class MLP(object): def __init__(self, rng, input, n_in, n_hidden, n_out,discriminant_threshold): self.hiddenLayer = HiddenLayer( rng=rng, input=input, n_in=n_in, n_out=n_hidden, activation=T.tanh ) self.logRegressionLayer = LogisticRegression( input=self.hiddenLayer.output, n_in=n_hidden, n_out=n_out, discriminant_threshold = discriminant_threshold ) self.L1 = ( abs(self.hiddenLayer.W).sum() + abs(self.logRegressionLayer.W).sum() ) self.L2_sqr = ( (self.hiddenLayer.W ** 2).sum() + (self.logRegressionLayer.W ** 2).sum() ) self.negative_log_likelihood = ( self.logRegressionLayer.negative_log_likelihood ) self.errors = self.logRegressionLayer.errors self.params = self.hiddenLayer.params + self.logRegressionLayer.params self.layers = [self.hiddenLayer, self.logRegressionLayer] self.ypred = self.logRegressionLayer.y_pred self.py_given_x = self.logRegressionLayer.p_y_given_x def output(self, x): print "x" #print x.eval() y = self.hiddenLayer.outputer(x) print "y" #print y.eval() z = self.logRegressionLayer.output(y) print "z" #print z return z def asimov_errors(self, y): # check if y has same dimension of y_pred if y.ndim != self.logRegressionLayer.y_pred.ndim: raise TypeError( 'y should have the same shape as self.y_pred', ('y', y.type, 'y_pred', self.y_pred.type) ) # check if y is of the correct datatype if y.dtype.startswith('int'): S = T.sum(T.eq(y,1)) B = T.sum(T.eq(y,0))#*10000 # TODO: cross-section scaling s = T.sum(T.and_(T.eq(y,1),T.eq(self.logRegressionLayer.y_pred,1))) b = T.sum(T.and_(T.eq(y,0),T.eq(self.logRegressionLayer.y_pred,1)))#*10000 TODO: cross-section scaling return(S,B,s,b) # represents a mistake in prediction else: raise NotImplementedError() def prediction(self,y): p_y_and_y = self.logRegressionLayer.p_y_given_x[:,1], y return p_y_and_y
class SdA(object): """Stacked denoising auto-encoder class (SdA) A stacked denoising autoencoder model is obtained by stacking several dAs. The hidden layer of the dA at layer `i` becomes the input of the dA at layer `i+1`. The first layer dA gets as input the input of the SdA, and the hidden layer of the last dA represents the output. Note that after pretraining, the SdA is dealt with as a normal MLP, the dAs are only used to initialize the weights. """ def __init__( self, numpy_rng, theano_rng=None, n_ins=5, hidden_layers_sizes=[28, 28],#original [500,500] n_outs=2, corruption_levels=[0.1, 0.1], discriminant_threshold = 0.5 ): self.sigmoid_layers = [] self.dA_layers = [] self.params = [] self.n_layers = len(hidden_layers_sizes) assert self.n_layers > 0 if not theano_rng: theano_rng = RandomStreams(numpy_rng.randint(2 ** 30)) # allocate symbolic variables for the data self.x = T.matrix('x') # the data is presented as rasterized images self.y = T.ivector('y') # the labels are presented as 1D vector of # [int] labels # end-snippet-1 # The SdA is an MLP, for which all weights of intermediate layers # are shared with a different denoising autoencoders # We will first construct the SdA as a deep multilayer perceptron, # and when constructing each sigmoidal layer we also construct a # denoising autoencoder that shares weights with that layer # During pretraining we will train these autoencoders (which will # lead to chainging the weights of the MLP as well) # During finetunining we will finish training the SdA by doing # stochastich gradient descent on the MLP # start-snippet-2 for i in xrange(self.n_layers): # construct the sigmoidal layer # the size of the input is either the number of hidden units of # the layer below or the input size if we are on the first layer if i == 0: input_size = n_ins else: input_size = hidden_layers_sizes[i - 1] # the input to this layer is either the activation of the hidden # layer below or the input of the SdA if you are on the first # layer if i == 0: layer_input = self.x else: layer_input = self.sigmoid_layers[-1].output sigmoid_layer = HiddenLayer(rng=numpy_rng, input=layer_input, n_in=input_size, n_out=hidden_layers_sizes[i], activation=T.nnet.sigmoid) # add the layer to our list of layers self.sigmoid_layers.append(sigmoid_layer) self.params.extend(sigmoid_layer.params) # Construct a denoising autoencoder that shared weights with this # layer dA_layer = dA(numpy_rng=numpy_rng, theano_rng=theano_rng, input=layer_input, n_visible=input_size, n_hidden=hidden_layers_sizes[i], W=sigmoid_layer.W, bhid=sigmoid_layer.b) self.dA_layers.append(dA_layer) # end-snippet-2 # We now need to add a logistic layer on top of the MLP self.logLayer = LogisticRegression( input=self.sigmoid_layers[-1].output, n_in=hidden_layers_sizes[-1], n_out=n_outs, discriminant_threshold = discriminant_threshold ) self.params.extend(self.logLayer.params) # construct a function that implements one step of finetunining # compute the cost for second phase of training, # defined as the negative log likelihood self.finetune_cost = self.logLayer.negative_log_likelihood(self.y) self.errors = self.logLayer.errors(self.y) self.py_given_x = self.logLayer.p_y_given_x def output(self, x): x = self.sigmoid_layers[0].outputer(x) x = self.sigmoid_layers[1].outputer(x) x = self.logLayer.output(x) return x def asimov_errors(self, y): # check if y has same dimension of y_pred if y.ndim != self.logLayer.y_pred.ndim: raise TypeError( 'y should have the same shape as self.y_pred', ('y', y.type, 