Added the ability to change the dimension of Glove embeddings

deepsegment/train.py

@@ -106,7 +106,7 @@ def generate_data(lines, max_sents_per_example=6, n_examples=1000):
 
     return x, y
 
-def train(x, y, vx, vy, epochs, batch_size, save_folder, glove_path):
+def train(x, y, vx, vy, epochs, batch_size, save_folder, glove_path,embedding_dim):
     """
         Trains a deepsegment model.
 
@@ -126,7 +126,9 @@ def train(x, y, vx, vy, epochs, batch_size, save_folder, glove_path):
 
         save_folder (str): path for the directory where checkpoints should be saved.
 
-        glove_path (str): path to 100d word vectors.
+        glove_path (str): path to word vectors.
+
+        dim_vectors (str): dimension of word vectors. (50, 100, 200, 300)
 
 
     """
@@ -141,7 +143,7 @@ def train(x, y, vx, vy, epochs, batch_size, save_folder, glove_path):
     checkpoint = ModelCheckpoint(checkpoint_path, verbose=1, save_best_only=True, mode='max', monitor='f1')
     earlystop = EarlyStopping(patience=3, monitor='f1', mode='max')
 
-    model = seqtag_keras.Sequence(embeddings=embeddings)
+    model = seqtag_keras.Sequence(embeddings=embeddings,word_embedding_dim=embedding_dim,word_lstm_size=embedding_dim)
 
     model.fit(x, y, x_valid=vx, y_valid=vy, epochs=epochs, batch_size=batch_size, callbacks=[checkpoint, earlystop])
 
@@ -154,7 +156,7 @@ def train(x, y, vx, vy, epochs, batch_size, save_folder, glove_path):
     'italian': 'it'
 }
 
-def finetune(lang_code, x, y, vx, vy, name=None, epochs=5, batch_size=16, lr=0.0001):
+def finetune(lang_code, x, y, vx, vy, name=None, epochs=5, batch_size=16, lr=0.0001,embedding_dim=100):
     """
         Finetunes an existing deepsegment model.
 
@@ -197,9 +199,9 @@ def finetune(lang_code, x, y, vx, vy, name=None, epochs=5, batch_size=16, lr=0.0
     model = BiLSTMCRF(char_vocab_size=p.char_vocab_size,
                           word_vocab_size=p.word_vocab_size,
                           num_labels=p.label_size,
-                          word_embedding_dim=100,
+                          word_embedding_dim=embedding_dim,
                           char_embedding_dim=25,
-                          word_lstm_size=100,
+                          word_lstm_size=embedding_dim,
                           char_lstm_size=25,
                           fc_dim=100,
                           dropout=0.2,