add ner model

sequence_tagging_for_ner/conll03.py

+# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""
+Conll03 dataset.
+"""
+
+import tarfile
+import gzip
+import itertools
+import re
+import numpy as np
+
+__all__ = ['train', 'test', 'get_dict', 'get_embedding']
+
+UNK_IDX = 0
+
+
+def canonicalize_digits(word):
+    if any([c.isalpha() for c in word]): return word
+    word = re.sub("\d", "DG", word)
+    if word.startswith("DG"):
+        word = word.replace(",", "")  # remove thousands separator
+    return word
+
+
+def canonicalize_word(word, wordset=None, digits=True):
+    word = word.lower()
+    if digits:
+        if (wordset != None) and (word in wordset): return word
+        word = canonicalize_digits(word)  # try to canonicalize numbers
+    if (wordset == None) or (word in wordset): return word
+    else: return "UUUNKKK"  # unknown token
+
+
+def load_dict(filename):
+    d = dict()
+    with open(filename, 'r') as f:
+        for i, line in enumerate(f):
+            d[line.strip()] = i
+    return d
+
+
+def get_dict(vocab_file='data/vocab.txt', target_file='data/target.txt'):
+    """
+    Get the word and label dictionary.
+    """
+    word_dict = load_dict(vocab_file)
+    label_dict = load_dict(target_file)
+    return word_dict, label_dict
+
+
+def get_embedding(emb_file='data/wordVectors.txt'):
+    """
+    Get the trained word vector.
+    """
+    return np.loadtxt(emb_file, dtype=float)
+
+
+def corpus_reader(filename='data/train'):
+    def reader():
+        sentence = []
+        labels = []
+        with open(filename) as f:
+            for line in f:
+                if re.match(r"-DOCSTART-.+", line) or (len(line.strip()) == 0):
+                    if len(sentence) > 0:
+                        yield sentence, labels
+                    sentence = []
+                    labels = []
+                else:
+                    segs = line.strip().split()
+                    sentence.append(segs[0])
+                    labels.append(segs[-1])
+
+        f.close()
+
+    return reader
+
+
+def reader_creator(corpus_reader=corpus_reader('data/train'),
+                   word_dict=load_dict('data/vocab.txt'),
+                   label_dict=load_dict('data/target.txt')):
+    """
+    Conll03 train set creator.
+
+    Because the training dataset is not free, the test dataset is used for
+    training. It returns a reader creator, each sample in the reader is nine
+    features, including sentence sequence, predicate, predicate context,
+    predicate context flag and tagged sequence.
+
+    :return: Training reader creator
+    :rtype: callable
+    """
+
+    def reader():
+        for sentence, labels in corpus_reader():
+            #word_idx = [word_dict.get(w, UNK_IDX) for w in sentence]
+            word_idx = [
+                word_dict.get(canonicalize_word(w, word_dict), UNK_IDX)
+                for w in sentence
+            ]
+            label_idx = [label_dict.get(w) for w in labels]
+            yield word_idx, label_idx
+
+    return reader
+
+
+def train():
+    return reader_creator(
+        corpus_reader('data/train'),
+        word_dict=load_dict('data/vocab.txt'),
+        label_dict=load_dict('data/target.txt'))
+
+
+def test():
+    return reader_creator(
+        corpus_reader('data/test'),
+        word_dict=load_dict('data/vocab.txt'),
+        label_dict=load_dict('data/target.txt'))

