add conv-knrm

examples/wikiqa/config/conv_knrm_wikiqa.config

@@ -0,0 +1,79 @@
+{
+ "net_name": "CONVKNRM",
+ "global":{
+ "model_type": "PY",
+ "weights_file": "examples/wikiqa/weights/knrm.wikiqa.weights",
+ "save_weights_iters": 10,
+ "num_iters": 100,
+ "display_interval": 100,
+ "test_weights_iters": 10,
+ "optimizer": "adam",
+ "learning_rate": 0.001
+ },
+ "inputs": {
+ "share": {
+ "text1_corpus": "./data/WikiQA/corpus_preprocessed.txt",
+ "text2_corpus": "./data/WikiQA/corpus_preprocessed.txt",
+ "use_dpool": false,
+ "embed_size": 300,
+ "embed_path": "./data/WikiQA/embed_glove_d300_norm",
+ "vocab_size": 18680,
+ "train_embed": false,
+ "target_mode": "ranking",
+ "text1_maxlen": 10,
+ "text2_maxlen": 40
+ },
+ "train": {
+ "input_type": "PairGenerator",
+ "phase": "TRAIN",
+ "use_iter": false,
+ "query_per_iter": 50,
+ "batch_per_iter": 5,
+ "batch_size": 16,
+ "relation_file": "./data/WikiQA/relation_train.txt"
+ },
+ "valid": {
+ "input_type": "ListGenerator",
+ "phase": "EVAL",
+ "batch_list": 10,
+ "relation_file": "./data/WikiQA/relation_valid.txt"
+ },
+ "test": {
+ "input_type": "ListGenerator", 
+ "phase": "EVAL",
+ "batch_list": 10,
+ "relation_file": "./data/WikiQA/relation_test.txt"
+ },
+ "predict": {
+ "input_type": "ListGenerator",
+ "phase": "PREDICT",
+ "batch_list": 10,
+ "relation_file": "./data/WikiQA/relation_test.txt"
+ }
+ },
+ "outputs": {
+ "predict": {
+ "save_format": "TREC",
+ "save_path": "predict.test.knrm_ranking.txt"
+ }
+ },
+ "model": {
+ "model_path": "./matchzoo/models/",
+ "model_py": "conv_knrm.CONVKNRM",
+ "setting": {
+ "kernel_num": 11,
+ "sigma": 0.1,
+ "exact_sigma": 0.001,
+ "max_ngram": 3,
+ "if_crossmatch": true
+ }
+ },
+ "losses": [ 
+ {
+ "object_name": "rank_hinge_loss",
+ "object_params": { "margin": 1.0 }
+ }
+ ],
+ "metrics": [ "ndcg@3", "ndcg@5", "map" ]
+}
+

examples/wikiqa/run_conv_knrm.sh

@@ -0,0 +1,10 @@
+cd ../../
+
+currpath=`pwd`
+# train the model
+python matchzoo/main.py --phase train --model_file ${currpath}/examples/wikiqa/config/conv_knrm_wikiqa.config
+
+
+# predict with the model
+
+python matchzoo/main.py --phase predict --model_file ${currpath}/examples/wikiqa/config/conv_knrm_wikiqa.config

