def mlp(learning_rate=0.01, L1_reg=0.00, L2_reg=0.0001, n_epochs=1000, batch_size=20, n_hidden=500): """ Stochastic gradient descent optimization for a multilayer perceptron :type learning_rate: float :param learning_rate: learning rate used (factor for stochatic gradient) :type L1_reg: float :param L1_reg: L1-norm's weight when added to the cost :type L2_reg: float :param L2_reg: L2-norm's weight when added to the cost :type n_epochs: int :param n_epochs: maximal number of epichs to run the optimizer :param batch_size: size of bach processed :param n_hidden: number of hidden outputs """ datasets = load_data(5000) train_set_x, train_set_y = datasets[0] valid_set_x, valid_set_y = datasets[1] test_set_x, test_set_y = datasets[2] # compute number of minibatches for training, validation and testing n_train_batches = train_set_x.get_value(borrow=True).shape[0] // batch_size n_valid_batches = valid_set_x.get_value(borrow=True).shape[0] // batch_size n_test_batches = test_set_x.get_value(borrow=True).shape[0] // batch_size ###################### # BUILD ACTUAL MODEL # ###################### print('... building the model ...') # allocate symbolic variables for the data index = T.lscalar() # index to a [mini]batch x = T.matrix('x') # the data is presented as rasterized images y = T.ivector('y') # labels as 1D vector of [int] labels rng = np.random.RandomState(13032016) # construct the MLP class classifier = MLP( rng=rng, input=x, n_in=prep.SIZE[0] * prep.SIZE[1] * 3, n_hidden=n_hidden, n_out=3 ) # start-snippet-4 # the cost we minimize during training is the negative log liklihood of # the model plus the regularization terms (L1 and L2) cost = ( classifier.negative_log_likelihood(y) + L1_reg * classifier.L1 + L2_reg * classifier.L2_sqr ) # end-snippet-4 # compiling a Theano function that computes the mistakes that are made # by the model on a minibatch test_model = theano.function( inputs=[index], outputs=classifier.errors(y), givens={ x: test_set_x[index * batch_size:(index + 1) * batch_size], y: test_set_y[index * batch_size:(index + 1) * batch_size] } ) validate_model = theano.function( inputs=[index], outputs=classifier.errors(y), givens={ x: valid_set_x[index * batch_size:(index + 1) * batch_size], y: valid_set_y[index * batch_size:(index + 1) * batch_size] } ) # start-snippet-5 # compute the gradient of cost with respect to theta (stored in params) # the resulting gradients will be stored in a list gparams gparams = [T.grad(cost, param) for param in classifier.params] # specify how to update the parameters of the model as a list of # (variable, update expression_ pairs updates = [ (param, param - learning_rate * gparam) for param, gparam in zip(classifier.params, gparams) ] # compiling a Theano function 'train model' that returns the cost, but # at the same time updates the parameter of the model based on the rules # defined in 'updates' train_model = theano.function( inputs=[index], outputs=cost, updates=updates, givens={ x: train_set_x[index * batch_size: (index + 1) * batch_size], y: train_set_y[index * batch_size: (index + 1) * batch_size] } ) # end-snippet-5 ############### # TRAIN MODEL # ############### print('... training ...') # early-stopping parameters patience = 10000 patience_increase = 2 # wait this long when a new best is found improvement_threshold = 0.995 # relative improvement considered significant validation_frequency = min(n_train_batches, patience // 2) # go thorugh this many minibatches before # checking the network on the validation set; # in this case we check every epoch best_validation_loss = np.inf best_iter = 0 test_score = 0. start_time = timeit.default_timer() epoch = 0 done_looping = False while (epoch < n_epochs) and (not done_looping): epoch += 1 for minibatch_index in range(n_train_batches): minibatch_avg_cost = train_model(minibatch_index) # iteration number iter = (epoch - 1) * n_train_batches + minibatch_index if (iter + 1) % validation_frequency == 0: # computer zero-one loss on validation set