Loads the dataset :type dataset: string :param dataset: the path to the dataset (here MNIST)

Return a float representing the number of errors in the minibatch over the total number of examples of the minibatch :type y: theano.

Just compute the error f_pred: Theano fct computing the prediction prepare_data: usual prepare_data for that dataset.

Return the negative log-likelihood of the prediction of this model under a given target distribution. .. math:: \frac{1}

Initialize a matrix shared variable with normally distributed elements.

Transform an array with one flattened image per row, into an array in which images are reshaped and layed out like tiles on a floor. Thi

Calculating the clustering accuracy. The predicted result must have the same number of clusters as the ground truth. ypred: 1-D numpy ve

Function to compute the free energy

This function computes the cost and the updates for one trainng step of the dA

Used to shuffle the dataset at each iteration.

win :: int corresponding to the size of the window given a list of indexes composing a sentence l :: array containing the word indexes

Computes the values of the hidden layer

Create the matrices from the datasets. This pad each sequence to the same lenght: the lenght of the longuest sequence or maxlen. if maxl

This function infers state of hidden units given visible units

Function that loads the dataset into shared variables The reason we store our dataset in shared variables is to allow Theano to copy

Initialize a vector shared variable with zero elements.

When we pickle the model. Needed for the GPU stuff.

Construct a k-step Gibbs chain starting at v for an RBM. v : Theano vector or matrix If a matrix, multiple chains will be run in parallel

INPUT: p :: predictions g :: groundtruth w :: corresponding words OUTPUT: filename :: name of the file where the predictions

Returns a new position obtained after `n_steps` of HMC simulation. Parameters ---------- kwargs: dictionary

Return a float representing the number of errors in the minibatch over the total number of examples of the minibatch ; zero one loss o

Demonstrates lenet on MNIST dataset :type learning_rate: float :param learning_rate: learning rate used (factor for the stochastic

Generate a sample sequence, plot the resulting piano-roll and save it as a MIDI file. filename : string A MIDI file will

This function keeps ``1-corruption_level`` entries of the inputs the same and zero-out randomly selected subset of size ``coruption_level``

Computes the reconstructed input given the values of the hidden layer

Returns the Hamiltonian (sum of potential and kinetic energy) for the given velocity and position. Parameters ---------- pos: th

This function infers state of visible units given hidden units

:type Init: string :param Init: a string contains the filename of a saved network, the saved network can be loaded to initialize

Function to perform a KMeans update on a batch of data, center is the centroid from last iteration.

Generates a function `train` that implements one step of finetuning, a function `validate` that computes the error on a batch from the

Generates a function `train` that implements one step of finetuning, a function `validate` that computes the error on a batch from the

Generates a function `train` that implements one step of finetuning, a function `validate` that computes the error on a batch from the

Construct a symbolic RNN-RBM and initialize parameters. n_visible : integer Number of visible units. n_hidden : integer Numbe

Demonstrate conjugate gradient optimization of a log-linear model This is demonstrated on MNIST. :type n_epochs: int :param n_epochs: nu

download the corresponding atis file from http://www-etud.iro.umontreal.ca/~mesnilgr/atis/

This function computes the cost and the updates for one trainng step of the dA

This functions implements one step of CD-k or PCD-k :param lr: learning rate used to train the RBM :param persistent: None for CD. F

This function computes the cost and the updates for one trainng step of the cA

Look for it as if it was a full path, if not, try local file, if not try in the data directory. Download dataset if it is not present

Computes the values of the hidden layer

Computes the values of the hidden layer

Computes the jacobian of the hidden layer with respect to the input, reshapes are necessary for broadcasting the element-wise product

run conlleval.pl perl script to obtain precision/recall and F1 score

Stochastic approximation to the pseudo-likelihood

Computes the reconstructed input given the values of the hidden layer

Computes the reconstructed input given the values of the hidden layer

Approximation to the reconstruction error Note that this function requires the pre-sigmoid activation as input. To understand why th

This function performs one-step of Hybrid Monte-Carlo sampling. We start by sampling a random velocity from a univariate Gaussian distributio

This function is executed after `n_steps` of HMC sampling (`hmc_move` function). It creates the updates dictionary used by the `simulate` func

Global (not LSTM) parameter. For the embeding and the classifier.

Instantiate the network kwargs contains all the input arguments of test_SdC

Returns the kinetic energy associated with the given velocity and mass of 1. Parameters ---------- vel: theano matrix Symboli

Loads the dataset :type path: String :param path: The path to the dataset (here IMDB) :type n_words: int :param n_words: The number o

Load the dataset and save it as shared-variable to be used by Theano

Performs a Metropolis-Hastings accept-reject move. Parameters ---------- energy_prev: theano vector Symbolic theano tensor w

Return the mean of the negative log-likelihood of the prediction of this model under a given target distribution. .. math::

Generates a list of functions, each of them implementing one step in trainnig the dA corresponding to the layer with same index. The

Generates a list of functions, for performing one step of gradient descent at a given layer. The function will require as input the mi

Generates a list of functions, each of them implementing one step in trainnig the dA corresponding to the layer with same index. The

This function propagates the hidden units activation downwards to the visible units Note that we return also the pre_sigmoid_activati

This function propagates the visible units activation upwards to the hidden units Note that we return also the pre-sigmoid activation

Scales all values in the ndarray ndar to be between 0 and 1

Demonstrate stochastic gradient descent optimization of a log-linear model This is demonstrated on MNIST. :type learning_rate: floa

Function that loads the dataset into shared variables The reason we store our dataset in shared variables is to allow Theano to copy

lol :: list of list as input seed :: seed the shuffling shuffle inplace each list in the same order

Return final (position, velocity) obtained after an `n_steps` leapfrog updates, using Hamiltonian dynamics. Parameters ----------

Demonstrates how to train and test a Deep Belief Network. This is demonstrated on MNIST. :type finetune_lr: float :param finetune_l