matchzoo/models/conv_knrm.py

@@ -0,0 +1,100 @@
+# -*- coding=utf-8 -*-
+import keras
+import keras.backend as K
+from keras.models import Sequential, Model
+from keras.layers import *
+from keras.layers import Input, Embedding, Dense, Activation, Lambda, Dot
+from keras.activations import softmax
+from keras.initializers import Constant, RandomNormal
+from model import BasicModel
+from utils.utility import *
+
+class CONVKNRM(BasicModel):
+ def __init__(self, config):
+ super(CONVKNRM, self).__init__(config)
+ self._name = 'CONVKNRM'
+ self.check_list = [ 'text1_maxlen', 'kernel_num','sigma','exact_sigma',
+ 'embed', 'embed_size', 'vocab_size', 'max_ngram', 'if_crossmatch']
+ self.setup(config)
+ if not self.check():
+ raise TypeError('[ConvKNRM] parameter check wrong')
+ print('[CONVKNRM] init done')
+
+ def setup(self, config):
+ self.set_default('kernel_num', 11)
+ self.set_default('sigma', 0.1)
+ self.set_default('exact_sigma', 0.001)
+ self.set_default('max_ngram', 3)
+ self.set_default('if_crossmatch', True)
+ if not isinstance(config, dict):
+ raise TypeError('parameter config should be dict:', config)
+ self.config.update(config)
+
+ def build(self):
+ def Kernel_layer(mu,sigma):
+ def kernel(x):
+ return K.tf.exp(-0.5 * (x - mu) * (x - mu) / sigma / sigma)
+ return Activation(kernel)
+
+ query = Input(name='query', shape=(self.config['text1_maxlen'],))
+ show_layer_info('Input', query)
+ doc = Input(name='doc', shape=(self.config['text2_maxlen'],))
+ show_layer_info('Input', doc)
+
+ embedding = Embedding(self.config['vocab_size'], self.config['embed_size'], weights=[self.config['embed']], trainable=self.config['train_embed'])
+
+ q_embed = embedding(query)
+ show_layer_info('Embedding', q_embed)
+ d_embed = embedding(doc)
+ show_layer_info('Embedding', d_embed)
+ convs = []
+ q_convs = []
+ d_convs = []
+ for i in range(self.config['max_ngram']):
+ c = keras.layers.Conv1D(128, i + 1, activation='relu', padding='same')
+ q_convs.append(c(q_embed) )
+ show_layer_info('Q N-gram Embedding', q_convs[i])
+ d_convs.append(c(d_embed) )
+ show_layer_info('D N-gram Embedding', d_convs[i])
+
+ KM = []
+ for qi in range(self.config['max_ngram']):
+ for di in range(self.config['max_ngram']):
+ # if not corssmatch, then do not match n-gram with different length
+ if not self.config['if_crossmatch'] and qi != di:
+ print ("non cross")
+ continue
+ q_ngram = q_convs[qi]
+ d_ngram = d_convs[di]
+ mm = Dot(axes=[2, 2], normalize=True)([q_ngram, d_ngram])
+ show_layer_info('Dot', mm)
+
+ for i in range(self.config['kernel_num']):
+ mu = 1. / (self.config['kernel_num'] - 1) + (2. * i) / (self.config['kernel_num'] - 1) - 1.0
+ sigma = self.config['sigma']
+ if mu > 1.0:
+ sigma = self.config['exact_sigma']
+ mu = 1.0
+ mm_exp = Kernel_layer(mu, sigma)(mm)
+ show_layer_info('Exponent of mm:', mm_exp)
+ mm_doc_sum = Lambda(lambda x: K.tf.reduce_sum(x,2))(mm_exp)
+ show_layer_info('Sum of document', mm_doc_sum)
+ mm_log = Activation(K.tf.log1p)(mm_doc_sum)
+ show_layer_info('Logarithm of sum', mm_log)
+ mm_sum = Lambda(lambda x: K.tf.reduce_sum(x, 1))(mm_log)
+ show_layer_info('Sum of all exponent', mm_sum)
+ KM.append(mm_sum)
+
+
+ Phi = Lambda(lambda x: K.tf.stack(x, 1))(KM)
+ show_layer_info('Stack', Phi)
+ if self.config['target_mode'] == 'classification':
+ out_ = Dense(2, activation='softmax', kernel_initializer=initializers.RandomUniform(minval=-0.014, maxval=0.014), bias_initializer='zeros')(Phi)
+ elif self.config['target_mode'] in ['regression', 'ranking']:
+ out_ = Dense(1, kernel_initializer=initializers.RandomUniform(minval=-0.014, maxval=0.014), bias_initializer='zeros')(Phi)
+ show_layer_info('Dense', out_)
+
+ model = Model(inputs=[query, doc], outputs=[out_])
+ return model
+
+