def model(label_vocab, embeds): import baseline.pytorch.tagger.model return create_tagger_model( embeds, label_vocab, crf=True, crf_mask=True, span_type=SPAN_TYPE, hsz=HSZ, cfiltsz=[3], wsz=WSZ, layers=2, rnntype="blstm" )
def model(label_vocab, embeds, mask): from baseline.tf import tagger model = create_tagger_model( embeds, label_vocab, crf=True, constraint=mask, hsz=HSZ, cfiltsz=[3], wsz=WSZ, layers=2, rnntype="blstm" ) model.create_loss() model.sess.run(tf.global_variables_initializer()) return model
def test_skip_mask(label_vocab, embeds, mask): from baseline.tf import tagger model = create_tagger_model( embeds, label_vocab, crf=True, hsz=HSZ, cfiltsz=[3], wsz=WSZ, layers=2, rnntype="blstm" ) model.create_loss() model.sess.run(tf.global_variables_initializer()) transition = model.sess.run(model.A) assert transition[label_vocab['O'], label_vocab[S]] != -1e4
def model(label_vocab, embeds, mask): from baseline.tf import tagger model = create_tagger_model(embeds, label_vocab, crf=True, constraint=mask, hsz=HSZ, cfiltsz=[3], wsz=WSZ, layers=2, rnntype="blstm") model.create_loss() model.sess.run(tf.global_variables_initializer()) return model
def test_skip_mask(label_vocab, embeds, mask): from baseline.tf import tagger model = create_tagger_model(embeds, label_vocab, crf=True, hsz=HSZ, cfiltsz=[3], wsz=WSZ, layers=2, rnntype="blstm") model.create_loss() model.sess.run(tf.global_variables_initializer()) transition = model.sess.run(model.A) assert transition[label_vocab['O'], label_vocab[S]] != -1e4
def create_model(labels, embeddings, **kwargs): return create_tagger_model(BASELINE_TAGGER_MODELS, labels, embeddings, **kwargs)
def create_model(labels, word_embedding, char_embedding, **kwargs): model = create_tagger_model(RNNTaggerModel.create, labels, word_embedding, char_embedding, **kwargs) return model
def create_model(labels, embeddings, **kwargs): model = create_tagger_model(RNNTaggerModel.create, labels, embeddings, **kwargs) return model