def __init__(self,
model_type,
filename,
labels=None,
feature_extractor=None,
model=None):
if model_type is None:
model_type = SPARSE_PERCEPTRON
self.model_type = model_type
self.filename = filename
if feature_extractor is not None and model is not None:
self.feature_extractor = feature_extractor
self.model = model
return
if model_type == SPARSE_PERCEPTRON:
from features.sparse_features import SparseFeatureExtractor
from linear.sparse_perceptron import SparsePerceptron
self.feature_extractor = SparseFeatureExtractor()
self.model = SparsePerceptron(filename, labels)
elif model_type == DENSE_PERCEPTRON:
from features.embedding import FeatureEmbedding
from linear.dense_perceptron import DensePerceptron
self.feature_extractor = self.dense_features_wrapper(
FeatureEmbedding)
self.model = DensePerceptron(
filename,
labels,
num_features=self.feature_extractor.num_features())
elif model_type == MLP_NN:
from features.enumerator import FeatureEnumerator
from nn.feedforward import MLP
self.feature_extractor = self.dense_features_wrapper(
FeatureEnumerator)
self.model = MLP(filename,
labels,
input_params=self.feature_extractor.params)
elif model_type == BILSTM_NN:
from features.enumerator import FeatureEnumerator
from features.indexer import FeatureIndexer
from nn.bilstm import BiLSTM
self.feature_extractor = FeatureIndexer(
self.dense_features_wrapper(FeatureEnumerator))
self.model = BiLSTM(filename,
labels,
input_params=self.feature_extractor.params)
else:
raise ValueError("Invalid model type: '%s'" % model_type)
def create_model(model_type, labels):
if model_type == "sparse":
from classifiers.sparse_perceptron import SparsePerceptron
from features.sparse_features import SparseFeatureExtractor
features = SparseFeatureExtractor()
model = SparsePerceptron(labels, min_update=Config().min_update)
elif model_type == "dense":
from features.embedding import FeatureEmbedding
from classifiers.dense_perceptron import DensePerceptron
features = dense_features_wrapper(FeatureEmbedding)
model = DensePerceptron(labels, num_features=features.num_features())
elif model_type == "nn":
from features.indexer import FeatureIndexer
from classifiers.neural_network import NeuralNetwork
features = dense_features_wrapper(FeatureIndexer)
model = NeuralNetwork(labels, inputs=features.feature_types)
else:
raise ValueError("Invalid model type: '%s'" % model_type)
return features, model
