SYMBOL INDEX (579 symbols across 46 files)

FILE: emolga/basic/activations.py
  function softmax (line 4) | def softmax(x):
  function vector_softmax (line 8) | def vector_softmax(x):
  function time_distributed_softmax (line 12) | def time_distributed_softmax(x):
  function softplus (line 18) | def softplus(x):
  function relu (line 22) | def relu(x):
  function tanh (line 26) | def tanh(x):
  function sigmoid (line 30) | def sigmoid(x):
  function hard_sigmoid (line 34) | def hard_sigmoid(x):
  function linear (line 38) | def linear(x):
  function maxout2 (line 45) | def maxout2(x):
  function get (line 68) | def get(identifier):

FILE: emolga/basic/initializations.py
  function get_fans (line 8) | def get_fans(shape):
  function uniform (line 16) | def uniform(shape, scale=0.1):
  function normal (line 20) | def normal(shape, scale=0.05):
  function lecun_uniform (line 24) | def lecun_uniform(shape):
  function glorot_normal (line 33) | def glorot_normal(shape):
  function glorot_uniform (line 41) | def glorot_uniform(shape):
  function he_normal (line 47) | def he_normal(shape):
  function he_uniform (line 55) | def he_uniform(shape):
  function orthogonal (line 61) | def orthogonal(shape, scale=1.1):
  function identity (line 73) | def identity(shape, scale=1):
  function zero (line 80) | def zero(shape):
  function one (line 84) | def one(shape):
  function get (line 88) | def get(identifier):

FILE: emolga/basic/objectives.py
  function mean_squared_error (line 13) | def mean_squared_error(y_true, y_pred):
  function mean_absolute_error (line 17) | def mean_absolute_error(y_true, y_pred):
  function mean_absolute_percentage_error (line 21) | def mean_absolute_percentage_error(y_true, y_pred):
  function mean_squared_logarithmic_error (line 25) | def mean_squared_logarithmic_error(y_true, y_pred):
  function squared_hinge (line 29) | def squared_hinge(y_true, y_pred):
  function hinge (line 33) | def hinge(y_true, y_pred):
  function categorical_crossentropy (line 37) | def categorical_crossentropy(y_true, y_pred):
  function binary_crossentropy (line 47) | def binary_crossentropy(y_true, y_pred):
  function poisson_loss (line 53) | def poisson_loss(y_true, y_pred):
  function gaussian_kl_divergence (line 59) | def gaussian_kl_divergence(mean, ln_var):
  function get (line 97) | def get(identifier):

FILE: emolga/basic/optimizers.py
  function clip_norm (line 17) | def clip_norm(g, c, n):
  function kl_divergence (line 23) | def kl_divergence(p, p_hat):
  class Optimizer (line 27) | class Optimizer(object):
    method __init__ (line 28) | def __init__(self, **kwargs):
    method add (line 33) | def add(self, v):
    method get_state (line 36) | def get_state(self):
    method set_state (line 39) | def set_state(self, value_list):
    method get_updates (line 44) | def get_updates(self, params, loss):
    method get_gradients (line 47) | def get_gradients(self, loss, params):
    method get_config (line 63) | def get_config(self):
  class SGD (line 67) | class SGD(Optimizer):
    method __init__ (line 69) | def __init__(self, lr=0.05, momentum=0.9, decay=0.01, nesterov=True, *...
    method get_updates (line 76) | def get_updates(self, params, loss):
    method get_config (line 94) | def get_config(self):
  class RMSprop (line 102) | class RMSprop(Optimizer):
    method __init__ (line 103) | def __init__(self, lr=0.001, rho=0.9, epsilon=1e-6, *args, **kwargs):
    method get_updates (line 110) | def get_updates(self, params, loss):
    method get_config (line 123) | def get_config(self):
  class Adagrad (line 130) | class Adagrad(Optimizer):
    method __init__ (line 131) | def __init__(self, lr=0.01, epsilon=1e-6, *args, **kwargs):
    method get_updates (line 136) | def get_updates(self, params, constraints, loss):
    method get_config (line 148) | def get_config(self):
  class Adadelta (line 154) | class Adadelta(Optimizer):
    method __init__ (line 158) | def __init__(self, lr=0.1, rho=0.95, epsilon=1e-6, *args, **kwargs):
    method get_updates (line 164) | def get_updates(self, params, loss):
    method get_config (line 187) | def get_config(self):
  class Adam (line 194) | class Adam(Optimizer):  # new Adam is designed for our purpose.
    method __init__ (line 201) | def __init__(self, lr=0.001, beta_1=0.9, beta_2=0.999, epsilon=1e-8, s...
    method add_noise (line 216) | def add_noise(self, param):
    method add_forget (line 221) | def add_forget(self, param):
    method get_updates (line 226) | def get_updates(self, params, loss):
  function get (line 276) | def get(identifier, kwargs=None):

FILE: emolga/config.py
  function setup_ptb2 (line 5) | def setup_ptb2():

