplyFunctionDataset for GradientLearner in the works. bengioy@bengiomac.local parents: 14 diff changeset	45 def __call__(self,training_set,train_stats_collector=None):
10 80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	46 """
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	47 Train a learner from scratch using the provided training set,
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	48 and return the learned function.
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	49 """
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	50 self.forget()
133 b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	51 return self.update(training_set,train_stats_collector)
1 2cd82666b9a7 Added statscollector and started writing dataset and learner. bengioy@esprit.iro.umontreal.ca parents: 0 diff changeset	52
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	53 def use(self,input_dataset,output_fieldnames=None,
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	54 test_stats_collector=None,copy_inputs=True,
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	55 put_stats_in_output_dataset=True,
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	56 output_attributes=[]):
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	57 """
132 f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	58 Once a L{Learner} has been trained by one or more call to 'update', it can
f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	59 be used with one or more calls to 'use'. The argument is an input L{DataSet} (possibly
f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	60 containing a single example) and the result is an output L{DataSet} of the same length.
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	61 If output_fieldnames is specified, it may be use to indicate which fields should
132 f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	62 be constructed in the output L{DataSet} (for example ['output','classification_error']).
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	63 Otherwise, self.defaultOutputFields is called to choose the output fields.
20 266c68cb6136 Minor editions, plus adding untested ApplyFunctionDataset for GradientLearner in the works. bengioy@bengiomac.local parents: 14 diff changeset	64 Optionally, if copy_inputs, the input fields (of the input_dataset) can be made
132 f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	65 visible in the output L{DataSet} returned by this method.
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	66 Optionally, attributes of the learner can be copied in the output dataset,
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	67 and statistics computed by the stats collector also put in the output dataset.
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	68 Note the distinction between fields (which are example-wise quantities, e.g. 'input')
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	69 and attributes (which are not, e.g. 'regularization_term').
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	70
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	71 We provide here a default implementation that does all this using
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	72 a sub-class defined method: minibatchwiseUseFunction.
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	73
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	74 @todo check if some of the learner attributes are actually SPECIFIED
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	75 as attributes of the input_dataset, and if so use their values instead
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	76 of the ones in the learner.
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	77
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	78 The learner tries to compute in the output dataset the output fields specified.
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	79 If None is specified then self.defaultOutputFields(input_dataset.fieldNames())
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	80 is called to determine the output fields.
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	81
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	82 Attributes of the learner can also optionally be copied into the output dataset.
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	83 If output_attributes is None then all of the attributes in self.AttributeNames()
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	84 are copied in the output dataset, but if it is [] (the default), then none are copied.
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	85 If a test_stats_collector is provided, then its attributes (test_stats_collector.AttributeNames())
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	86 are also copied into the output dataset attributes.
10 80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners. bengioy@esprit.iro.umontreal.ca parents: 1 diff changeset	87 """
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	88 minibatchwise_use_function = self.minibatchwiseUseFunction(input_dataset.fieldNames(),
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	89 output_fieldnames,
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	90 test_stats_collector)
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	91 virtual_output_dataset = ApplyFunctionDataSet(input_dataset,
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	92 minibatchwise_use_function,
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	93 True,DataSet.numpy_vstack,
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	94 DataSet.numpy_hstack)
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	95 # actually force the computation
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	96 output_dataset = CachedDataSet(virtual_output_dataset,True)
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	97 if copy_inputs:
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	98 output_dataset = input_dataset \| output_dataset
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	99 # copy the wanted attributes in the dataset
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	100 if output_attributes is None:
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	101 output_attributes = self.attributeNames()
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	102 if output_attributes:
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	103 assert set(attribute_names) <= set(self.attributeNames())
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	104 output_dataset.setAttributes(output_attributes,
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	105 self.names2attributes(output_attributes,return_copy=True))
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	106 if test_stats_collector:
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	107 test_stats_collector.update(output_dataset)
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	108 if put_stats_in_output_dataset:
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	109 output_dataset.setAttributes(test_stats_collector.attributeNames(),
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	110 test_stats_collector.attributes())
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	111 return output_dataset
77 1e2bb5bad636 toying with different ways to implement learners bengioy@bengiomac.local parents: 75 diff changeset	112
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	113 def minibatchwiseUseFunction(self, input_fields, output_fields, stats_collector):
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	114 """
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	115 Returns a function that can map the given input fields to the given output fields
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	116 and to the attributes that the stats collector needs for its computation.
