documentation updated

docs/tutorial_nnlib.rst

@@ -43,7 +43,7 @@ Then we choose a neural network. We choose ConvNet:
 
 Initialize the classifier:
 
->>> classifier = shorttext.classifiers.VarNNEmbeddedVecClassifier(wvmodel, vecsize=300)
+>>> classifier = shorttext.classifiers.VarNNEmbeddedVecClassifier(wvmodel)
 
 Then train the classifier:
 
@@ -99,11 +99,11 @@ Then load the training data
 
 Then we choose a neural network. We choose ConvNet and set `with_gensim` as `True`:
 
->>> kmodel = shorttext.classifiers.frameworks.CNNWordEmbed(wvmodel=wvmodel, nb_labels=len(trainclassdict.keys()), vecsize=100, with_gensim=True)
+>>> kmodel = shorttext.classifiers.frameworks.CNNWordEmbed(wvmodel=wvmodel, nb_labels=len(trainclassdict.keys()) with_gensim=True)
 
 Initialize the classifier and set `with_gensim` as `True`:
 
->>> classifier = shorttext.classifiers.VarNNEmbeddedVecClassifier(wvmodel, with_gensim=True, vecsize=100)
+>>> classifier = shorttext.classifiers.VarNNEmbeddedVecClassifier(wvmodel, with_gensim=True)
 
 Then train the classifier:
 
@@ -151,62 +151,30 @@ as demonstrated in Kim's paper.
 
 The function in the frameworks returns a :class:`keras.models.Sequential` or :class:`keras.models.Model`. Its input parameters are:
 
-`CNNWordEmbed(nb_labels, wvmodel=None, nb_filters=1200, n_gram=2, maxlen=15, vecsize=100, cnn_dropout=0.0, final_activation='softmax', dense_wl2reg=0.0, dense_bl2reg=0.0, optimizer='adam', with_gensim=False)`
-
-* nb_labels (int) – number of class labels
-* wvmodel (gensim.models.keyedvectors.KeyedVectors) – pre-trained Gensim word2vec model
-* nb_filters (int) – number of filters (Default: 1200)
-* n_gram (int) – n-gram, or window size of CNN/ConvNet (Default: 2)
-* maxlen (int) – maximum number of words in a sentence (Default: 15)
-* vecsize (int) – length of the embedded vectors in the model (Default: 100)
-* cnn_dropout (float) – dropout rate for CNN/ConvNet (Default: 0.0)
-* final_activation (str) – activation function. Options: softplus, softsign, relu, tanh, sigmoid, hard_sigmoid, linear. (Default: ‘softmax’)
-* dense_wl2reg (float) – L2 regularization coefficient (Default: 0.0)
-* dense_bl2reg (float) – L2 regularization coefficient for bias (Default: 0.0)
-* optimizer (str) – optimizer for gradient descent. Options: sgd, rmsprop, adagrad, adadelta, adam, adamax, nadam. (Default: adam)
-* with_gensim (bool) – boolean variable to indicate if the word-embeddings being used derived from a Gensim’s Word2Vec model. (Default: True)
-
 The parameter `maxlen` defines the maximum length of the sentences. If the sentence has less than `maxlen`
 words, then the empty words will be filled with zero vectors.
 
->>> kmodel = fr.CNNWordEmbed(len(trainclassdict.keys()))
+>>> kmodel = fr.CNNWordEmbed(len(trainclassdict.keys()), vecsize=wvmodel.vector_size)
 
 Or if you want to include word-embedding layer, do this: (`shorttext` >= 0.4.0)
 
 >>> wvmodel = shorttext.utils.load_word2vec_model('/path/to/gensim/w2vmodel')
->>> kmodel = fr.CNNWordEmbed(wvmodel=wvmodel, nb_labels=len(trainclassdict.keys()), vecsize=100, with_gensim=True)
+>>> kmodel = fr.CNNWordEmbed(wvmodel=wvmodel, nb_labels=len(trainclassdict.keys()), vecsize=wvmodel.vector_size, with_gensim=True)
 
