Add xDeepFM & version check

* Add xDeepFM * Add auto version check * Fix some bugs when model's input field size is 1 * Encapsulate some duplicate code * Update docs
shenweichen · Dec 22, 2018 · aad52c4 · aad52c4
1 parent 0bab07a
commit aad52c4
Show file tree

Hide file tree

Showing 44 changed files with 1,142 additions and 1,821 deletions.
diff --git a/.travis.yml b/.travis.yml
@@ -36,7 +36,6 @@ install:
     - pip install -q pytest-cov==2.4.0 #>=2.4.0,<2.6
     - pip install -q python-coveralls
     - pip install -q codacy-coverage
-    - pip install -q h5py
     - pip install -q tensorflow==$TF_VERSION
     - pip install -e .
 # command to run tests

diff --git a/README.md b/README.md
@@ -14,7 +14,7 @@
 [![Codacy Badge](https://api.codacy.com/project/badge/Grade/d4099734dc0e4bab91d332ead8c0bdd0)](https://www.codacy.com/app/wcshen1994/DeepCTR?utm_source=github.com&amp;utm_medium=referral&amp;utm_content=shenweichen/DeepCTR&amp;utm_campaign=Badge_Grade)
 [![License](https://img.shields.io/github/license/shenweichen/deepctr.svg)](https://github.com/shenweichen/deepctr/blob/master/LICENSE)
 
-DeepCTR is a **Easy-to-use**,**Modular** and **Extendible** package of deep-learning based CTR models along with lots of core components layer  which can be used to build your own custom model easily.You can use any complex model with `model.fit()`and`model.predict()` just like any other keras model.And the layers are compatible with tensorflow.
+DeepCTR is a **Easy-to-use**,**Modular** and **Extendible** package of deep-learning based CTR models along with lots of core components layer  which can be used to build your own custom model easily.You can use any complex model with `model.fit()`and`model.predict()` .And the layers are compatible with tensorflow.
 
 Through  `pip install deepctr`  get the package and [**Get Started!**](https://deepctr-doc.readthedocs.io/en/latest/Quick-Start.html)
 
@@ -32,3 +32,4 @@ Through  `pip install deepctr`  get the package and [**Get Started!**](https://d
 |Attentional Factorization Machine|[IJCAI 2017][Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networks](http://www.ijcai.org/proceedings/2017/435)|
 |Neural Factorization Machine|[SIGIR 2017][Neural Factorization Machines for Sparse Predictive Analytics](https://arxiv.org/pdf/1708.05027.pdf)|
 |Deep Interest Network|[KDD 2018][Deep Interest Network for Click-Through Rate Prediction](https://arxiv.org/pdf/1706.06978.pdf)|
+|xDeepFM|[KDD 2018][xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems](https://arxiv.org/pdf/1803.05170.pdf)|
diff --git a/deepctr/__init__.py b/deepctr/__init__.py
@@ -1,4 +1,7 @@
 from .import activations
 from .import layers
 from .import sequence
-from .import models
+from . import models
+from .utils import check_version
+__version__ = '0.2.0a1'
+check_version(__version__)
diff --git a/deepctr/activations.py b/deepctr/activations.py
@@ -46,3 +46,15 @@ def get_config(self,):
 
     def compute_output_shape(self, input_shape):
         return input_shape
+
+
+def activation_fun(activation, fc):
+
+    if isinstance(activation, str):
+        fc = tf.keras.layers.Activation(activation)(fc)
+    elif issubclass(activation, Layer):
+        fc = activation()(fc)
+    else:
+        raise ValueError(
+            "Invalid activation,found %s.You should use a str or a Activation Layer Class." % (activation))
+    return fc
diff --git a/deepctr/layers.py b/deepctr/layers.py
diff --git a/deepctr/models/__init__.py b/deepctr/models/__init__.py
@@ -7,6 +7,7 @@
 from .fnn import FNN
 from .pnn import PNN
 from .wdl import WDL
+from .xdeepfm import xDeepFM
 
 __all__ = ["AFM", "DCN", "MLR",  "DeepFM",
-           "MLR", "NFM", "DIN", "FNN", "PNN", "WDL"]
+           "MLR", "NFM", "DIN", "FNN", "PNN", "WDL", "xDeepFM"]
diff --git a/deepctr/models/afm.py b/deepctr/models/afm.py
@@ -10,12 +10,11 @@
 
 """
 
-from tensorflow.python.keras.layers import Dense, Embedding, Concatenate, Reshape, add
+from tensorflow.python.keras.layers import Dense, Concatenate, Reshape, add
 from tensorflow.python.keras.models import Model
-from tensorflow.python.keras.initializers import RandomNormal
 from tensorflow.python.keras.regularizers import l2
 
-from ..utils import get_input
+from ..utils import get_input, get_share_embeddings
 from ..layers import PredictionLayer, AFMLayer, FM
 