sequence_tagging_for_ner/data/target.txt

+B-LOC
+B-MISC
+B-ORG
+B-PER
+I-LOC
+I-MISC
+I-ORG
+I-PER
+O

sequence_tagging_for_ner/ner_final.py

+import math
+import gzip
+import paddle.v2 as paddle
+import paddle.v2.evaluator as evaluator
+import conll03
+import itertools
+
+word_dict, label_dict = conll03.get_dict()
+word_dict_len = len(word_dict)
+label_dict_len = len(label_dict)
+
+word_dim = 50
+caps_dim = 5
+context_length = 5
+hidden_dim = 300
+
+mix_hidden_lr = 1e-3
+default_std = 1 / math.sqrt(hidden_dim) / 3.0
+emb_para = paddle.attr.Param(
+    name='emb', initial_std=math.sqrt(1. / word_dim), is_static=True)
+std_0 = paddle.attr.Param(initial_std=0.)
+std_default = paddle.attr.Param(initial_std=default_std)
+
+
+def d_type(size):
+    return paddle.data_type.integer_value_sequence(size)
+
+
+def ner_net():
+    word = paddle.layer.data(name='word', type=d_type(word_dict_len))
+    #ws = paddle.layer.data(name='ws', type=d_type(num_ws))
+
+    word_embedding = paddle.layer.mixed(
+        name='word_embedding',
+        size=word_dim,
+        input=paddle.layer.table_projection(input=word, param_attr=emb_para))
+    #ws_embedding = paddle.layer.mixed(name='ws_embedding', size=caps_dim, 
+    #                    input=paddle.layer.table_projection(input=ws))
+    emb_layers = [word_embedding]  #[word_embedding, ws_embedding]
+
+    word_caps_vector = paddle.layer.concat(
+        name='word_caps_vector', input=emb_layers)
+    hidden_1 = paddle.layer.mixed(
+        name='hidden1',
+        size=hidden_dim,
+        bias_attr=std_default,
+        input=[
+            paddle.layer.full_matrix_projection(
+                input=word_caps_vector, param_attr=std_default)
+        ])
+
+    lstm_para_attr = paddle.attr.Param(initial_std=0.0, learning_rate=0.1)
+    hidden_para_attr = paddle.attr.Param(
+        initial_std=default_std, learning_rate=mix_hidden_lr)
+
+    lstm_1_1 = paddle.layer.lstmemory(
+        name='rnn1-1',
+        input=hidden_1,
+        act=paddle.activation.Relu(),
+        gate_act=paddle.activation.Sigmoid(),
+        state_act=paddle.activation.Sigmoid(),
+        bias_attr=std_0,
+        param_attr=lstm_para_attr)
+    lstm_1_2 = paddle.layer.lstmemory(
+        name='rnn1-2',
+        input=hidden_1,
+        act=paddle.activation.Relu(),
+        gate_act=paddle.activation.Sigmoid(),
+        state_act=paddle.activation.Sigmoid(),
+        reverse=1,
+        bias_attr=std_0,
+        param_attr=lstm_para_attr)
+
+    hidden_2_1 = paddle.layer.mixed(
+        size=hidden_dim,
+        bias_attr=std_default,
+        input=[
+            paddle.layer.full_matrix_projection(
+                input=hidden_1, param_attr=hidden_para_attr),
+            paddle.layer.full_matrix_projection(
+                input=lstm_1_1, param_attr=lstm_para_attr)
+        ])
+    hidden_2_2 = paddle.layer.mixed(
+        size=hidden_dim,
+        bias_attr=std_default,
+        input=[
+            paddle.layer.full_matrix_projection(
+                input=hidden_1, param_attr=hidden_para_attr),
+            paddle.layer.full_matrix_projection(
+                input=lstm_1_2, param_attr=lstm_para_attr)
+        ])
+
+    lstm_2_1 = paddle.layer.lstmemory(
+        name='rnn2-1',
+        input=hidden_2_1,
+        act=paddle.activation.Relu(),
+        gate_act=paddle.activation.Sigmoid(),
+        state_act=paddle.activation.Sigmoid(),
+        reverse=1,
+        bias_attr=std_0,
+        param_attr=lstm_para_attr)
+    lstm_2_2 = paddle.layer.lstmemory(
+        name='rnn2-2',
+        input=hidden_2_2,
+        act=paddle.activation.Relu(),
+        gate_act=paddle.activation.Sigmoid(),
+        state_act=paddle.activation.Sigmoid(),
+        bias_attr=std_0,
+        param_attr=lstm_para_attr)
+
+    hidden_3 = paddle.layer.mixed(
+        name='hidden3',
+        size=hidden_dim,
+        bias_attr=std_default,
+        input=[
+            paddle.layer.full_matrix_projection(
+                input=hidden_2_1, param_attr=hidden_para_attr),
+            paddle.layer.full_matrix_projection(
