v0.8.1 release (shenweichen#276)

- Improve the reproducibility 
- Fix some bugs

Author: 浅梦

.github/workflows/ci.yml

@@ -18,7 +18,7 @@ jobs:
     strategy:
       matrix:
         python-version: [3.6,3.7]
-        tf-version: [1.4.0,1.15.0,2.1.0,2.2.0]
+        tf-version: [1.4.0,1.15.0,2.1.0,2.3.0]
 
         exclude:
           - python-version: 3.7

README.md

@@ -53,12 +53,28 @@ Let's [**Get Started!**](https://deepctr-doc.readthedocs.io/en/latest/Quick-Star
 |                FiBiNET                 | [RecSys 2019][FiBiNET: Combining Feature Importance and Bilinear feature Interaction for Click-Through Rate Prediction](https://arxiv.org/pdf/1905.09433.pdf)   |
 |                FLEN                    | [arxiv 2019][FLEN: Leveraging Field for Scalable CTR Prediction](https://arxiv.org/pdf/1911.04690.pdf)   |
 
+## Citation
 
+- Weichen Shen. (2018). DeepCTR: Easy-to-use,Modular and Extendible package of deep-learning based CTR models. https://github.com/shenweichen/deepctr.
 
-## DisscussionGroup  
+
+If you find this code useful in your research, please cite it using the following BibTeX:
+
+```bibtex
+@misc{shen2018deepctr,
+  author = {Weichen Shen},
+  title = {DeepCTR: Easy-to-use,Modular and Extendible package of deep-learning based CTR models},
+  year = {2018},
+  publisher = {GitHub},
+  journal = {GitHub Repository},
+  howpublished = {\url{https://github.com/shenweichen/deepctr}},
+}
+```
+
+## DisscussionGroup 交流群
 
 Please follow our wechat to join group:  
 - 公众号：**浅梦的学习笔记**  
 - wechat ID: **deepctrbot**
 
-  ![wechat](./docs/pics/weichennote.png)
+  ![wechat](./docs/pics/code.png)

deepctr/__init__.py

@@ -1,4 +1,4 @@
 from .utils import check_version
 
-__version__ = '0.8.0'
+__version__ = '0.8.1'
 check_version(__version__)

deepctr/estimator/inputs.py

@@ -23,7 +23,10 @@ def input_fn_tfrecord(filenames, feature_description, label=None, batch_size=256
                       shuffle_factor=10, prefetch_factor=1,
                       ):
     def _parse_examples(serial_exmp):
-        features = tf.parse_single_example(serial_exmp, features=feature_description)
+        try:
+            features = tf.parse_single_example(serial_exmp, features=feature_description)
+        except AttributeError:
+            features = tf.io.parse_single_example(serial_exmp, features=feature_description)
         if label is not None:
             labels = features.pop(label)
             return features, labels
@@ -39,8 +42,10 @@ def input_fn():
 
         if prefetch_factor > 0:
             dataset = dataset.prefetch(buffer_size=batch_size * prefetch_factor)
-
-        iterator = dataset.make_one_shot_iterator()
+        try:
+            iterator = dataset.make_one_shot_iterator()
+        except AttributeError:
+            iterator = tf.compat.v1.data.make_one_shot_iterator(dataset)
 
         return iterator.get_next()
 

deepctr/estimator/models/autoint.py

@@ -77,15 +77,15 @@ def _model_fn(features, labels, mode, config):
                                dnn_use_bn, seed)(dnn_input, training=train_flag)
                 stack_out = tf.keras.layers.Concatenate()([att_output, deep_out])
                 final_logit = tf.keras.layers.Dense(
-                    1, use_bias=False, activation=None)(stack_out)
+                    1, use_bias=False, kernel_initializer=tf.keras.initializers.glorot_normal(seed))(stack_out)
             elif len(dnn_hidden_units) > 0:  # Only Deep
                 deep_out = DNN(dnn_hidden_units, dnn_activation, l2_reg_dnn, dnn_dropout,
                                dnn_use_bn, seed)(dnn_input, training=train_flag)
                 final_logit = tf.keras.layers.Dense(
-                    1, use_bias=False, activation=None)(deep_out)
+                    1, use_bias=False, kernel_initializer=tf.keras.initializers.glorot_normal(seed))(deep_out)
             elif att_layer_num > 0:  # Only Interacting Layer
                 final_logit = tf.keras.layers.Dense(
-                    1, use_bias=False, activation=None)(att_output)
+                    1, use_bias=False, kernel_initializer=tf.keras.initializers.glorot_normal(seed))(att_output)
             else:  # Error
                 raise NotImplementedError
 