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	117 That function is expected to operate on minibatches.
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	118 The function returned makes use of the self.useInputAttributes() and
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	119 sets the attributes specified by self.useOutputAttributes().
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	120 """
143 b7ca3545186b added missing raise Frederic Bastien <bastienf@iro.umontreal.ca> parents: 134 diff changeset	121 raise AbstractFunction()
b7ca3545186b added missing raise Frederic Bastien <bastienf@iro.umontreal.ca> parents: 134 diff changeset	122
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	123 def attributeNames(self):
78 3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	124 """
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	125 A Learner may have attributes that it wishes to export to other objects. To automate
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	126 such export, sub-classes should define here the names (list of strings) of these attributes.
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	127
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	128 @todo By default, attributeNames looks for all dictionary entries whose name does not start with _.
78 3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	129 """
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	130 return []
77 1e2bb5bad636 toying with different ways to implement learners bengioy@bengiomac.local parents: 75 diff changeset	131
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	132 def attributes(self,return_copy=False):
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	133 """
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	134 Return a list with the values of the learner's attributes (or optionally, a deep copy).
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	135 """
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	136 return self.names2attributes(self.attributeNames(),return_copy)
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	137
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	138 def names2attributes(self,names):
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	139 """
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	140 Private helper function that maps a list of attribute names to a list
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	141 of (optionally copies) values of attributes.
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	142 """
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	143 res=[]
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	144 for name in names:
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	145 assert name in names
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	146 res.append(self.__getattribute__(name))
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	147 return res
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	148
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	149 def useInputAttributes(self):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	150 """
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	151 A subset of self.attributeNames() which are the names of attributes needed by use() in order
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	152 to do its work.
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	153 """
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	154 raise AbstractFunction()
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	155
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	156 def useOutputAttributes(self):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	157 """
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	158 A subset of self.attributeNames() which are the names of attributes modified/created by use() in order
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	159 to do its work.
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	160 """
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	161 raise AbstractFunction()
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	162
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	163
78 3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	164 class TLearner(Learner):
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	165 """
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	166 TLearner is a virtual class of Learners that attempts to factor out of the definition
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	167 of a learner the steps that are common to many implementations of learning algorithms,
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	168 so as to leave only 'the equations' to define in particular sub-classes, using Theano.
78 3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	169
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	170 In the default implementations of use and update, it is assumed that the 'use' and 'update' methods
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	171 visit examples in the input dataset sequentially. In the 'use' method only one pass through the dataset is done,
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	172 whereas the sub-learner may wish to iterate over the examples multiple times. Subclasses where this
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	173 basic model is not appropriate can simply redefine update or use.
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	174
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	175 Sub-classes must provide the following functions and functionalities:
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	176 - attributeNames(): defines all the names of attributes which can be used as fields or
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	177 attributes in input/output datasets or in stats collectors.
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	178 All these attributes are expected to be theano.Result objects
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	179 (with a .data property and recognized by theano.Function for compilation).
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	180 The sub-class constructor defines the relations between
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	181 the Theano variables that may be used by 'use' and 'update'
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	182 or by a stats collector.
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	183 - defaultOutputFields(input_fields): return a list of default dataset output fields when
3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	184 None are provided by the caller of use.
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	185 The following naming convention is assumed and important.