 
@@ -50,7 +49,7 @@ def AFM(feature_dim_dict, embedding_size=8, use_attention=True, attention_factor
             feature_dim_dict['dense']))
 
     sparse_input, dense_input = get_input(feature_dim_dict, None)
-    sparse_embedding, linear_embedding, = get_embeddings(
+    sparse_embedding, linear_embedding, = get_share_embeddings(
         feature_dim_dict, embedding_size, init_std, seed, l2_reg_embedding, l2_reg_linear)
 
     embed_list = [sparse_embedding[i](sparse_input[i])
@@ -87,19 +86,3 @@ def AFM(feature_dim_dict, embedding_size=8, use_attention=True, attention_factor
     output = PredictionLayer(final_activation)(final_logit)
     model = Model(inputs=sparse_input + dense_input, outputs=output)
     return model
-
-
-def get_embeddings(feature_dim_dict, embedding_size, init_std, seed, l2_rev_V, l2_reg_w):
-    sparse_embedding = [Embedding(feature_dim_dict["sparse"][feat], embedding_size,
-                                  embeddings_initializer=RandomNormal(
-                                      mean=0.0, stddev=init_std, seed=seed),
-                                  embeddings_regularizer=l2(l2_rev_V),
-                                  name='sparse_emb_' + str(i) + '-' + feat) for i, feat in
-                        enumerate(feature_dim_dict["sparse"])]
-    linear_embedding = [Embedding(feature_dim_dict["sparse"][feat], 1,
-                                  embeddings_initializer=RandomNormal(mean=0.0, stddev=init_std,
-                                                                      seed=seed), embeddings_regularizer=l2(l2_reg_w),
-                                  name='linear_emb_' + str(i) + '-' + feat) for
-                        i, feat in enumerate(feature_dim_dict["sparse"])]
-
-    return sparse_embedding, linear_embedding
diff --git a/deepctr/models/dcn.py b/deepctr/models/dcn.py
@@ -4,7 +4,7 @@
     Weichen Shen,wcshen1994@163.com
 
 Reference:
-    [1] Deep & Cross Network for Ad Click Predictions (https://arxiv.org/abs/1708.05123)
+    [1] Wang R, Fu B, Fu G, et al. Deep & cross network for ad click predictions[C]//Proceedings of the ADKDD'17. ACM, 2017: 12. (https://arxiv.org/abs/1708.05123)
 """
 from tensorflow.python.keras.layers import Dense, Embedding, Concatenate, Flatten
 from tensorflow.python.keras.models import Model

diff --git a/deepctr/models/deepfm.py b/deepctr/models/deepfm.py
@@ -4,17 +4,14 @@
     Weichen Shen,wcshen1994@163.com
 
 Reference:
-    [1] DeepFM: A Factorization-Machine based Neural Network for CTR Prediction(https://arxiv.org/abs/1703.04247)
+    [1] Guo H, Tang R, Ye Y, et al. Deepfm: a factorization-machine based neural network for ctr prediction[J]. arXiv preprint arXiv:1703.04247, 2017.(https://arxiv.org/abs/1703.04247)
 
 """
 
-from tensorflow.python.keras.layers import Dense, Embedding, Concatenate, Reshape, Flatten, add
+from tensorflow.python.keras.layers import Dense, Concatenate, Reshape, Flatten, add
 from tensorflow.python.keras.models import Model
-from tensorflow.python.keras.initializers import RandomNormal
 from tensorflow.python.keras.regularizers import l2
-
-
-from ..utils import get_input
+from ..utils import get_input, get_share_embeddings
 from ..layers import PredictionLayer, MLP, FM
 
 
@@ -50,7 +47,7 @@ def DeepFM(feature_dim_dict, embedding_size=8,
             feature_dim_dict['dense']))
 
     sparse_input, dense_input = get_input(feature_dim_dict, None)
-    sparse_embedding, linear_embedding, = get_embeddings(
+    sparse_embedding, linear_embedding, = get_share_embeddings(
         feature_dim_dict, embedding_size, init_std, seed, l2_reg_embedding, l2_reg_linear)
 
     embed_list = [sparse_embedding[i](sparse_input[i])
@@ -97,19 +94,3 @@ def DeepFM(feature_dim_dict, embedding_size=8,
     output = PredictionLayer(final_activation)(final_logit)
     model = Model(inputs=sparse_input + dense_input, outputs=output)
     return model
-
-
-def get_embeddings(feature_dim_dict, embedding_size, init_std, seed, l2_rev_V, l2_reg_w):
-    sparse_embedding = [Embedding(feature_dim_dict["sparse"][feat], embedding_size,
-                                  embeddings_initializer=RandomNormal(
-                                      mean=0.0, stddev=init_std, seed=seed),
-                                  embeddings_regularizer=l2(l2_rev_V),
-                                  name='sparse_emb_' + str(i) + '-' + feat) for i, feat in
-                        enumerate(feature_dim_dict["sparse"])]
-    linear_embedding = [Embedding(feature_dim_dict["sparse"][feat], 1,
-                                  embeddings_initializer=RandomNormal(mean=0.0, stddev=init_std,
-                                                                      seed=seed), embeddings_regularizer=l2(l2_reg_w),
-                                  name='linear_emb_' + str(i) + '-' + feat) for
-                        i, feat in enumerate(feature_dim_dict["sparse"])]
-
-    return sparse_embedding, linear_embedding
diff --git a/deepctr/models/din.py b/deepctr/models/din.py
@@ -4,7 +4,7 @@
     Weichen Shen,wcshen1994@163.com
 