deepctr/estimator/models/ccpm.py

@@ -82,7 +82,7 @@ def _model_fn(features, labels, mode, config):
             flatten_result = tf.keras.layers.Flatten()(pooling_result)
             dnn_out = DNN(dnn_hidden_units, l2_reg=l2_reg_dnn,
                           dropout_rate=dnn_dropout, seed=seed)(flatten_result, training=train_flag)
-            dnn_logit = tf.keras.layers.Dense(1, use_bias=False)(dnn_out)
+            dnn_logit = tf.keras.layers.Dense(1, use_bias=False, kernel_initializer=tf.keras.initializers.glorot_normal(seed))(dnn_out)
 
         logits = linear_logits + dnn_logit
 

deepctr/estimator/models/dcn.py

@@ -68,16 +68,16 @@ def _model_fn(features, labels, mode, config):
                 cross_out = CrossNet(cross_num, l2_reg=l2_reg_cross)(dnn_input)
                 stack_out = tf.keras.layers.Concatenate()([cross_out, deep_out])
                 final_logit = tf.keras.layers.Dense(
-                    1, use_bias=False, activation=None)(stack_out)
+                    1, use_bias=False, kernel_initializer=tf.keras.initializers.glorot_normal(seed))(stack_out)
             elif len(dnn_hidden_units) > 0:  # Only Deep
                 deep_out = DNN(dnn_hidden_units, dnn_activation, l2_reg_dnn, dnn_dropout,
                                dnn_use_bn, seed)(dnn_input, training=train_flag)
                 final_logit = tf.keras.layers.Dense(
-                    1, use_bias=False, activation=None)(deep_out)
+                    1, use_bias=False, kernel_initializer=tf.keras.initializers.glorot_normal(seed))(deep_out)
             elif cross_num > 0:  # Only Cross
                 cross_out = CrossNet(cross_num, l2_reg=l2_reg_cross)(dnn_input)
                 final_logit = tf.keras.layers.Dense(
-                    1, use_bias=False, activation=None)(cross_out)
+                    1, use_bias=False, kernel_initializer=tf.keras.initializers.glorot_normal(seed))(cross_out)
             else:  # Error
                 raise NotImplementedError
 

deepctr/estimator/models/deepfm.py

@@ -10,8 +10,6 @@
 
 import tensorflow as tf
 
-from tensorflow.python.keras.initializers import glorot_normal
-
 from ..feature_column import get_linear_logit, input_from_feature_columns
 from ..utils import deepctr_model_fn, DNN_SCOPE_NAME, variable_scope
 from ...layers.core import DNN
@@ -68,7 +66,7 @@ def _model_fn(features, labels, mode, config):
             dnn_output = DNN(dnn_hidden_units, dnn_activation, l2_reg_dnn, dnn_dropout,
                              dnn_use_bn, seed)(dnn_input, training=train_flag)
             dnn_logit = tf.keras.layers.Dense(
-                1, use_bias=False, activation=None, kernel_initializer=glorot_normal(seed=seed))(dnn_output)
+                1, use_bias=False, kernel_initializer=tf.keras.initializers.glorot_normal(seed=seed))(dnn_output)
 
         logits = linear_logits + fm_logit + dnn_logit
 

deepctr/estimator/models/fibinet.py

@@ -71,7 +71,7 @@ def _model_fn(features, labels, mode, config):
             dnn_out = DNN(dnn_hidden_units, dnn_activation, l2_reg_dnn, dnn_dropout,
                           False, seed)(dnn_input, training=train_flag)
             dnn_logit = Dense(
-                1, use_bias=False, activation=None)(dnn_out)
+                1, use_bias=False, kernel_initializer=tf.keras.initializers.glorot_normal(seed))(dnn_out)
 
         logits = linear_logits + dnn_logit
 

deepctr/estimator/models/fnn.py

@@ -56,7 +56,7 @@ def _model_fn(features, labels, mode, config):
             deep_out = DNN(dnn_hidden_units, dnn_activation, l2_reg_dnn,
                            dnn_dropout, False, seed)(dnn_input, training=train_flag)
             dnn_logit = tf.keras.layers.Dense(
-                1, use_bias=False, activation=None)(deep_out)
+                1, use_bias=False, kernel_initializer=tf.keras.initializers.glorot_normal(seed))(deep_out)
 
         logits = linear_logits + dnn_logit