 Double ConvNet
 ^^^^^^^^^^^^^^
 
 This neural network is nothing more than two ConvNet layers. The function in the frameworks returns a :class:`keras.models.Sequential` or :class:`keras.models.Model`. Its input parameters are:
 
-`DoubleCNNWordEmbed(nb_labels, wvmodel=None, nb_filters_1=1200, nb_filters_2=600, n_gram=2, filter_length_2=10, maxlen=15, vecsize=100, cnn_dropout_1=0.0, cnn_dropout_2=0.0, final_activation='softmax', dense_wl2reg=0.0, dense_bl2reg=0.0, optimizer='adam', with_gensim=False)`
-
-* nb_labels (int) – number of class labels
-* wvmodel (gensim.models.keyedvectors.KeyedVectors) – pre-trained Gensim word2vec model
-* nb_filters_1 (int) – number of filters for the first CNN/ConvNet layer (Default: 1200)
-* nb_filters_2 (int) – number of filters for the second CNN/ConvNet layer (Default: 600)
-* n_gram (int) – n-gram, or window size of first CNN/ConvNet (Default: 2)
-* filter_length_2 (int) – window size for second CNN/ConvNet layer (Default: 10)
-* maxlen (int) – maximum number of words in a sentence (Default: 15)
-* vecsize (int) – length of the embedded vectors in the model (Default: 100)
-* cnn_dropout_1 (float) – dropout rate for the first CNN/ConvNet layer (Default: 0.0)
-* cnn_dropout_2 (float) – dropout rate for the second CNN/ConvNet layer (Default: 0.0)
-* final_activation (str) – activation function. Options: softplus, softsign, relu, tanh, sigmoid, hard_sigmoid, linear. (Default: ‘softmax’)
-* dense_wl2reg (float) – L2 regularization coefficient (Default: 0.0)
-* dense_bl2reg (float) – L2 regularization coefficient for bias (Default: 0.0)
-* optimizer (str) – optimizer for gradient descent. Options: sgd, rmsprop, adagrad, adadelta, adam, adamax, nadam. (Default: adam)
-
 The parameter `maxlen` defines the maximum length of the sentences. If the sentence has less than `maxlen`
 words, then the empty words will be filled with zero vectors.
 
->>> kmodel = fr.DoubleCNNWordEmbed(len(trainclassdict.keys()))
+>>> kmodel = fr.DoubleCNNWordEmbed(len(trainclassdict.keys()), vecsize=wvmodel.vector_size)
 
 Or if you want to include word-embedding layer, do this: (`shorttext` >= 0.4.0)
 
 >>> wvmodel = shorttext.utils.load_word2vec_model('/path/to/gensim/w2vmodel')
->>> kmodel = fr.DoubleCNNWordEmbed(wvmodel=wvmodel, nb_labels=len(trainclassdict.keys()), vecsize=100, with_gensim=True)
+>>> kmodel = fr.DoubleCNNWordEmbed(wvmodel=wvmodel, nb_labels=len(trainclassdict.keys()), vecsize=wvmodel.vector_size, with_gensim=True)
 
 C-LSTM (Convolutional Long Short-Term Memory)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
@@ -219,31 +187,15 @@ and then followed by LSTM (long short-term memory), a type of recurrent neural n
 
 The function in the frameworks returns a :class:`keras.models.Sequential` or :class:`keras.models.Model`.
 