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	186 Attributes whose names are listed in attributeNames() can be of any type,
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	187 but those that can be referenced as input/output dataset fields or as
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	188 output attributes in 'use' or as input attributes in the stats collector
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	189 should be associated with a Theano Result variable. If the exported attribute
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	190 name is <name>, the corresponding Result name (an internal attribute of
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	191 the TLearner, created in the sub-class constructor) should be _<name>.
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	192 Typically <name> will be numpy ndarray and _<name> will be the corresponding
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	193 Theano Tensor (for symbolic manipulation).
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	194
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	195 @todo pousser dans Learner toute la poutine qui peut l'etre sans etre
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	196 dependant de Theano
78 3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	197 """
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	198
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	199 def __init__(self):
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	200 Learner.__init__(self)
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	201 self.use_functions_dictionary={}
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	202
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	203 def defaultOutputFields(self, input_fields):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	204 """
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	205 Return a default list of output field names (to put in the output dataset).
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	206 This will be used when None are provided (as output_fields) by the caller of the 'use' method.
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	207 This may involve looking at the input_fields (names) available in the
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	208 input_dataset.
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	209 """
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	210 raise AbstractFunction()
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	211
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	212 def minibatchwiseUseFunction(self, input_fields, output_fields, stats_collector):
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	213 """
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	214 Implement minibatchwiseUseFunction by exploiting Theano compilation
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	215 and the expression graph defined by a sub-class constructor.
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	216 """
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	217 if not output_fields:
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	218 output_fields = self.defaultOutputFields(input_fields)
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	219 if stats_collector:
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	220 stats_collector_inputs = stats_collector.input2UpdateAttributes()
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	221 for attribute in stats_collector_inputs:
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	222 if attribute not in input_fields:
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	223 output_fields.append(attribute)
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	224 key = (tuple(input_fields),tuple(output_fields))
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	225 if key not in self.use_functions_dictionary:
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	226 use_input_attributes = self.useInputAttributes()
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	227 use_output_attributes = self.useOutputAttributes()
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	228 complete_f = compile.function(self.names2OpResults(input_fields+use_input_attributes),
d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	229 self.names2OpResults(output_fields+use_output_attributes))
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	230 def f(*input_field_values):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	231 input_attribute_values = self.names2attributes(use_input_attributes)
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	232 results = complete_f(*(list(input_field_values) + input_attribute_values))
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	233 output_field_values = results[0:len(output_fields)]
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	234 output_attribute_values = results[len(output_fields):len(results)]
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	235 if use_output_attributes:
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	236 self.setAttributes(use_output_attributes,output_attribute_values)
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	237 return output_field_values
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	238 self.use_functions_dictionary[key]=f
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	239 return self.use_functions_dictionary[key]
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	240
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	241 def names2OpResults(self,names):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	242 """
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	243 Private helper function that maps a list of attribute names to a list
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	244 of corresponding Op Results (with the same name but with a '_' prefix).
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	245 """
133 b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	246 return [self.__getattribute__('_'+name) for name in names]
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	247
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	248
111 88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	249 class MinibatchUpdatesTLearner(TLearner):
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	250 """
132 f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	251 This adds to L{TLearner} a
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	252 - updateStart(), updateEnd(), updateMinibatch(minibatch), isLastEpoch():
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	253 functions executed at the beginning, the end, in the middle
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	254 (for each minibatch) of the update method, and at the end
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	255 of each epoch. This model only
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	256 works for 'online' or one-shot learning that requires
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	257 going only once through the training data. For more complicated
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	258 models, more specialized subclasses of TLearner should be used
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	259 or a learning-algorithm specific update method should be defined.
111 88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	260
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	261 - a 'parameters' attribute which is a list of parameters (whose names are
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	262 specified by the user's subclass with the parameterAttributes() method)
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	263
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	264 """
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	265
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	266 def __init__(self):
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	267 TLearner.__init__(self)
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	268 self.update_minibatch_function = compile.function(self.names2OpResults(self.updateMinibatchOutputAttributes()+
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	269 self.updateMinibatchInputFields()),
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	270 self.names2OpResults(self.updateMinibatchOutputAttributes()))
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	271 self.update_end_function = compile.function(self.names2OpResults(self.updateEndInputAttributes()),
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	272 self.names2OpResults(self.updateEndOutputAttributes()))
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	273
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	274 def allocate(self, minibatch):
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	275 """
132 f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	276 This function is called at the beginning of each L{updateMinibatch}
128 ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	277 and should be used to check that all required attributes have been
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	278 allocated and initialized (usually this function calls forget()
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	279 when it has to do an initialization).
