Mercurial > pylearn

annotate learner.py @ 413:f63dfb0ac7dc

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Added cost functions
author Joseph Turian <turian@iro.umontreal.ca>
date Fri, 11 Jul 2008 15:26:41 -0400
parents fe57b96f33d4
children
rev   line source
211
bd728c83faff in __get__, problem if the i.stop was None, i being the slice, added one line replacing None by the len(self)
Thierry Bertin-Mahieux <bertinmt@iro.umontreal.ca>
parents: 209
diff changeset
1
1
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents: 0
diff changeset
2
172
fb4837eed1a6 fixed import of AbstractFunction
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 167
diff changeset
3 from exceptions import *
209
50a8302addaf template statscollector
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 193
diff changeset
4 from dataset import AttributesHolder
180
2698c0feeb54 mlp seems to work!
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 177
diff changeset
5
262
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
6 class OfflineLearningAlgorithm(object):
132
f6505ec32dc3 Updated documentation slightly
Joseph Turian <turian@gmail.com>
parents: 131
diff changeset
7 """
262
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
8 Base class for offline learning algorithms, provides an interface
1
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents: 0
diff changeset
9 that allows various algorithms to be applicable to generic learning
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
10 algorithms. It is only given here to define the expected semantics.
1
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents: 0
diff changeset
11
262
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
12 An offline learning algorithm can be seen as a function that when
135
0d8e721cc63c Fixed bugs in dataset to make test_mlp.py work
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 134
diff changeset
13 applied to training data returns a learned function (which is an object that
0d8e721cc63c Fixed bugs in dataset to make test_mlp.py work
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 134
diff changeset
14 can be applied to other data and return some output data).
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
15
262
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
16 The offline learning scenario is the standard and most common one
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
17 in machine learning: an offline learning algorithm is applied
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
18 to a training dataset,
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
19
326
fe57b96f33d4 made ExampleWiseMean Op
Olivier Breuleux <breuleuo@iro.umontreal.ca>
parents: 287
diff changeset
20 model = learning_algorithm(training_set)
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
21
262
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
22 resulting in a fully trained model that can be applied to another dataset
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
23 in order to perform some desired computation:
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
24
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
25 output_dataset = model(input_dataset)
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
26
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
27 Note that the application of a dataset has no side-effect on the model.
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
28 In that example, the training set may for example have 'input' and 'target'
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
29 fields while the input dataset may have only 'input' (or both 'input' and
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
30 'target') and the output dataset would contain some default output fields defined
262
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
31 by the learning algorithm (e.g. 'output' and 'error'). The user may specifiy
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
32 what the output dataset should contain either by setting options in the
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
33 model, by the presence of particular fields in the input dataset, or with
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
34 keyword options of the __call__ method of the model (see LearnedModel.__call__).
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
35
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
36 """
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
37
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
38 def __init__(self): pass
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
39
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
40 def __call__(self, training_dataset):
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
41 """
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
42 Return a fully trained TrainedModel.
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
43 """
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
44 raise AbstractFunction()
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
45
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
46 class TrainedModel(AttributesHolder):
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
47 """
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
48 TrainedModel is a base class for models returned by instances of an
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
49 OfflineLearningAlgorithm subclass. It is only given here to define the expected semantics.
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
50 """
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
51 def __init__(self):
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
52 pass
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
53
262
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
54 def __call__(self,input_dataset,output_fieldnames=None,
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
55 test_stats_collector=None,copy_inputs=False,
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
56 put_stats_in_output_dataset=True,
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
57 output_attributes=[]):
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
58 """
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
59 A L{TrainedModel} can be used with
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
60 with one or more calls to it. The main argument is an input L{DataSet} (possibly
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
61 containing a single example) and the result is an output L{DataSet} of the same length.
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
62 If output_fieldnames is specified, it may be use to indicate which fields should
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
63 be constructed in the output L{DataSet} (for example ['output','classification_error']).
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
64 Otherwise, some default output fields are produced (possibly depending on the input
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
65 fields available in the input_dataset).
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
66 Optionally, if copy_inputs, the input fields (of the input_dataset) can be made
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
67 visible in the output L{DataSet} returned by this method.
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
68 Optionally, attributes of the learner can be copied in the output dataset,
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
69 and statistics computed by the stats collector also put in the output dataset.
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
70 Note the distinction between fields (which are example-wise quantities, e.g. 'input')
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
71 and attributes (which are not, e.g. 'regularization_term').
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
72 """
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
73 raise AbstractFunction()
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
74
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
75
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
76 class OnlineLearningAlgorithm(object):
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
77 """
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
78 Base class for online learning algorithms, provides an interface
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
79 that allows various algorithms to be applicable to generic online learning
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
80 algorithms. It is only given here to define the expected semantics.
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
81
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
82 The basic setting is that the training data are only revealed in pieces
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
83 (maybe one example or a batch of example at a time):
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
84
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
85 model = learning_algorithm()
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
86
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
87 results in a fresh model. The model can be adapted by presenting
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
88 it with some training data,
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
89
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
90 model.update(some_training_data)
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
91 ...
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
92 model.update(some_more_training_data)
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
93 ...
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
94 model.update(yet_more_training_data)
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
95
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
96 and at any point one can use the model to perform some computation:
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
97
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
98 output_dataset = model(input_dataset)
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
99
262
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
100 The model should be a LearnerModel subclass instance, and LearnerModel
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
101 is a subclass of LearnedModel.
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
102
1
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents: 0
diff changeset
103 """
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
104
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
105 def __init__(self): pass
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
106
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
107 def __call__(self, training_dataset=None):
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
108 """
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
109 Return a LearnerModel, either fresh (if training_dataset is None) or fully trained (otherwise).
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
110 """
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
111 raise AbstractFunction()
1
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents: 0
diff changeset
112
287
78cc8fe3bbe9 bugfix, bad object name
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 262
diff changeset
113 class LearnerModel(TrainedModel):
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
114 """
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
115 LearnerModel is a base class for models returned by instances of a LearningAlgorithm subclass.
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
116 It is only given here to define the expected semantics.
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
117 """
10
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners.
bengioy@esprit.iro.umontreal.ca
parents: 1
diff changeset
118 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
119 pass
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners.
bengioy@esprit.iro.umontreal.ca
parents: 1
diff changeset
120
14
5ede27026e05 Working on gradient_based_learner
bengioy@bengiomac.local
parents: 13
diff changeset
121 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
122 """
262
14b9779622f9 Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 211
diff changeset
123 Continue training a learner model, with the evidence provided by the given training set.
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
124 Hence update can be called multiple times. This is the main method used for training in the
10
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners.
bengioy@esprit.iro.umontreal.ca
parents: 1
diff changeset
125 on-line setting or the sequential (Bayesian or not) settings.
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
126
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
127 This function has as side effect that self(data) will behave differently,
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
128 according to the adaptation achieved by update().
75
90e4c0784d6e Added draft of LinearRegression learner
bengioy@bengiomac.local
parents: 20
diff changeset
129
132
f6505ec32dc3 Updated documentation slightly
Joseph Turian <turian@gmail.com>
parents: 131
diff changeset
130 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
131 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
132 training.
10
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners.
bengioy@esprit.iro.umontreal.ca
parents: 1
diff changeset
133 """
193
cb6b945acf5a Complete redesign of learner...
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 180
diff changeset
134 raise AbstractFunction()
10
80bf5492e571 Rewrote learner.py according to the specs in the wiki for learners.
bengioy@esprit.iro.umontreal.ca
parents: 1
diff changeset
135