validation_losses = [validate_model(i) for i in range(n_valid_batches)] this_validation_loss = np.mean(validation_losses) print( 'epoch %i, minibatch %i/%i, validation error %f %%' % ( epoch, minibatch_index + 1, n_train_batches, this_validation_loss * 100. ) ) # if we get the best validation score until now if this_validation_loss < best_validation_loss: # improve patience if loss improvement is good enough if ( this_validation_loss < best_validation_loss * improvement_threshold ): patience = max(patience, iter * patience_increase) best_validation_loss = this_validation_loss best_iter = iter # test it on on the test set test_losses = [test_model(i) for i in range(n_test_batches)] test_score = np.mean(test_losses) print((' epoch %i, minibatch %i/%i, test error of ' 'best model %f %%') % (epoch, minibatch_index + 1, n_train_batches, test_score * 100.)) # save the best model with open('best_mlp_model.pkl', 'wb') as f: pickle.dump(classifier, f) if patience <= iter: done_looping = True break end_time = timeit.default_timer() print(('Optimization complete. Best validation score of %f %% ' 'obtained at iteration %i, with test performance %f %%') % (best_validation_loss * 100., best_iter + 1, test_score * 100.)) print(('The code for file ' + os.path.split(__file__)[1] + ' ran for %.2fm' % ((end_time - start_time) / 60.)), file=sys.stderr)
def evaluate_lenet5(learning_rate=0.1, n_epochs=200, nkerns=[20, 50], batch_size=500): """ :type learning_rate: float :param learning_rate: factor for stochastic gradient :type n_epochs: int :param n_epochs: maximal number of epochs to run the optimizer :type nkerns: list of ints :param nkerns: number of kernels on each layer :param batch_size: size of batch """ rng = np.random.RandomState(20160313) datasets = load_data(3500) train_set_x, train_set_y = datasets[0] valid_set_x, valid_set_y = datasets[1] test_set_x, test_set_y = datasets[2] # compute number of minibatches for training, validation and testing n_train_batches = train_set_x.get_value(borrow=True).shape[0] n_valid_batches = valid_set_x.get_value(borrow=True).shape[0] n_test_batches = test_set_x.get_value(borrow=True).shape[0] n_train_batches //= batch_size n_valid_batches //= batch_size n_test_batches //= batch_size # allocate symbolic variables for the data index = T.lscalar() # index to a [mini]batch # start-snippet-1 x = T.matrix('x') # the data is presented as rasterized images y = T.ivector('y') # the labels are presented as 1D vector of # [int] labels ###################### # BUILD ACTUAL MODEL # ###################### print('... building the model') # Reshape matrix of rasterized images of shape (batch_size, image w*h) # to a 4D tensor, compatible with our LeNetConvPoolLayer layer0_input = x.reshape((batch_size, 1, prep.SIZE[0], prep.SIZE[1])) # Construct the first convolutional pooling layer: # filtering reduces the image size to (size-5+1 , size-5+1) = (24, 24) # maxpooling reduces this by factor of 1/2 # 4D output tensor is thus of shape (batch_size, nkerns[0], 12, 12) layer0 = LeNetConvPoolLayer( rng, input=layer0_input, image_shape=(batch_size, 1, prep.SIZE[0], prep.SIZE[1]), filter_shape=(nkerns[0], 1, 5, 5), poolsize=(2, 2) ) new_size = ((prep.SIZE[0] - 5 + 1) // 2, (prep.SIZE[1] - 5 + 1 // 2)) # Construct the second convolutional pooling layer # filtering reduces the image size as previous layer1 = LeNetConvPoolLayer( rng, input=layer0.output, image_shape=(batch_size, nkerns[0], new_size[0], new_size[1]), filter_shape=(nkerns[1], nkerns[0], 5, 5), poolsize=(2, 2) ) new_size = ((new_size[0] -5 + 1) // 2, (new_size[1] -5 + 1 // 2)) # the HiddenLayer being fully-connected, it operates on 2D matrices of # shape (batch_size, num_pixels) (i.e matrix of rasterized images). # This will generate a matrix of shape (batch_size, nkerns[1] * new_size), # or (500, 50 * new_size[0] * new_size[1]) = (500, 800) # with the default values. layer2_input = layer1.output.flatten(2) # construct a fully-connected sigmoidal layer layer2 = HiddenLayer( rng, input=layer2_input, n_in=nkerns[1] * new_size[0] * new_size[1], n_out=500, activation=T.tanh ) # classify the values of the fully-connected sigmoidal layer layer3 = LogisticRegression(input=layer2.output, n_in=500, n_out=3) # the cost we minimize during training is the NLL of the model cost = layer3.negative_log_likelihood(y) # create a function to compute the mistakes that are made by the model test_model = theano.function( [index], layer3.errors(y), givens={ x: test_set_x[index * batch_size: (index + 1) * batch_size], y: test_set_y[index * batch_size: (index + 1) * batch_size] } ) validate_model = theano.function( [index], layer3.errors(y), givens={ x: valid_set_x[index * batch_size: (index + 1) * batch_size], y: valid_set_y[index * batch_size: (index + 1) * batch_size] } ) # create a list of all model parameters to be fit by gradient descent params = layer3.params + layer2.params + layer1.params + layer0.params # create a list of gradients for all model parameters grads = T.grad(cost, params) # train_model is a function that updates the model parameters by # SGD Since this model has many parameters, it would be tedious to # manually create an update rule for each model parameter. We thus # create the updates list by automatically looping over all # (params[i], grads[i]) pairs. updates = [ (param_i, param_i - learning_rate * grad_i) for param_i, grad_i in zip(params, grads) ] train_model = theano.function( [index], cost, updates=updates, givens={ x: train_set_x[index * batch_size: (index + 1) * batch_size], y: train_set_y[index * batch_size: (index + 1) * batch_size] } ) # end-snippet-1 ############### # TRAIN MODEL # ############### print('... training') # early-stopping parameters patience = 10000 # look as this many examples regardless patience_increase = 2 # wait this much longer when a new best is # found improvement_threshold = 0.995 # a relative improvement of this much is # considered significant validation_frequency = min(n_train_batches, patience // 2) # go through this many # minibatche before checking the network # on the validation set; in this case we # check every epoch best_validation_loss = np.inf best_iter = 0 test_score = 0. start_time = timeit.default_timer() epoch = 0 done_looping = False while (epoch < n_epochs) and (not done_looping): epoch = epoch + 1 for minibatch_index in range(n_train_batches): iter = (epoch - 1) * n_train_batches + minibatch_index if iter % 100 == 0: print('training @ iter = ', iter) cost_ij = train_model(minibatch_index) if (iter + 1) % validation_frequency == 0: # compute zero-one loss on validation set validation_losses = [validate_model(i) for i in range(n_valid_batches)] this_validation_loss = np.mean(validation_losses) print('epoch %i, minibatch %i/%i, validation error %f %%' % (epoch, minibatch_index + 1, n_train_batches, this_validation_loss * 100.)) # if we got the best validation score until now if this_validation_loss < best_validation_loss: #improve patience if loss improvement is good enough if this_validation_loss < best_validation_loss * \ improvement_threshold: patience = max(patience, iter * patience_increase) # save best validation score and iteration number best_validation_loss = this_validation_loss best_iter = iter # test it on the test set test_losses = [ test_model(i) for i in range(n_test_batches) ] test_score = np.mean(test_losses) print((' epoch %i, minibatch %i/%i, test error of ' 'best model %f %%') % (epoch, minibatch_index + 1, n_train_batches, test_score * 100.)) # save the best model # with open('best_mlp_model.pkl', 'wb') as f: # pickle.dump(classifier, f) if patience <= iter: done_looping = True break end_time = timeit.default_timer() print('Optimization complete.') print('Best validation score of %f %% obtained at iteration %i, ' 'with test performance %f %%' % (best_validation_loss * 100., best_iter + 1, test_score * 100.)) print(('The code for file ' + os.path.split(__file__)[1] + ' ran for %.2fm' % ((end_time - start_time) / 60.)), file=sys.stderr)