def build_finetune_functions(self, sValidFeatMask, sTestFeatMask, 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) = datasets[0] #(valid_set_x, valid_set_y) = datasets[1] #(test_set_x, test_set_y) = datasets[2] def shared_dataset(data_xy, borrow=True): data_x, data_y = data_xy shared_x = theano.shared(numpy.asarray(data_x, dtype=theano.config.floatX), borrow=borrow) shared_y = theano.shared( numpy.asarray(data_y, dtype='int32'), #theano.config.floatX), borrow=borrow) return shared_x, shared_y #T.cast(shared_y, 'int32') with open(sValidFeatMask, 'r') as fFeatMask: lLines = fFeatMask.read().splitlines() lValidFeatMaskFiles = [tuple(line.split()) for line in lLines] with open(sTestFeatMask, 'r') as fFeatMask: lLines = fFeatMask.read().splitlines() lTestFeatMaskFiles = [tuple(line.split()) for line in lLines] print('Loading Validation Data.....') valid_set_x, valid_set_y = shared_dataset( GetXy(lValidFeatMaskFiles, iAttributeIndx, iNumFrames, iFeatSize, True)) print('Done') print('Loading Testing Data.....') test_set_x, test_set_y = shared_dataset( GetXy(lTestFeatMaskFiles, iAttributeIndx, iNumFrames, iFeatSize, True)) print('Done') # compute number of minibatches for training, validation and testing n_valid_samples = valid_set_x.get_value(borrow=True).shape[0] n_valid_batches = n_valid_samples / batch_size n_test_samples = test_set_x.get_value(borrow=True).shape[0] n_test_batches = (n_test_samples / batch_size) + (n_test_samples % batch_size > 0) print(( 'Number of Validation Samples = %i , Number of Validation batches = %i , Number of Test Samples = %i , Number of Test batches = %i' ) % (n_valid_samples, n_valid_batches, n_test_samples, n_test_batches)) index = T.lscalar('index') # index to a [mini]batch Alr = T.scalar('Alr') # learning rate to use X = T.matrix('x') # the data is presented as rasterized images y = T.ivector('y') # compute the gradients with respect to the model parameters gparams = T.grad(self.finetune_cost, self.params) # compute list of fine-tuning updates updates = [] for param, gparam in zip(self.params, gparams): updates.append((param, param - gparam * Alr)) train_fn = theano.function(inputs=[X, y, Alr], outputs=self.finetune_cost, updates=updates, givens={ self.x: X, self.y: y }) test_score_i = theano.function( [index], [self.errors, self.y_pred], 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] }) 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] }) #yPred = theano.function([self.x],self.y_pred) # 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)[0] for i in xrange(n_test_batches)] def yPred(): yP = numpy.asarray([]) for i in xrange(n_test_batches): yP = numpy.concatenate((yP, test_score_i(i)[1])) return yP def GetRefY(): return test_set_y.get_value(borrow=True) return train_fn, valid_score, test_score, yPred, GetRefY
def test_DBN(finetune_lr=lr, pretraining_epochs=100, pretrain_lr=0.0025, k=2, training_epochs=NOfepoch, sValidList=sValidList, sTestingList=sTestingList, batch_size=batchSize): """ Demonstrates how to train and test a Deep Belief Network. This is demonstrated on MNIST. :type learning_rate: float :param learning_rate: learning rate used in the finetune stage :type pretraining_epochs: int :param pretraining_epochs: number of epoch to do pretraining :type pretrain_lr: float :param pretrain_lr: learning rate to be used during pre-training :type k: int :param k: number of Gibbs steps in CD/PCD :type training_epochs: int :param training_epochs: maximal number of iterations ot run the optimizer :type dataset: string :param dataset: path the the pickled dataset :type batch_size: int :param batch_size: the size of a minibatch """ # compute number of minibatches for training, validation and testing with open(sTrainingList, 'r') as fFeatMask: lLines = fFeatMask.read().splitlines() lFeatMaskFiles = [tuple(line.split()) for line in lLines] n_train_chuncks = int( math.ceil(len(lFeatMaskFiles) / float(iNumFiles) )) #train_set_x.get_value(borrow=True).shape[0] / batch_size # numpy random generator numpy_rng = numpy.random.RandomState(123) print('... building the model') # construct the Deep Belief Network dbn = DBN(numpy_rng=numpy_rng, n_ins=n_inp, hidden_layers_sizes=[n_hus for i in range(n_hls)], n_outs=n_out) ######################### # PRETRAINING THE MODEL # ######################### #print ('... getting the pretraining functions') #pretraining_fns = dbn.pretraining_functions(train_set_x=train_set_x, # batch_size=batch_size, # k=k) #print ('... pre-training the model') #start_time = time.clock() ## Pre-train layer-wise #for i in xrange(dbn.n_layers): # # go through pretraining epochs # for epoch in xrange(pretraining_epochs): # # go through the training set # c = [] # for batch_index in xrange(n_train_batches): # c.append(pretraining_fns[i](index=batch_index, # lr=pretrain_lr)) # print ('Pre-training layer %i, epoch %d, cost ' % (i, epoch)) # print (numpy.mean(c)) # # end_time = time.clock() # print (('The pretraining code for file ' + # os.path.split(__file__)[1] + # ' ran for %.2fm' % ((end_time - start_time) / 60.))) ######################## # FINETUNING THE MODEL # ######################## # get the training, validation and testing function for the model print('... getting the finetuning functions') train_fn, validate_model, test_model, gety_pred, get_y = dbn.build_finetune_functions( sValidFeatMask=sValidList, sTestFeatMask=sTestingList, batch_size=batch_size, learning_rate=finetune_lr) print('... finetunning the model') # early-stopping parameters #patience = 4 * n_train_batches # 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 print('Number of Training Chunks = ', n_train_chuncks) y = get_y() best_params = [] ybest = numpy.asarray([]) best_validation_loss = numpy.inf prev_validation_loss = 200 test_score = 0. start_time = time.clock() Alrc = 0.1 AlrE = 0.00001 done_looping = False epoch = 0 epochC = 0 iter = 0 for param in dbn.params: best_params.append(param.get_value()) while (epoch < training_epochs) and (not done_looping): epoch = epoch + 1 epochC = epochC + 1 #for minibatch_index in xrange(n_train_batches): for chunck_index in xrange(n_train_chuncks): #print n_train_batches, epoch, minibatch_index #minibatch_avg_cost = train_fn(minibatch_index,Alrc) lCurFeatMask = lFeatMaskFiles[chunck_index * iNumFiles:(chunck_index + 1) * iNumFiles] arCur_X, vCur_y = GetXy(lCurFeatMask, iAttributeIndx, iNumFrames, iFeatSize, True) print(('Chunk Index = %i , Number of Samples = %i') % (chunck_index, arCur_X.shape[0])) #minibatch_avg_cost = train_fn(train_set_x_NS[minibatch_index*batch_size:(minibatch_index+1)*batch_size],train_set_y_NS[minibatch_index*batch_size:(minibatch_index+1)*batch_size],Alrc) n_curTrain_batches = int( math.ceil(arCur_X.shape[0] / float(batch_size))) for minibatch_index in xrange(n_curTrain_batches): minibatch_avg_cost = train_fn( arCur_X[minibatch_index * batch_size:(minibatch_index + 1) * batch_size], vCur_y[minibatch_index * batch_size:(minibatch_index + 1) * batch_size], Alrc) #iter = (epoch - 1) * n_train_chuncks + chunck_index * minibatch_index validation_losses = validate_model() this_validation_loss = numpy.mean(validation_losses) lossratio = (this_validation_loss - prev_validation_loss) / (prev_validation_loss + 1) print(lossratio) print('epoch %i, validation error %f, lr %f %%' % \ (epoch,this_validation_loss * 100., Alrc)) # if we got the best validation score until now #if this_validation_loss < best_validation_loss: if lossratio <= 0.0: #print '*******************1**************' #print dbn.params[0].get_value() for i in range(len(dbn.params)): best_params[i] = dbn.params[i].get_value() # save best validation score and iteration number best_validation_loss = this_validation_loss prev_validation_loss = this_validation_loss best_iter = iter # test it on the test set test_losses = test_model() test_score = numpy.mean(test_losses) yP = gety_pred() acc = accuracy_score(y, yP) conf_matrix = confusion_matrix(y, yP) conf_matrix_norm = conf_matrix / numpy.sum( conf_matrix, axis=1, dtype='float', keepdims=True) acc_str = ' '.join([ str(conf_matrix_norm[i, i]) for i in range(conf_matrix_norm.shape[0]) ]) conf_str = ' '.join(conf_matrix.flatten().astype('str')) ybest = yP print(( ' epoch %i, test error of best model %f, acc = %s , conf = %s %%' ) % (epoch, test_score * 100., acc_str, conf_str)) else: if Alrc <= AlrE: done_looping = True break elif epochC > 20: Alrc = Alrc / 2 for param, best_param in zip(dbn.params, best_params): param.set_value(best_param) epochC = 0 end_time = time.clock() print(('Optimization complete with best validation score of %f %%,' 'with test performance %f %%') % (best_validation_loss * 100., test_score * 100.)) if not os.path.isdir('Temp'): os.makedirs('Temp') fbestParam = gzip.open( 'Temp/Bestparam_' + str(iAttributeIndx) + '_' + str(iNumFrames) + '_' + str(n_hls) + '_' + str(n_hus) + '_' + str(n_out) + '_' + str(iFeatSize) + '.pkl.gz', 'wb') fbestP = open( 'Temp/Bestp_' + str(iAttributeIndx) + '_' + str(iNumFrames) + '_' + str(n_hls) + '_' + str(n_hus) + '_' + str(n_out) + '_' + str(iFeatSize) + '.txt', 'w') numpy.savetxt(fbestP, ybest) fbestP.close() cPickle.dump(best_params, fbestParam) fbestParam.close() OF = open(outFile, 'a') print(iAttributeIndx, iNumFrames, lr, n_hls, n_hus, n_out, batchSize, NOfepoch, n_inp, outFile, test_score * 100., acc_str, conf_str, file=OF) OF.close()
import sys from GetDataFromList import GetXy with open(sys.argv[1]) as fIn: lLines = fIn.read().splitlines() lFeatMask = [tuple(sLine.split()) for sLine in lLines] a,b,c = GetXy(lFeatMask[:300],int(sys.argv[2]),9,78,True) print(c)