'y_pred', self.y_pred.type) ) # check if y is of the correct datatype if y.dtype.startswith('int'): S = T.sum(T.eq(y,1)) B = T.sum(T.eq(y,0))#*10000 # TODO: cross-section scaling s = T.sum(T.and_(T.eq(y,1),T.eq(self.logLayer.y_pred,1))) b = T.sum(T.and_(T.eq(y,0),T.eq(self.logLayer.y_pred,1)))#*10000 TODO: cross-section scaling return(S,B,s,b) # represents a mistake in prediction else: raise NotImplementedError() def prediction(self,y): p_y_and_y = self.logLayer.p_y_given_x[:,1], y return p_y_and_y def pretraining_functions(self, train_set_x, batch_size): ''' Generates a list of functions, each of them implementing one step in trainnig the dA corresponding to the layer with same index. The function will require as input the minibatch index, and to train a dA you just need to iterate, calling the corresponding function on all minibatch indexes. :type train_set_x: theano.tensor.TensorType :param train_set_x: Shared variable that contains all datapoints used for training the dA :type batch_size: int :param batch_size: size of a [mini]batch :type learning_rate: float :param learning_rate: learning rate used during training for any of the dA layers ''' # index to a [mini]batch index = T.lscalar('index') # index to a minibatch corruption_level = T.scalar('corruption') # % of corruption to use learning_rate = T.scalar('lr') # learning rate to use # begining of a batch, given `index` batch_begin = index * batch_size # ending of a batch given `index` batch_end = batch_begin + batch_size pretrain_fns = [] for dA in self.dA_layers: # get the cost and the updates list cost, updates = dA.get_cost_updates(corruption_level, learning_rate) # compile the theano function fn = theano.function( inputs=[ index, theano.Param(corruption_level, default=0.2), theano.Param(learning_rate, default=0.1) ], outputs=cost, updates=updates, givens={ self.x: train_set_x[batch_begin: batch_end] } ) # append `fn` to the list of functions pretrain_fns.append(fn) return pretrain_fns def build_finetune_functions(self, datasets, batch_size, learning_rate): '''Generates a function `train` that implements one step of finetuning, a function `validate` that computes the error on a batch from the validation set, and a function `test` that computes the error on a batch from the testing set :type datasets: list of pairs of theano.tensor.TensorType :param datasets: It is a list that contain all the datasets; the has to contain three pairs, `train`, `valid`, `test` in this order, where each pair is formed of two Theano variables, one for the datapoints, the other for the labels :type batch_size: int :param batch_size: size of a minibatch :type learning_rate: float :param learning_rate: learning rate used during finetune stage ''' (train_set_x, train_set_y, train_set_w) = datasets[0] (valid_set_x, valid_set_y, valid_set_w) = datasets[1] (test_set_x, test_set_y, test_set_w) = datasets[2] # compute number of minibatches for training, validation and testing n_valid_batches = valid_set_x.get_value(borrow=True).shape[0] n_valid_batches /= batch_size n_test_batches = test_set_x.get_value(borrow=True).shape[0] n_test_batches /= batch_size index = T.lscalar('index') # index to a [mini]batch # compute the gradients with respect to the model parameters gparams = T.grad(self.finetune_cost, self.params) # compute list of fine-tuning updates updates = [ (param, param - gparam * learning_rate) for param, gparam in zip(self.params, gparams) ] train_fn = theano.function( inputs=[index], outputs=self.finetune_cost, updates=updates, givens={ self.x: train_set_x[ index * batch_size: (index + 1) * batch_size ], self.y: train_set_y[ index * batch_size: (index + 1) * batch_size ] }, name='train' ) test_score_i = theano.function( [index], self.errors, givens={ self.x: test_set_x[ index * batch_size: (index + 1) * batch_size ], self.y: test_set_y[ index * batch_size: (index + 1) * batch_size ] }, name='test' ) test_asimov_model = theano.function( inputs=[index], outputs=self.asimov_errors(self.y), givens={ self.x: test_set_x[index * batch_size: (index + 1) * batch_size], self.y: test_set_y[index * batch_size: (index + 1) * batch_size] #w: test_set_w[index * batch_size: (index + 1) * batch_size] #JP } ) get_prediction_model = theano.function( inputs=[index], outputs=self.prediction(self.y), givens={ self.x: test_set_x[index * batch_size: (index + 1) * batch_size], self.y: test_set_y[index * batch_size: (index + 1) * batch_size] #w: test_set_w[index * batch_size: (index + 1) * batch_size] #JP } ) valid_score_i = theano.function( [index], self.errors, givens={ self.x: valid_set_x[ index * batch_size: (index + 1) * batch_size ], self.y: valid_set_y[ index * batch_size: (index + 1) * batch_size ] }, name='valid' ) # Create a function that scans the entire validation set def valid_score(): return [valid_score_i(i) for i in xrange(n_valid_batches)] # Create a function that scans the entire test set def test_score(): return [test_score_i(i) for i in xrange(n_test_batches)] def asimov_score(): return np.sum([test_asimov_model(i) for i in xrange(n_test_batches)], axis = 0) def prediction_score(): return np.vstack([np.reshape(np.ravel(get_prediction_model(i), order='F'),(-1,2)) for i in xrange(n_test_batches)]) return train_fn, valid_score, test_score, asimov_score, prediction_score