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	280 """
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	281 raise AbstractFunction()
ee5507af2c60 minor edits Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 126 diff changeset	282
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	283 def updateMinibatchInputFields(self):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	284 raise AbstractFunction()
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	285
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	286 def updateMinibatchInputAttributes(self):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	287 raise AbstractFunction()
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	288
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	289 def updateMinibatchOutputAttributes(self):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	290 raise AbstractFunction()
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	291
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	292 def updateEndInputAttributes(self):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	293 raise AbstractFunction()
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	294
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	295 def updateEndOutputAttributes(self):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	296 raise AbstractFunction()
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	297
111 88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	298 def parameterAttributes(self):
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	299 raise AbstractFunction()
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	300
133 b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	301 def updateStart(self,training_set):
b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	302 pass
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	303
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	304 def updateEnd(self):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	305 self.setAttributes(self.updateEndOutputAttributes(),
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	306 self.update_end_function(*self.names2attributes(self.updateEndInputAttributes())))
111 88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	307 self.parameters = self.names2attributes(self.parameterAttributes())
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	308
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	309 def updateMinibatch(self,minibatch):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	310 # make sure all required fields are allocated and initialized
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	311 self.allocate(minibatch)
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	312 input_attributes = self.names2attributes(self.updateMinibatchInputAttributes())
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	313 input_fields = minibatch(*self.updateMinibatchInputFields())
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	314 self.setAttributes(self.updateMinibatchOutputAttributes(),
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	315 # concatenate the attribute values and field values and then apply update fn
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	316 self.update_minibatch_function(*(input_attributes+input_fields)))
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	317
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	318 def isLastEpoch(self):
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	319 """
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	320 This method is called at the end of each epoch (cycling over the training set).
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	321 It returns a boolean to indicate if this is the last epoch.
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	322 By default just do one epoch.
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	323 """
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	324 return True
78 3499918faa9d In the middle of designing TLearner bengioy@bengiomac.local parents: 77 diff changeset	325
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	326 def update(self,training_set,train_stats_collector=None):
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	327 """
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	328 @todo check if some of the learner attributes are actually SPECIFIED
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	329 in as attributes of the training_set.
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	330 """
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	331 self.updateStart(training_set)
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	332 stop=False
133 b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	333 if hasattr(self,'_minibatch_size') and self._minibatch_size:
b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	334 minibatch_size=self._minibatch_size
b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	335 else:
b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	336 minibatch_size=min(100,len(training_set))
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	337 while not stop:
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	338 if train_stats_collector:
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	339 train_stats_collector.forget() # restart stats collectin at the beginning of each epoch
133 b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	340 for minibatch in training_set.minibatches(minibatch_size=minibatch_size):
b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	341 self.updateMinibatch(minibatch)
107 c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	342 if train_stats_collector:
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	343 minibatch_set = minibatch.examples()
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	344 minibatch_set.setAttributes(self.attributeNames(),self.attributes())
c4916445e025 Comments from Pascal V. Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 92 diff changeset	345 train_stats_collector.update(minibatch_set)
110 8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	346 stop = self.isLastEpoch()
8fa1ef2411a0 Worked on OneShotTLearner and implementation of LinearRegression bengioy@bengiomac.local parents: 109 diff changeset	347 self.updateEnd()
92 c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	348 return self.use
c4726e19b8ec Finished first draft of TLearner Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 78 diff changeset	349
129 4c2280edcaf5 Fixed typos in learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 128 diff changeset	350 class OnlineGradientTLearner(MinibatchUpdatesTLearner):
111 88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	351 """
132 f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	352 Specialization of L{MinibatchUpdatesTLearner} in which the minibatch updates
111 88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	353 are obtained by performing an online (minibatch-based) gradient step.