FILE: emolga/config_variant.py
  function setup_bienc (line 10) | def setup_bienc(config=None):
  function setup_dim (line 20) | def setup_dim(config=None):
  function setup_rep (line 35) | def setup_rep(config=None):
  function setup_opt (line 44) | def setup_opt(config=None):

FILE: emolga/dataset/build_dataset.py
  function serialize_to_file (line 15) | def serialize_to_file(obj, path, protocol=cPickle.HIGHEST_PROTOCOL):
  function show_txt (line 21) | def show_txt(array, path):
  function divide_dataset (line 29) | def divide_dataset(dataset, test_size, max_size):
  function deserialize_from_file (line 40) | def deserialize_from_file(path):
  function build_fuel (line 47) | def build_fuel(data):
  function obtain_stream (line 55) | def obtain_stream(dataset, batch_size, size=1):
  function build_ptb (line 70) | def build_ptb():
  function filter_unk (line 99) | def filter_unk(X, min_freq=5):
  function build_msr (line 125) | def build_msr():

FILE: emolga/layers/attention.py
  class Attention (line 11) | class Attention(Layer):
    method __init__ (line 12) | def __init__(self, target_dim, source_dim, hidden_dim,
    method __call__ (line 42) | def __call__(self, X, S,
  class CosineAttention (line 80) | class CosineAttention(Layer):
    method __init__ (line 81) | def __init__(self, target_dim, source_dim,
    method __call__ (line 115) | def __call__(self, X, S, Smask=None, return_log=False):

FILE: emolga/layers/core.py
  class Layer (line 8) | class Layer(object):
    method __init__ (line 9) | def __init__(self):
    method init_updates (line 15) | def init_updates(self):
    method _monitoring (line 18) | def _monitoring(self):
    method __call__ (line 26) | def __call__(self, X, *args, **kwargs):
    method _add (line 29) | def _add(self, layer):
    method supports_masked_input (line 34) | def supports_masked_input(self):
    method get_output_mask (line 40) | def get_output_mask(self, train=None):
    method set_weights (line 56) | def set_weights(self, weights):
    method get_weights (line 62) | def get_weights(self):
    method get_params (line 68) | def get_params(self):
    method set_name (line 71) | def set_name(self, name):
  class MaskedLayer (line 80) | class MaskedLayer(Layer):
    method supports_masked_input (line 85) | def supports_masked_input(self):
  class Identity (line 89) | class Identity(Layer):
    method __init__ (line 90) | def __init__(self, name='Identity'):
    method __call__ (line 95) | def __call__(self, X):
  class Dense (line 99) | class Dense(Layer):
    method __init__ (line 100) | def __init__(self, input_dim, output_dim, init='glorot_uniform', activ...
    method set_name (line 126) | def set_name(self, name):
    method __call__ (line 130) | def __call__(self, X):
    method reverse (line 134) | def reverse(self, Y):
  class Dense2 (line 141) | class Dense2(Layer):
    method __init__ (line 142) | def __init__(self, input_dim1, input_dim2, output_dim, init='glorot_un...
    method set_name (line 167) | def set_name(self, name):
    method __call__ (line 172) | def __call__(self, X1, X2):
  class Constant (line 177) | class Constant(Layer):
    method __init__ (line 178) | def __init__(self, input_dim, output_dim, init=None, activation='tanh'...
    method set_name (line 194) | def set_name(self, name):
    method __call__ (line 197) | def __call__(self, X=None):
  class MemoryLinear (line 205) | class MemoryLinear(Layer):
    method __init__ (line 206) | def __init__(self, input_dim, input_wdth, init='glorot_uniform',
    method __call__ (line 225) | def __call__(self, X=None):
  class Dropout (line 232) | class Dropout(MaskedLayer):
    method __init__ (line 236) | def __init__(self, rng=None, p=1., name=None):
    method __call__ (line 241) | def __call__(self, X, train=True):
  class Activation (line 251) | class Activation(MaskedLayer):
    method __init__ (line 255) | def __init__(self, activation):
    method __call__ (line 259) | def __call__(self, X):

FILE: emolga/layers/embeddings.py
  class Embedding (line 8) | class Embedding(Layer):
    method __init__ (line 17) | def __init__(self, input_dim, output_dim, init='uniform', name=None):
    method get_output_mask (line 31) | def get_output_mask(self, X):
    method __call__ (line 34) | def __call__(self, X, mask_zero=False, context=None):
  class Zero (line 68) | class Zero(Layer):
    method __call__ (line 69) | def __call__(self, X):
  class Bias (line 74) | class Bias(Layer):
    method __call__ (line 75) | def __call__(self, X):

