label2Idx[bioTag+nerClass+subtype] = idx
idx += 1
#Inverse label mapping
idx2Label = {v: k for k, v in label2Idx.items()}
#Casing matrix
caseLookup = {'numeric': 0, 'allLower':1, 'allUpper':2, 'initialUpper':3, 'other':4, 'mainly_numeric':5, 'contains_digit': 6, 'PADDING':7}
caseMatrix = np.identity(len(caseLookup), dtype=theano.config.floatX)
# 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_case_x, train_y = GermEvalReader.createNumpyArrayWithCasing(train_sentences, windowSize, word2Idx, label2Idx, caseLookup)
dev_x, dev_case_x, dev_y = GermEvalReader.createNumpyArrayWithCasing(dev_sentences, windowSize, word2Idx, label2Idx, caseLookup)
test_x, test_case_x, test_y = GermEvalReader.createNumpyArrayWithCasing(test_sentences, windowSize, word2Idx, label2Idx, caseLookup)
#####################################
#
# Create the Network
#
#####################################
"numeric": 0,
"allLower": 1,
"allUpper": 2,
"initialUpper": 3,
"other": 4,
"mainly_numeric": 5,
"contains_digit": 6,
"PADDING": 7,
}
caseMatrix = np.identity(len(caseLookup), dtype=theano.config.floatX)
# 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_case_x, train_y = GermEvalReader.createNumpyArrayWithCasing(
train_sentences, windowSize, word2Idx, label2Idx, caseLookup
)
dev_x, dev_case_x, dev_y = GermEvalReader.createNumpyArrayWithCasing(
dev_sentences, windowSize, word2Idx, label2Idx, caseLookup
)
test_x, test_case_x, test_y = GermEvalReader.createNumpyArrayWithCasing(
test_sentences, windowSize, word2Idx, label2Idx, caseLookup
)