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	354
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	355 Sub-classes must define the following:
132 f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	356 - self._learning_rate (may be changed by the sub-class between epochs or minibatches)
f6505ec32dc3 Updated documentation slightly Joseph Turian <turian@gmail.com> parents: 131 diff changeset	357 - self.lossAttribute() = name of the loss field
111 88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	358 """
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	359 def __init__(self,truly_online=False):
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	360 """
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	361 If truly_online then only one pass is made through the training set passed to update().
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	362
d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	363 SUBCLASSES SHOULD CALL THIS CONSTRUCTOR ONLY AFTER HAVING DEFINED ALL THEIR THEANO FORMULAS
111 88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	364 """
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	365 self.truly_online=truly_online
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	366
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	367 # create the formulas for the gradient update
129 4c2280edcaf5 Fixed typos in learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 128 diff changeset	368 old_params = [self.__getattribute__("_"+name) for name in self.parameterAttributes()]
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	369 new_params_names = ["_new_"+name for name in self.parameterAttributes()]
129 4c2280edcaf5 Fixed typos in learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 128 diff changeset	370 loss = self.__getattribute__("_"+self.lossAttribute())
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	371 self.setAttributes(new_params_names,
129 4c2280edcaf5 Fixed typos in learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 128 diff changeset	372 [t.add_inplace(param,self._learning_rate*t.grad(loss,param))
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	373 for param in old_params])
129 4c2280edcaf5 Fixed typos in learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 128 diff changeset	374 MinibatchUpdatesTLearner.__init__(self)
4c2280edcaf5 Fixed typos in learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 128 diff changeset	375
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	376
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	377 def namesOfAttributesToComputeOutputs(self,output_names):
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	378 """
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	379 The output_names are attribute names (not the corresponding Result names, which have leading _).
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	380 Return the corresponding input names
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	381 """
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	382 all_inputs = t.gof.graph.inputs(self.names2OpResults(output_names))
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	383 # remove constants and leading '_' in name
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	384
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	385 return [r.name for r in all_inputs if isinstance(r,theano.Result) and \
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	386 not isinstance(r,theano.Constant) and not isinstance(r,theano.Value)]
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	387 #inputs = []
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	388 #for r in all_inputs:
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	389 # if isinstance(r,theano.Result) and \
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	390 # not isinstance(r,theano.Constant) and not isinstance(r,theano.Value):
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	391 # inputs.append(r.name)
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	392 #return inputs
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	393
111 88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	394 def isLastEpoch(self):
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	395 return self.truly_online
88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	396
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	397 def updateMinibatchInputAttributes(self):
d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	398 return self.parameterAttributes()
d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	399
d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	400 def updateMinibatchOutputAttributes(self):
133 b4657441dd65 Corrected typos Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 132 diff changeset	401 return ["new_"+name for name in self.parameterAttributes()]
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	402
d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	403 def updateEndInputAttributes(self):
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	404 return self.namesOfAttributesToComputeOutputs(self.updateEndOutputAttributes())
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	405
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	406 def useInputAttributes(self):
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	407 return self.parameterAttributes()
111 88257dfedf8c Added another work in progress, for mlp's bengioy@bengiomac.local parents: 110 diff changeset	408
134 3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	409 def useOutputAttributes(self):
3f4e5c9bdc5e Fixes to ApplyFunctionDataSet and other things to make learner and mlp work Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 133 diff changeset	410 return []
118 d0a1bd0378c6 Finished draft of OneHiddenLayerNNetClassifier to debut learner.py Yoshua Bengio <bengioy@iro.umontreal.ca> parents: 111 diff changeset	411

Mercurial > pylearn

annotate learner.py @ 143:b7ca3545186b