for nerClass in ['PER', 'LOC', 'ORG', 'OTH']:
for subtype in ['', 'deriv', 'part']:
label2Idx[bioTag + nerClass + subtype] = idx
idx += 1
#Inverse label mapping
idx2Label = {v: k for k, v in label2Idx.items()}
# Read in data
print "Read in data and create matrices"
train_sentences = GermEvalReader.readFile(trainFile)
dev_sentences = GermEvalReader.readFile(devFile)
test_sentences = GermEvalReader.readFile(testFile)
# Create numpy arrays
train_x, train_y = GermEvalReader.createNumpyArray(train_sentences, windowSize,
word2Idx, label2Idx)
dev_x, dev_y = GermEvalReader.createNumpyArray(dev_sentences, windowSize,
word2Idx, label2Idx)
test_x, test_y = GermEvalReader.createNumpyArray(test_sentences, windowSize,
word2Idx, label2Idx)
#####################################
#
# Create the Lasagne Network
#
#####################################
def build_network(input_var, n_in, n_hidden, n_out, embedding_matrix):
### -----> Put your code here to build the network <-------
# l_out: Your network (see MNIST example from last week)
label2Idx[bioTag+nerClass+subtype] = idx
idx += 1
#Inverse label mapping
idx2Label = {v: k for k, v in label2Idx.items()}
# Read in data
print "Read in data and create matrices"
train_sentences = GermEvalReader.readFile(trainFile)
dev_sentences = GermEvalReader.readFile(devFile)
test_sentences = GermEvalReader.readFile(testFile)
# Create numpy arrays
train_x, train_y = GermEvalReader.createNumpyArray(train_sentences, windowSize, word2Idx, label2Idx)
dev_x, dev_y = GermEvalReader.createNumpyArray(dev_sentences, windowSize, word2Idx, label2Idx)
test_x, test_y = GermEvalReader.createNumpyArray(test_sentences, windowSize, word2Idx, label2Idx)
#####################################
#
# Create the Lasagne Network
#
#####################################
def build_network(input_var, n_in, n_hidden, n_out, embedding_matrix):
### -----> Put your code here to build the network <-------
# l_out: Your network (see MNIST example from last week)
# params: The parameters you would like to train