FILE: emolga/layers/gridlstm.py
  class Grid (line 13) | class Grid(Recurrent):
    method __init__ (line 41) | def __init__(self,
    method build (line 86) | def build(self):
    method lstm_ (line 140) | def lstm_(self, k, H, m, x, identity=False):
    method grid_ (line 195) | def grid_(self,
  class GridLSTM3D (line 243) | class GridLSTM3D(Grid):
    method __init__ (line 248) | def __init__(self,
    method _step (line 335) | def _step(self, *args):
    method __call__ (line 386) | def __call__(self, X, init_H=None, init_M=None,
  class SequentialGridLSTM (line 450) | class SequentialGridLSTM(Grid):
    method __init__ (line 460) | def __init__(self,
    method _step (line 569) | def _step(self, *args):
    method __call__ (line 620) | def __call__(self, X, init_H=None, init_M=None,
  class PyramidGridLSTM2D (line 683) | class PyramidGridLSTM2D(Grid):
    method __init__ (line 687) | def __init__(self,
    method _step (line 772) | def _step(self, *args):
    method __call__ (line 825) | def __call__(self, X, init_x=None, init_y=None,
  class PyramidLSTM (line 876) | class PyramidLSTM(Layer):
    method __init__ (line 880) | def __init__(self,
    method _step (line 975) | def _step(self, *args):
    method __call__ (line 1033) | def __call__(self, X, init_x=None, init_y=None,

FILE: emolga/layers/ntm_minibatch.py
  class Reader (line 17) | class Reader(Layer):
    method __init__ (line 22) | def __init__(self, input_dim, memory_width, shift_width, shift_conv,
    method __call__ (line 67) | def __call__(self, X, w_temp, m_temp):
  class Writer (line 93) | class Writer(Reader):
    method __init__ (line 98) | def __init__(self, input_dim, memory_width, shift_width, shift_conv,
    method get_fixer (line 119) | def get_fixer(self, X):
  class Controller (line 125) | class Controller(Recurrent):
    method __init__ (line 134) | def __init__(self,
    method _controller (line 216) | def _controller(self, input_t, read_t, controller_tm1=None):
    method _read (line 231) | def _read(w_read, memory):
    method _write (line 239) | def _write(w_write, memory, erase, add):
    method _step (line 252) | def _step(self, input_t, mask_t,
    method __call__ (line 302) | def __call__(self, X, mask=None, M=None, init_ww=None,
  class AttentionReader (line 369) | class AttentionReader(Layer):
    method __init__ (line 374) | def __init__(self, input_dim, memory_width, shift_width, shift_conv,
    method __call__ (line 424) | def __call__(self, X, w_temp, m_temp):
  class AttentionWriter (line 454) | class AttentionWriter(AttentionReader):
    method __init__ (line 459) | def __init__(self, input_dim, memory_width, shift_width, shift_conv,
    method get_fixer (line 480) | def get_fixer(self, X):
  class BernoulliController (line 487) | class BernoulliController(Recurrent):
    method __init__ (line 496) | def __init__(self,
    method _controller (line 579) | def _controller(self, input_t, read_t, controller_tm1=None):
    method _read (line 594) | def _read(w_read, memory):
    method _write (line 602) | def _write(w_write, memory, erase, add):
    method _step (line 618) | def _step(self, input_t, mask_t,
    method __call__ (line 668) | def __call__(self, X, mask=None, M=None, init_ww=None,

FILE: emolga/layers/recurrent.py
  class Recurrent (line 6) | class Recurrent(MaskedLayer):
    method get_padded_shuffled_mask (line 12) | def get_padded_shuffled_mask(mask, pad=0):
  class GRU (line 31) | class GRU(Recurrent):
    method __init__ (line 54) | def __init__(self,
    method _step (line 115) | def _step(self,
    method _step_gate (line 128) | def _step_gate(self,
    method __call__ (line 141) | def __call__(self, X, mask=None, C=None, init_h=None,
  class JZS3 (line 226) | class JZS3(Recurrent):
    method __init__ (line 246) | def __init__(self,
    method _step (line 307) | def _step(self,
    method __call__ (line 318) | def __call__(self, X, mask=None, C=None, init_h=None, return_sequence=...
  class LSTM (line 373) | class LSTM(Recurrent):
    method __init__ (line 374) | def __init__(self,
    method _step (line 448) | def _step(self,
    method input_embed (line 468) | def input_embed(self, X, C=None):
    method __call__ (line 487) | def __call__(self, X, mask=None, C=None, init_h=None, init_c=None, ret...

FILE: emolga/models/core.py
  class Model (line 12) | class Model(object):
    method __init__ (line 13) | def __init__(self):
    method _add (line 19) | def _add(self, layer):
    method _monitoring (line 24) | def _monitoring(self):
    method compile_monitoring (line 32) | def compile_monitoring(self, inputs, updates=None):
    method set_weights (line 44) | def set_weights(self, weights):
    method get_weights (line 56) | def get_weights(self):
    method set_name (line 67) | def set_name(self, name):
    method save (line 75) | def save(self, filename):
    method load (line 90) | def load(self, filename):