-`CLSTMWordEmbed(nb_labels, wvmodel=None, nb_filters=1200, n_gram=2, maxlen=15, vecsize=100, cnn_dropout=0.0, nb_rnnoutdim=1200, rnn_dropout=0.2, final_activation='softmax', dense_wl2reg=0.0, dense_bl2reg=0.0, optimizer='adam', with_gensim=False)`
-
-* nb_labels (int) – number of class labels
-* wvmodel (gensim.models.keyedvectors.KeyedVectors) – pre-trained Gensim word2vec model
-* nb_filters (int) – number of filters (Default: 1200)
-* n_gram (int) – n-gram, or window size of CNN/ConvNet (Default: 2)
-* maxlen (int) – maximum number of words in a sentence (Default: 15)
-* vecsize (int) – length of the embedded vectors in the model (Default: 100)
-* cnn_dropout (float) – dropout rate for CNN/ConvNet (Default: 0.0)
-* nb_rnnoutdim (int) – output dimension for the LSTM networks (Default: 1200)
-* rnn_dropout (float) – dropout rate for LSTM (Default: 0.2)
-* final_activation (str) – activation function. Options: softplus, softsign, relu, tanh, sigmoid, hard_sigmoid, linear. (Default: ‘softmax’)
-* dense_wl2reg (float) – L2 regularization coefficient (Default: 0.0)
-* dense_bl2reg (float) – L2 regularization coefficient for bias (Default: 0.0)
-* optimizer (str) – optimizer for gradient descent. Options: sgd, rmsprop, adagrad, adadelta, adam, adamax, nadam. (Default: adam)
-
 The parameter `maxlen` defines the maximum length of the sentences. If the sentence has less than `maxlen`
 words, then the empty words will be filled with zero vectors.
 
->>> kmodel = fr.CLSTMWordEmbed(len(trainclassdict.keys()))
+>>> kmodel = fr.CLSTMWordEmbed(len(trainclassdict.keys()), vecsize=wvmodel.vector_size)
 
 Or if you want to include word-embedding layer, do this: (`shorttext` >= 0.4.0)
 
 >>> wvmodel = shorttext.utils.load_word2vec_model('/path/to/gensim/w2vmodel')
->>> kmodel = fr.CLSTMWordEmbed(wvmodel=wvmodel, nb_labels=len(trainclassdict.keys()), vecsize=100, with_gensim=True)
+>>> kmodel = fr.CLSTMWordEmbed(wvmodel=wvmodel, nb_labels=len(trainclassdict.keys()), vecsize=wvmodel.vector_size, with_gensim=True)
 
 User-Defined Neural Network
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^

docs/tutorial_sumvec.rst

@@ -32,7 +32,7 @@ Then we load a set of data:
 
 Then initialize the classifier:
 
->>> classifier = shorttext.classifiers.SumEmbeddedVecClassifier(wvmodel, vecsize=300)   # for Google model, the vector size is 300 (default: 100)
+>>> classifier = shorttext.classifiers.SumEmbeddedVecClassifier(wvmodel)   # for Google model, the vector size is 300 (default: 100)
 >>> classifier.train(nihtraindata)
 
 This classifier takes relatively little time to train compared with others

docs/tutorial_topic.rst

@@ -197,8 +197,9 @@ A list include:
     shorttext.classifiers.load_autoencoder_topic  ->  shorttext.generators.load_autoencoder_topicmodel
 
 
-For backward compatibility, developers can still call the topic models as if there were no such changes,
-although they are advised to make this change.
+Before release 0.5.6, for backward compatibility, developers can still call the topic models as if there were no such changes,
+although they are advised to make this change. However, *effective release 0.5.7, this backward compatibility is no longer
+available.*
 
 Classification Using Cosine Similarity
 --------------------------------------

docs/tutorial_wordembed.rst

@@ -75,7 +75,11 @@ in the documentation of `gensim`: `Converting GloVe to Word2Vec
 FastText
 --------
 
-FastText is a similar word-embedding model from Facebook.
+FastText is a similar word-embedding model from Facebook. You can download pre-trained models here:
+
+`Pre-trained word vectors
+<https://github.com/facebookresearch/fastText/blob/master/pretrained-vectors.md>`_
+
 
 Links
 -----