From d9b2e7d09ec2d74f1c2a689e45838c95008006a2 Mon Sep 17 00:00:00 2001
From: Yu Yang <yuyang18@baidu.com>
Date: Wed, 1 Mar 2017 17:27:31 +0800
Subject: [PATCH] Complete recommendation demo in API.v2

---
 demo/recommendation/api_train_v2.py   | 69 ++++++++++++++++++++++++---
 python/paddle/v2/dataset/movielens.py | 19 +++++++-
 2 files changed, 80 insertions(+), 8 deletions(-)

diff --git a/demo/recommendation/api_train_v2.py b/demo/recommendation/api_train_v2.py
index c726fa5bd..3ff86c2c4 100644
--- a/demo/recommendation/api_train_v2.py
+++ b/demo/recommendation/api_train_v2.py
@@ -2,6 +2,7 @@ import paddle.v2 as paddle
 
 
 def main():
+    paddle.init(use_gpu=False, trainer_count=3)
     movie_title_dict = paddle.dataset.movielens.get_movie_title_dict()
     uid = paddle.layer.data(
         name='user_id',
@@ -17,10 +18,17 @@ def main():
         name='age_id',
         type=paddle.data_type.integer_value(
             len(paddle.dataset.movielens.age_table)))
-    usr_age_emb = paddle.embedding(input=usr_age_id, size=16)
+    usr_age_emb = paddle.layer.embedding(input=usr_age_id, size=16)
 
-    usr_combined_features = paddle.fc(
-        input=[usr_emb, usr_gender_emb, usr_age_emb],
+    usr_job_id = paddle.layer.data(
+        name='job_id',
+        type=paddle.data_type.integer_value(paddle.dataset.movielens.max_job_id(
+        ) + 1))
+
+    usr_job_emb = paddle.layer.embedding(input=usr_job_id, size=16)
+
+    usr_combined_features = paddle.layer.fc(
+        input=[usr_emb, usr_gender_emb, usr_age_emb, usr_job_emb],
         size=200,
         act=paddle.activation.Tanh())
 
@@ -30,12 +38,59 @@ def main():
             paddle.dataset.movielens.max_movie_id() + 1))
     mov_emb = paddle.layer.embedding(input=mov_id, size=32)
 
+    mov_categories = paddle.layer.data(
+        name='category_id',
+        type=paddle.data_type.sparse_binary_vector(
+            len(paddle.dataset.movielens.movie_categories())))
+
+    mov_categories_hidden = paddle.layer.fc(input=mov_categories, size=32)
+
     mov_title_id = paddle.layer.data(
         name='movie_title',
-        type=paddle.data_type.integer_value(len(movie_title_dict)))
-    mov_title_emb = paddle.embedding(input=mov_title_id, size=32)
-    with paddle.layer.mixed() as mixed:
-        pass
+        type=paddle.data_type.integer_value_sequence(len(movie_title_dict)))
+    mov_title_emb = paddle.layer.embedding(input=mov_title_id, size=32)
+    mov_title_conv = paddle.networks.sequence_conv_pool(
+        input=mov_title_emb, hidden_size=32, context_len=3)
+
+    mov_combined_features = paddle.layer.fc(
+        input=[mov_emb, mov_categories_hidden, mov_title_conv],
+        size=200,
+        act=paddle.activation.Tanh())
+
+    inference = paddle.layer.cos_sim(
+        a=usr_combined_features, b=mov_combined_features, size=1, scale=5)
+    cost = paddle.layer.regression_cost(
+        input=inference,
+        label=paddle.layer.data(
+            name='score', type=paddle.data_type.dense_vector(1)))
+
+    parameters = paddle.parameters.create(cost)
+
+    trainer = paddle.trainer.SGD(cost=cost,
+                                 parameters=parameters,
+                                 update_equation=paddle.optimizer.Adam(
+                                     learning_rate=1e-4))
+
+    def event_handler(event):
+        if isinstance(event, paddle.event.EndIteration):
+            if event.batch_id % 100 == 0:
+                print "Pass %d Batch %d Cost %.2f" % (
+                    event.pass_id, event.batch_id, event.cost)
+
+    trainer.train(
+        reader=paddle.reader.batched(
+            paddle.dataset.movielens.train(), batch_size=256),
+        event_handler=event_handler,
+        reader_dict={
+            'user_id': 0,
+            'gender_id': 1,
+            'age_id': 2,
+            'job_id': 3,
+            'movie_id': 4,
+            'category_id': 5,
+            'movie_title': 6,
+            'score': 7
+        })
 
 
 if __name__ == '__main__':
diff --git a/python/paddle/v2/dataset/movielens.py b/python/paddle/v2/dataset/movielens.py
index 6efe42adb..058400502 100644
--- a/python/paddle/v2/dataset/movielens.py
+++ b/python/paddle/v2/dataset/movielens.py
@@ -6,7 +6,7 @@ import functools
 
 __all__ = [
     'train', 'test', 'get_movie_title_dict', 'max_movie_id', 'max_user_id',
-    'age_table'
+    'age_table', 'movie_categories', 'max_job_id'
 ]
 
 age_table = [1, 18, 25, 35, 45, 50, 56]
@@ -135,6 +135,23 @@ def max_user_id():
     return reduce(__max_index_info__, USER_INFO.viewvalues()).index
 
 
+def __max_job_id_impl__(a, b):
+    if a.job_id > b.job_id:
+        return a
+    else:
+        return b
+
+
+def max_job_id():
+    __initialize_meta_info__()
+    return reduce(__max_job_id_impl__, USER_INFO.viewvalues()).job_id
+
+
+def movie_categories():
+    __initialize_meta_info__()
+    return CATEGORIES_DICT
+
+
 def unittest():
     for train_count, _ in enumerate(train()()):
         pass
-- 
GitLab