FILE: emolga/models/covc_encdec.py
  class Encoder (line 22) | class Encoder(Model):
    method __init__ (line 28) | def __init__(self,
    method build_encoder (line 102) | def build_encoder(self, source, context=None, return_embed=False,
    method compile_encoder (line 175) | def compile_encoder(self, with_context=False, return_embed=False, retu...
  class Decoder (line 203) | class Decoder(Model):
    method __init__ (line 213) | def __init__(self,
    method _grab_prob (line 379) | def _grab_prob(probs, X, block_unk=False):
    method prepare_xy (line 392) | def prepare_xy(self, target):
    method build_decoder (line 407) | def build_decoder(self, target, context=None,
    method _step_sample (line 485) | def _step_sample(self, prev_word, prev_stat, context):
    method build_sampler (line 539) | def build_sampler(self):
    method build_stochastic_sampler (line 581) | def build_stochastic_sampler(self):
    method get_sample (line 592) | def get_sample(self, context, k=1, maxlen=30, stochastic=True, argmax=...
  class DecoderAtt (line 703) | class DecoderAtt(Decoder):
    method __init__ (line 708) | def __init__(self,
    method prepare_xy (line 768) | def prepare_xy(self, target, cc_matrix):
    method build_decoder (line 791) | def build_decoder(self,
    method _step_sample (line 918) | def _step_sample(self,
    method build_sampler (line 998) | def build_sampler(self):
    method get_sample (line 1052) | def get_sample(self,
  class FnnDecoder (line 1242) | class FnnDecoder(Model):
    method __init__ (line 1243) | def __init__(self, config, rng, prefix='fnndec'):
    method _grab_prob (line 1271) | def _grab_prob(probs, X):
    method build_decoder (line 1281) | def build_decoder(self, target, context):
    method build_sampler (line 1289) | def build_sampler(self):
    method get_sample (line 1296) | def get_sample(self, context, argmax=True):
  class RNNLM (line 1308) | class RNNLM(Model):
    method __init__ (line 1313) | def __init__(self,
    method build_ (line 1324) | def build_(self):
    method compile_ (line 1341) | def compile_(self, mode='train', contrastive=False):
    method compile_train (line 1367) | def compile_train(self):
    method compile_train_CE (line 1408) | def compile_train_CE(self):
    method compile_sample (line 1411) | def compile_sample(self):
    method compile_inference (line 1417) | def compile_inference(self):
    method default_context (line 1420) | def default_context(self):
    method generate_ (line 1428) | def generate_(self, context=None, max_len=None, mode='display'):
  class AutoEncoder (line 1456) | class AutoEncoder(RNNLM):
    method __init__ (line 1461) | def __init__(self,
    method build_ (line 1472) | def build_(self):
    method compile_train (line 1518) | def compile_train(self, mode='train'):
  class NRM (line 1561) | class NRM(Model):
    method __init__ (line 1566) | def __init__(self,
    method build_ (line 1583) | def build_(self, lr=None, iterations=None):
    method compile_ (line 1609) | def compile_(self, mode='all', contrastive=False):
    method compile_train (line 1632) | def compile_train(self):
    method compile_sample (line 1682) | def compile_sample(self):
    method compile_inference (line 1691) | def compile_inference(self):
    method generate_ (line 1694) | def generate_(self, inputs, mode='display', return_attend=False, retur...
    method evaluate_ (line 1737) | def evaluate_(self, inputs, outputs, idx2word, inputs_unk=None, encode...
    method analyse_ (line 1831) | def analyse_(self, inputs, outputs, idx2word, inputs_unk=None, return_...
    method analyse_cover (line 1888) | def analyse_cover(self, inputs, outputs, idx2word, inputs_unk=None, re...

FILE: emolga/models/encdec.py
  class Encoder (line 172) | class Encoder(Model):
    method __init__ (line 178) | def __init__(self,
    method build_encoder (line 252) | def build_encoder(self, source, context=None, return_embed=False, retu...
    method compile_encoder (line 288) | def compile_encoder(self, with_context=False, return_embed=False, retu...
  class Decoder (line 306) | class Decoder(Model):
    method __init__ (line 316) | def __init__(self,
    method _grab_prob (line 474) | def _grab_prob(probs, X):
    method prepare_xy (line 487) | def prepare_xy(self, target):
    method build_decoder (line 502) | def build_decoder(self, target, context=None,
    method _step_sample (line 580) | def _step_sample(self, prev_word, prev_stat, context):
    method build_sampler (line 634) | def build_sampler(self):
    method build_stochastic_sampler (line 677) | def build_stochastic_sampler(self):
    method get_sample (line 688) | def get_sample(self, context, k=1, maxlen=30, stochastic=True, argmax=...
  class DecoderAtt (line 799) | class DecoderAtt(Decoder):
    method __init__ (line 804) | def __init__(self,
    method prepare_xy (line 849) | def prepare_xy(self, target, context=None):
    method build_decoder (line 865) | def build_decoder(self,
    method _step_sample (line 948) | def _step_sample(self, prev_word, prev_stat, context, c_mask):
    method build_sampler (line 1010) | def build_sampler(self):
    method get_sample (line 1052) | def get_sample(self, context, c_mask, k=1, maxlen=30, stochastic=True,...
  class FnnDecoder (line 1164) | class FnnDecoder(Model):
    method __init__ (line 1165) | def __init__(self, config, rng, prefix='fnndec'):
    method _grab_prob (line 1193) | def _grab_prob(probs, X):
    method build_decoder (line 1203) | def build_decoder(self, target, context):
    method build_sampler (line 1211) | def build_sampler(self):
    method get_sample (line 1218) | def get_sample(self, context, argmax=True):
  class RNNLM (line 1230) | class RNNLM(Model):
    method __init__ (line 1235) | def __init__(self,
    method build_ (line 1246) | def build_(self):
    method compile_ (line 1263) | def compile_(self, mode='train', contrastive=False):
    method compile_train (line 1289) | def compile_train(self):
    method compile_train_CE (line 1330) | def compile_train_CE(self):
    method compile_sample (line 1333) | def compile_sample(self):
    method compile_inference (line 1339) | def compile_inference(self):
    method default_context (line 1342) | def default_context(self):
    method generate_ (line 1350) | def generate_(self, context=None, max_len=None, mode='display'):
  class AutoEncoder (line 1378) | class AutoEncoder(RNNLM):
    method __init__ (line 1383) | def __init__(self,
    method build_ (line 1394) | def build_(self):
    method compile_train (line 1440) | def compile_train(self, mode='train'):
  class NRM (line 1483) | class NRM(Model):
    method __init__ (line 1488) | def __init__(self,
    method build_ (line 1505) | def build_(self):
    method compile_ (line 1527) | def compile_(self, mode='all', contrastive=False):
    method compile_train (line 1550) | def compile_train(self):
    method compile_sample (line 1588) | def compile_sample(self):
    method compile_inference (line 1597) | def compile_inference(self):
    method generate_ (line 1600) | def generate_(self, inputs, mode='display', return_all=False):
    method evaluate_ (line 1662) | def evaluate_(self, inputs, outputs, idx2word, inputs_unk=None):
    method analyse_ (line 1689) | def analyse_(self, inputs, outputs, idx2word):
    method analyse_cover (line 1706) | def analyse_cover(self, inputs, outputs, idx2word):