 Reference:
-    [1] Deep Interest Network for Click-Through Rate Prediction (https://arxiv.org/pdf/1706.06978.pdf)
+    [1] Zhou G, Zhu X, Song C, et al. Deep interest network for click-through rate prediction[C]//Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. ACM, 2018: 1059-1068. (https://arxiv.org/pdf/1706.06978.pdf)
 """
 
 from tensorflow.python.keras.layers import Input, Dense, Embedding, Concatenate, Reshape

diff --git a/deepctr/models/fnn.py b/deepctr/models/fnn.py
@@ -4,16 +4,15 @@
     Weichen Shen,wcshen1994@163.com
 
 Reference:
-    [1] Zhang, Weinan, Tianming Du, and Jun Wang. "Deep learning over multi-field categorical data." European conference on information retrieval. Springer, Cham, 2016.(https://arxiv.org/pdf/1601.02376.pdf)
+    [1] Zhang W, Du T, Wang J. Deep learning over multi-field categorical data[C]//European conference on information retrieval. Springer, Cham, 2016: 45-57.(https://arxiv.org/pdf/1601.02376.pdf)
 """
 
-from tensorflow.python.keras.layers import Dense, Embedding, Concatenate, Reshape, add
+from tensorflow.python.keras.layers import Dense, Concatenate, Reshape, add
 from tensorflow.python.keras.models import Model
-from tensorflow.python.keras.initializers import RandomNormal
 from tensorflow.python.keras.regularizers import l2
 
 from ..layers import PredictionLayer, MLP
-from ..utils import get_input
+from ..utils import get_input, get_share_embeddings
 
 
 def FNN(feature_dim_dict, embedding_size=8,
@@ -42,12 +41,8 @@ def FNN(feature_dim_dict, embedding_size=8,
             "feature_dim must be a dict like {'sparse':{'field_1':4,'field_2':3,'field_3':2},'dense':['field_5',]}")
 
     sparse_input, dense_input = get_input(feature_dim_dict, None)
-    # sparse_embedding = [Embedding(feature_dim_dict["sparse"][feat], embedding_size,
-    #                              embeddings_initializer=RandomNormal( mean=0.0, stddev=init_std, seed=seed),
-    #   embeddings_regularizer=l2( l2_reg_embedding),name='sparse_emb_' + str(i) + '-' + feat) for i, feat in
-    #   enumerate(feature_dim_dict["sparse"])]
-    sparse_embedding, linear_embedding, = get_embeddings(feature_dim_dict, embedding_size, init_std, seed, l2_reg_embedding,
-                                                         l2_reg_linear)
+    sparse_embedding, linear_embedding, = get_share_embeddings(feature_dim_dict, embedding_size, init_std, seed, l2_reg_embedding,
+                                                               l2_reg_linear)
 
     embed_list = [sparse_embedding[i](sparse_input[i])
                   for i in range(len(feature_dim_dict["sparse"]))]
@@ -56,10 +51,11 @@ def FNN(feature_dim_dict, embedding_size=8,
                    for i in range(len(sparse_input))]
     if len(linear_term) > 1:
         linear_term = add(linear_term)
-    elif len(linear_term) > 0:
+    elif len(linear_term) == 1:
         linear_term = linear_term[0]
+    else:
+        linear_term = 0
 
-    #linear_term = add([linear_embedding[i](sparse_input[i]) for i in range(len(feature_dim_dict["sparse"]))])
     if len(dense_input) > 0:
         continuous_embedding_list = list(
             map(Dense(embedding_size, use_bias=False, kernel_regularizer=l2(l2_reg_embedding), ),
@@ -85,19 +81,3 @@ def FNN(feature_dim_dict, embedding_size=8,
     model = Model(inputs=sparse_input + dense_input,
                   outputs=output)
     return model
-
-
-def get_embeddings(feature_dim_dict, embedding_size, init_std, seed, l2_rev_V, l2_reg_w):
-    sparse_embedding = [Embedding(feature_dim_dict["sparse"][feat], embedding_size,
-                                  embeddings_initializer=RandomNormal(
-                                      mean=0.0, stddev=init_std, seed=seed),
-                                  embeddings_regularizer=l2(l2_rev_V),
-                                  name='sparse_emb_' + str(i) + '-' + feat) for i, feat in
-                        enumerate(feature_dim_dict["sparse"])]
-    linear_embedding = [Embedding(feature_dim_dict["sparse"][feat], 1,
-                                  embeddings_initializer=RandomNormal(mean=0.0, stddev=init_std,
-                                                                      seed=seed), embeddings_regularizer=l2(l2_reg_w),
-                                  name='linear_emb_' + str(i) + '-' + feat) for
-                        i, feat in enumerate(feature_dim_dict["sparse"])]
-
-    return sparse_embedding, linear_embedding