+                input=lstm_2_1,
+                param_attr=lstm_para_attr), paddle.layer.full_matrix_projection(
+                    input=hidden_2_2, param_attr=hidden_para_attr),
+            paddle.layer.full_matrix_projection(
+                input=lstm_2_2, param_attr=lstm_para_attr)
+        ])
+
+    output = paddle.layer.mixed(
+        name='output',
+        size=label_dict_len,
+        bias_attr=False,
+        input=[
+            paddle.layer.full_matrix_projection(
+                input=hidden_3, param_attr=std_default)
+        ])
+
+    target = paddle.layer.data(name='target', type=d_type(label_dict_len))
+
+    crf_cost = paddle.layer.crf(
+        size=label_dict_len,
+        input=output,
+        label=target,
+        param_attr=paddle.attr.Param(
+            name='crfw', initial_std=default_std, learning_rate=mix_hidden_lr))
+
+    predict = paddle.layer.crf_decoding(
+        size=label_dict_len,
+        input=output,
+        param_attr=paddle.attr.Param(name='crfw'))
+
+    return output, target, crf_cost, predict
+
+
+def ner_net_train(data_reader=conll03.train(), num_passes=1):
+    # define network topology
+    feature_out, target, crf_cost, predict = ner_net()
+    crf_dec = paddle.layer.crf_decoding(
+        size=label_dict_len,
+        input=feature_out,
+        label=target,
+        param_attr=paddle.attr.Param(name='crfw'))
+    evaluator.sum(input=crf_dec)
+
+    # create parameters
+    parameters = paddle.parameters.create(crf_cost)
+    parameters.set('emb', conll03.get_embedding())
+
+    # create optimizer
+    optimizer = paddle.optimizer.Momentum(
+        momentum=0,
+        learning_rate=2e-4,
+        regularization=paddle.optimizer.L2Regularization(rate=8e-4),
+        gradient_clipping_threshold=25,
+        model_average=paddle.optimizer.ModelAverage(
+            average_window=0.5, max_average_window=10000), )
+
+    trainer = paddle.trainer.SGD(
+        cost=crf_cost,
+        parameters=parameters,
+        update_equation=optimizer,
+        extra_layers=crf_dec)
+
+    reader = paddle.batch(
+        paddle.reader.shuffle(data_reader, buf_size=8192), batch_size=256)
+
+    feeding = {'word': 0, 'target': 1}
+
+    def event_handler(event):
+        if isinstance(event, paddle.event.EndIteration):
+            if event.batch_id % 100 == 0:
+                print "Pass %d, Batch %d, Cost %f, %s" % (
+                    event.pass_id, event.batch_id, event.cost, event.metrics)
+            if event.batch_id % 1000 == 0:
+                result = trainer.test(reader=reader, feeding=feeding)
+                print "\nTest with Pass %d, Batch %d, %s" % (
+                    event.pass_id, event.batch_id, result.metrics)
+
+        if isinstance(event, paddle.event.EndPass):
+            # save parameters
+            with gzip.open('params_pass_%d.tar.gz' % event.pass_id, 'w') as f:
+                parameters.to_tar(f)
+
+            result = trainer.test(reader=reader, feeding=feeding)
+            print "\nTest with Pass %d, %s" % (event.pass_id, result.metrics)
+
+    trainer.train(
+        reader=reader,
+        event_handler=event_handler,
+        num_passes=num_passes,
+        feeding=feeding)
+
+    return parameters
+
+
+def ner_net_infer(parameters=paddle.parameters.Parameters.from_tar(
+        gzip.open('ner_params_pass_99.tar.gz')),
+                  data_reader=conll03.test()):
+    test_creator = data_reader
+    test_data = []
+    for item in test_creator():
+        test_data.append([item[0]])
+        if len(test_data) == 10:
+            break
+
+    feature_out, target, crf_cost, predict = ner_net()
+
+    lab_ids = paddle.infer(
+        output_layer=predict,
+        parameters=parameters,
+        input=test_data,
+        field='id')
+
+    labels_reverse = {}
+    for (k, v) in label_dict.items():
+        labels_reverse[v] = k
+    pre_lab = [labels_reverse[lab_id] for lab_id in lab_ids]
+    print pre_lab
+
+
+if __name__ == '__main__':
+    paddle.init(use_gpu=False, trainer_count=1)
+    ner_net_train()
+    ner_net_infer()