FILE: emolga/models/ntm_encdec.py
  class RecurrentBase (line 20) | class RecurrentBase(Model):
    method __init__ (line 24) | def __init__(self, config, model='RNN', prefix='enc', use_contxt=True,...
    method get_context (line 105) | def get_context(self, context):
    method loop (line 127) | def loop(self, X, X_mask, info=None, return_sequence=False, return_ful...
    method step (line 135) | def step(self, X, prev_info):
    method build_ (line 153) | def build_(self):
    method get_init (line 209) | def get_init(self, context):
    method get_next_state (line 228) | def get_next_state(self, prev_X, prev_info):
  class Encoder (line 255) | class Encoder(Model):
    method __init__ (line 261) | def __init__(self,
    method build_encoder (line 306) | def build_encoder(self, source, context=None):
  class Decoder (line 347) | class Decoder(Model):
    method __init__ (line 357) | def __init__(self,
    method _grab_prob (line 462) | def _grab_prob(probs, X):
    method prepare_xy (line 475) | def prepare_xy(self, target):
    method build_decoder (line 490) | def build_decoder(self, target, context=None, return_count=False):
    method _step_embed (line 527) | def _step_embed(self, prev_word):
    method _step_sample (line 540) | def _step_sample(self, X, next_stat, context):
    method build_sampler (line 563) | def build_sampler(self):
    method get_sample (line 595) | def get_sample(self, context, k=1, maxlen=30, stochastic=True, argmax=...
  class RNNLM (line 717) | class RNNLM(Model):
    method __init__ (line 722) | def __init__(self,
    method build_ (line 733) | def build_(self):
    method compile_ (line 749) | def compile_(self, mode='train', contrastive=False):
    method compile_train (line 775) | def compile_train(self):
    method compile_train_CE (line 815) | def compile_train_CE(self):
    method compile_sample (line 818) | def compile_sample(self):
    method compile_inference (line 823) | def compile_inference(self):
    method default_context (line 826) | def default_context(self):
    method generate_ (line 834) | def generate_(self, context=None, mode='display', max_len=None):
  class Helmholtz (line 862) | class Helmholtz(RNNLM):
    method __init__ (line 871) | def __init__(self,
    method build_ (line 882) | def build_(self):
    method compile_train (line 951) | def compile_train(self):
    method compile_sample (line 1053) | def compile_sample(self):
    method compile_inference (line 1077) | def compile_inference(self):
    method default_context (line 1100) | def default_context(self):
  class BinaryHelmholtz (line 1105) | class BinaryHelmholtz(RNNLM):
    method __init__ (line 1114) | def __init__(self,
    method build_ (line 1125) | def build_(self):
    method compile_train (line 1175) | def compile_train(self):
    method compile_sample (line 1277) | def compile_sample(self):
    method compile_inference (line 1300) | def compile_inference(self):
    method default_context (line 1323) | def default_context(self):
  class AutoEncoder (line 1328) | class AutoEncoder(RNNLM):
    method __init__ (line 1334) | def __init__(self,
    method build_ (line 1345) | def build_(self):
    method compile_train (line 1368) | def compile_train(self, mode='train'):
    method compile_sample (line 1404) | def compile_sample(self):

