modify CNN model

moneyDboat · moneyDboat · commit dca0bed78c2f · 2017-04-11T23:16:46.000+08:00
diff --git a/models.py b/models.py
@@ -53,8 +53,10 @@ def ConvMF(res_dir, train_user, train_item, valid_user, test_user,
     # return the theta of CNN
     theta = cnn_module.get_projection_layer(CNN_X)
     np.random.seed(133)
+    # dimension is the k
     U = np.random.uniform(size=(num_user, dimension))
     V = theta
+    print("Jay::theta over")
 
     endure_count = 5
     count = 0
diff --git a/run_test_ConvMF.sh b/run_test_ConvMF.sh
@@ -3,8 +3,8 @@ python ./run.py \
 -d ./data/preprocessed/ml-1m/0.2/ \
 -a ./data/preprocessed/ml-1m/ \
 -o ./test/ml-1m/result/1_100_200 \
--e 200 \
--p ./data/preprocessed/glove/glove.6B.200d.txt \
+-e 50 \
+-p ./data/preprocessed/glove/glove.6B.50d.txt \
 -u 10 \
 -v 100 \
 -g True
diff --git a/text_analysis/models.py b/text_analysis/models.py
@@ -38,8 +38,6 @@ def __init__(self, output_dimesion, vocab_size, dropout_rate, emb_dim, max_len,
         # self.qual_model = Model()
         self.qual_conv_set = {}
 
-        filter_lengths = [3, 4, 5]
-
         # my code
         '''Input'''
         input_data = Input(shape=(max_len, ), name='input', dtype='int32')
@@ -169,8 +167,8 @@ def train(self, X_train, V, item_weight, seed):
         item_weight = np.random.permutation(item_weight)
 
         print("Train...CNN module")
-        history = self.model.fit({'input': X_train, 'output': V},
-                                 verbose=0, batch_size=self.batch_size, nb_epoch=self.nb_epoch, sample_weight={'output': item_weight})
+        history = self.model.fit(X_train, V, verbose=0, batch_size=self.batch_size,
+                                 nb_epoch=self.nb_epoch, sample_weight=item_weight)
 
         # cnn_loss_his = history.history['loss']
         # cmp_cnn_loss = sorted(cnn_loss_his)[::-1]
@@ -180,6 +178,5 @@ def train(self, X_train, V, item_weight, seed):
 
     def get_projection_layer(self, X_train):
         X_train = sequence.pad_sequences(X_train, maxlen=self.max_len)
-        Y = self.model.predict(
-            {'input': X_train}, batch_size=len(X_train))['output']
+        Y = self.model.predict(X_train, batch_size=len(X_train))
         return Y