FILE: emolga/models/pointers.py
  class PtrDecoder (line 19) | class PtrDecoder(Model):
    method __init__ (line 23) | def __init__(self,
    method grab_prob (line 64) | def grab_prob(probs, X):
    method grab_source (line 75) | def grab_source(source, target):
    method build_decoder (line 91) | def build_decoder(self,
    method _step_sample (line 139) | def _step_sample(self, prev_idx, prev_stat,
    method build_sampler (line 156) | def build_sampler(self):
    method get_sample (line 199) | def get_sample(self, context, inputs, source, smask,
  class PointerDecoder (line 316) | class PointerDecoder(Model):
    method __init__ (line 321) | def __init__(self,
    method grab_prob (line 370) | def grab_prob(probs, X):
    method grab_source (line 381) | def grab_source(source, target):
    method build_decoder (line 397) | def build_decoder(self,
    method _step_sample (line 473) | def _step_sample(self,
    method build_sampler (line 495) | def build_sampler(self):
    method get_sample (line 540) | def get_sample(self, context, inputs, source, smask,
  class MemNet (line 616) | class MemNet(Model):
    method __init__ (line 621) | def __init__(self,
    method __call__ (line 653) | def __call__(self, key, memory=None, mem_mask=None, out_memory=None):
  class PtrNet (line 675) | class PtrNet(Model):
    method __init__ (line 679) | def __init__(self, config, n_rng, rng,
    method build_ (line 689) | def build_(self, encoder=None):
    method build_train (line 743) | def build_train(self, memory=None, out_memory=None, compile_train=Fals...
    method build_sampler (line 881) | def build_sampler(self, memory=None, out_mem=None):
    method build_predict_sampler (line 925) | def build_predict_sampler(self):
    method generate_ (line 973) | def generate_(self, inputs, context, source, smask):

FILE: emolga/models/variational.py
  class VAE (line 19) | class VAE(RNNLM):
    method __init__ (line 33) | def __init__(self,
    method _add_tag (line 45) | def _add_tag(self, layer, tag):
    method build_ (line 52) | def build_(self):
    method compile_train (line 115) | def compile_train(self):
    method compile_sample (line 176) | def compile_sample(self):
    method compile_inference (line 192) | def compile_inference(self):
    method default_context (line 208) | def default_context(self):
  class Helmholtz (line 212) | class Helmholtz(VAE):
    method __init__ (line 220) | def __init__(self,
    method build_ (line 235) | def build_(self):
    method dynamic (line 301) | def dynamic(self):
    method compile_ (line 330) | def compile_(self, mode='train', contrastive=False):
    method compile_train (line 356) | def compile_train(self):
    method build_dynamics (line 477) | def build_dynamics(self, states, action, Y):
    method compile_sample (line 487) | def compile_sample(self):
    method compile_inference (line 523) | def compile_inference(self):
    method evaluate_ (line 539) | def evaluate_(self, inputs):
    method compile_train_CE (line 564) | def compile_train_CE(self):
  class HarX (line 695) | class HarX(Helmholtz):
    method __init__ (line 705) | def __init__(self,
    method build_ (line 720) | def build_(self):
    method compile_ (line 798) | def compile_(self, mode='train', contrastive=False):
    method compile_train (line 824) | def compile_train(self):
    method generate_ (line 1015) | def generate_(self, context=None, max_len=None, mode='display'):
  class THarX (line 1023) | class THarX(Helmholtz):
    method __init__ (line 1033) | def __init__(self,
    method build_ (line 1048) | def build_(self):
    method compile_ (line 1126) | def compile_(self, mode='train', contrastive=False):
    method compile_train (line 1152) | def compile_train(self):
    method generate_ (line 1350) | def generate_(self, context=None, max_len=None, mode='display'):
  class NVTM (line 1358) | class NVTM(Helmholtz):
    method __init__ (line 1364) | def __init__(self,
    method build_ (line 1376) | def build_(self):
    method compile_ (line 1492) | def compile_(self, mode='train', contrastive=False):
    method compile_train (line 1518) | def compile_train(self):
    method generate_ (line 1701) | def generate_(self, context=None, max_len=None, mode='display'):

FILE: emolga/run.py
  function simulator (line 79) | def simulator(M=25, display=False):
  function learner (line 162) | def learner(data, fr=1., fs=1., fb=1.):
  function SL_learner (line 182) | def SL_learner(data, batch_size=25):
  function main (line 226) | def main():
  function check_answer (line 246) | def check_answer(x, y, g):
  function display_session (line 258) | def display_session(x, y, g, t, acc, cov):
  function main_sl (line 282) | def main_sl():

FILE: emolga/test_lm.py
  function init_logging (line 18) | def init_logging(logfile):
  function prepare_batch (line 98) | def prepare_batch(batch):

FILE: emolga/test_nvtm.py
  function init_logging (line 18) | def init_logging(logfile):
  function prepare_batch (line 98) | def prepare_batch(batch):

FILE: emolga/test_run.py
  function simulator (line 82) | def simulator(M=25, display=False):
  function learner (line 165) | def learner(data, fr=1., fs=1., fb=1.):
  function SL_learner (line 185) | def SL_learner(data, batch_size=25, eval_freq=0, eval_train=None, eval_t...
  function main (line 261) | def main():
  function check_answer (line 281) | def check_answer(x, y, g):
  function display_session (line 293) | def display_session(x, y, g, t, acc, cov):
  function SL_test (line 317) | def SL_test(test_set):
  function main_sl (line 346) | def main_sl():

FILE: emolga/utils/generic_utils.py
  function get_from_module (line 10) | def get_from_module(identifier, module_params, module_name, instantiate=...
  function make_tuple (line 24) | def make_tuple(*args):
  function printv (line 28) | def printv(v, prefix=''):
  function make_batches (line 50) | def make_batches(size, batch_size):
  function slice_X (line 55) | def slice_X(X, start=None, stop=None):
  class Progbar (line 68) | class Progbar(object):
    method __init__ (line 69) | def __init__(self, target, width=30, verbose=1):
    method update (line 82) | def update(self, current, values=[]):
    method add (line 152) | def add(self, n, values=[]):
    method clear (line 155) | def clear(self):
  function print_sample (line 162) | def print_sample(idx2word, idx):
  function visualize_ (line 172) | def visualize_(subplots, data, w=None, h=None, name=None,
  function vis_Gaussian (line 249) | def vis_Gaussian(subplot, mean, std, name=None, display='off', size=10):

FILE: emolga/utils/io_utils.py
  class HDF5Matrix (line 8) | class HDF5Matrix():
    method __init__ (line 11) | def __init__(self, datapath, dataset, start, end, normalizer=None):
    method __len__ (line 22) | def __len__(self):
    method __getitem__ (line 25) | def __getitem__(self, key):
    method shape (line 52) | def shape(self):
  function save_array (line 56) | def save_array(array, name):
  function load_array (line 65) | def load_array(name):
  function save_config (line 75) | def save_config():
  function load_config (line 79) | def load_config():

FILE: emolga/utils/np_utils.py
  function to_categorical (line 8) | def to_categorical(y, nb_classes=None):
  function normalize (line 21) | def normalize(a, axis=-1, order=2):
  function binary_logloss (line 27) | def binary_logloss(p, y):
  function multiclass_logloss (line 36) | def multiclass_logloss(P, Y):
  function accuracy (line 43) | def accuracy(p, y):
  function probas_to_classes (line 47) | def probas_to_classes(y_pred):
  function categorical_probas_to_classes (line 53) | def categorical_probas_to_classes(p):

FILE: emolga/utils/test_utils.py
  function get_test_data (line 4) | def get_test_data(nb_train=1000, nb_test=500, input_shape=(10,), output_...

FILE: emolga/utils/theano_utils.py
  function floatX (line 10) | def floatX(X):
  function sharedX (line 14) | def sharedX(X, dtype=theano.config.floatX, name=None):
  function shared_zeros (line 18) | def shared_zeros(shape, dtype=theano.config.floatX, name=None):
  function shared_scalar (line 22) | def shared_scalar(val=0., dtype=theano.config.floatX, name=None):
  function shared_ones (line 26) | def shared_ones(shape, dtype=theano.config.floatX, name=None):
  function alloc_zeros_matrix (line 30) | def alloc_zeros_matrix(*dims):
  function alloc_ones_matrix (line 34) | def alloc_ones_matrix(*dims):
  function ndim_tensor (line 38) | def ndim_tensor(ndim):
  function ndim_itensor (line 51) | def ndim_itensor(ndim, name=None):
  function dot (line 62) | def dot(inp, matrix, bias=None):
  function logSumExp (line 87) | def logSumExp(x, axis=None, mask=None, status='theano', c=None, err=1e-7):
  function softmax (line 121) | def softmax(x):
  function masked_softmax (line 125) | def masked_softmax(x, mask, err=1e-9):
  function cosine_sim (line 133) | def cosine_sim(k, M):
  function cosine_sim2d (line 144) | def cosine_sim2d(k, M):
  function dot_2d (line 160) | def dot_2d(k, M, b=None, g=None):
  function shift_convolve (line 182) | def shift_convolve(weight, shift, shift_conv):
  function shift_convolve2d (line 187) | def shift_convolve2d(weight, shift, shift_conv):

FILE: experiments/bst_dataset.py
  class BSTnode (line 7) | class BSTnode(object):
    method __init__ (line 12) | def __init__(self, parent, t):
    method update_stats (line 20) | def update_stats(self):
    method insert (line 24) | def insert(self, t, NodeType):
    method find (line 42) | def find(self, t):
    method rank (line 57) | def rank(self, t):
    method minimum (line 73) | def minimum(self):
    method successor (line 81) | def successor(self):
    method delete (line 90) | def delete(self):
    method check (line 111) | def check(self, lokey, hikey):
    method __repr__ (line 128) | def __repr__(self):
  class BST (line 132) | class BST(object):
    method __init__ (line 140) | def __init__(self, NodeType=BSTnode):
    method reroot (line 145) | def reroot(self):
    method insert (line 148) | def insert(self, t):
    method find (line 157) | def find(self, t):
    method rank (line 164) | def rank(self, t):
    method delete (line 171) | def delete(self, t):
    method check (line 178) | def check(self):
    method __str__ (line 182) | def __str__(self):
  function printsizes (line 233) | def printsizes(node):
  function test (line 240) | def test(args=None, BSTtype=BST):
  function generate (line 263) | def generate():
  function obtain_dataset (line 290) | def obtain_dataset():

FILE: experiments/bst_vest.py
  function init_logging (line 28) | def init_logging(logfile):
  function build_data (line 72) | def build_data(data):
  function output_stream (line 127) | def output_stream(dataset, batch_size, size=1):
  function prepare_batch (line 136) | def prepare_batch(batch, mask, fix_len=None):
  function cc_martix (line 151) | def cc_martix(source, target):
  function unk_filter (line 160) | def unk_filter(data):
  function analysis_ (line 220) | def analysis_(data_plain, t_idx, mode='Training'):

FILE: experiments/config.py
  function setup (line 6) | def setup():
  function setup_syn (line 92) | def setup_syn():
  function setup_bst (line 280) | def setup_bst():
  function setup_lcsts (line 381) | def setup_lcsts():
  function setup_weibo (line 499) | def setup_weibo():

FILE: experiments/copynet.py
  function init_logging (line 20) | def init_logging(logfile):
  function build_data (line 64) | def build_data(source, target):
  function output_stream (line 91) | def output_stream(dataset, batch_size, size=1):
  function prepare_batch (line 100) | def prepare_batch(batch, mask):

FILE: experiments/copynet_input.py
  function init_logging (line 22) | def init_logging(logfile):
  function unk_filter (line 65) | def unk_filter(data):

FILE: experiments/dataset.py
  function repeat_name (line 33) | def repeat_name(l):
  function replace (line 52) | def replace(word):
  function print_str (line 82) | def print_str(data):

FILE: experiments/lcsts_dataset.py
  function build_data (line 117) | def build_data(data):

FILE: experiments/lcsts_rouge.py
  function build_evaluation (line 24) | def build_evaluation(train_set, segment):
  function init_logging (line 73) | def init_logging(logfile):
  function unk_filter (line 141) | def unk_filter(data):

FILE: experiments/lcsts_sample.py
  function init_logging (line 25) | def init_logging(logfile):
  function build_data (line 68) | def build_data(data):
  function unk_filter (line 79) | def unk_filter(data):

FILE: experiments/lcsts_test.py
  function init_logging (line 24) | def init_logging(logfile):
  function build_data (line 64) | def build_data(data):
  function output_stream (line 107) | def output_stream(dataset, batch_size, size=1):
  function prepare_batch (line 116) | def prepare_batch(batch, mask, fix_len=None):

FILE: experiments/lcsts_vest.py
  function init_logging (line 26) | def init_logging(logfile):
  function build_data (line 70) | def build_data(data):
  function output_stream (line 126) | def output_stream(dataset, batch_size, size=1):
  function prepare_batch (line 135) | def prepare_batch(batch, mask, fix_len=None):
  function cc_martix (line 150) | def cc_martix(source, target):
  function unk_filter (line 159) | def unk_filter(data):

FILE: experiments/lcsts_vest_new.py
  function init_logging (line 26) | def init_logging(logfile):
  function build_data (line 70) | def build_data(data):
  function output_stream (line 126) | def output_stream(dataset, batch_size, size=1):
  function prepare_batch (line 135) | def prepare_batch(batch, mask, fix_len=None):
  function cc_martix (line 150) | def cc_martix(source, target):
  function unk_filter (line 159) | def unk_filter(data):

FILE: experiments/movie_dataset.py
  function mark (line 15) | def mark(line):
  function build_data (line 75) | def build_data(data):

FILE: experiments/syn_vest.py
  function init_logging (line 27) | def init_logging(logfile):
  function build_data (line 71) | def build_data(data):
  function output_stream (line 122) | def output_stream(dataset, batch_size, size=1):
  function prepare_batch (line 131) | def prepare_batch(batch, mask, fix_len=None):
  function cc_martix (line 146) | def cc_martix(source, target):
  function unk_filter (line 155) | def unk_filter(data):
  function judge_rule (line 207) | def judge_rule(rule):
  function analysis_ (line 217) | def analysis_(data_plain, mode='Training'):

FILE: experiments/syntest.py
  function init_logging (line 23) | def init_logging(logfile):
  function build_data (line 74) | def build_data(data):
  function output_stream (line 118) | def output_stream(dataset, batch_size, size=1):
  function prepare_batch (line 127) | def prepare_batch(batch, mask, fix_len=None):

FILE: experiments/synthetic.py
  function ftr (line 37) | def ftr(v):
  function build_instance (line 46) | def build_instance():

FILE: experiments/weibo_dataset.py
  function build_data (line 75) | def build_data(data):

FILE: experiments/weibo_vest.py
  function init_logging (line 25) | def init_logging(logfile):
  function build_data (line 69) | def build_data(data):
  function output_stream (line 116) | def output_stream(dataset, batch_size, size=1):
  function prepare_batch (line 125) | def prepare_batch(batch, mask, fix_len=None):
  function cc_martix (line 140) | def cc_martix(source, target):
  function unk_filter (line 149) | def unk_filter(data):