Merge branch 'develop' of https://github.com/PaddlePaddle/models into pr_04221554

PaddleRec/din/_ce.py → PaddleRec/ctr/din/_ce.py

@@ -6,18 +6,19 @@ sys.path.append(os.environ['ceroot'])
 from kpi import CostKpi
 from kpi import DurationKpi
 
-
-each_pass_duration_card1_kpi = DurationKpi('each_pass_duration_card1', 0.08, 0, actived=True)
+each_pass_duration_card1_kpi = DurationKpi(
+    'each_pass_duration_card1', 0.08, 0, actived=True)
 train_loss_card1_kpi = CostKpi('train_loss_card1', 0.08, 0)
-each_pass_duration_card4_kpi = DurationKpi('each_pass_duration_card4', 0.08, 0, actived=True)
+each_pass_duration_card4_kpi = DurationKpi(
+    'each_pass_duration_card4', 0.08, 0, actived=True)
 train_loss_card4_kpi = CostKpi('train_loss_card4', 0.08, 0)
 
 tracking_kpis = [
-        each_pass_duration_card1_kpi,
-        train_loss_card1_kpi,
-        each_pass_duration_card4_kpi,
-        train_loss_card4_kpi,
-        ]
+    each_pass_duration_card1_kpi,
+    train_loss_card1_kpi,
+    each_pass_duration_card4_kpi,
+    train_loss_card4_kpi,
+]
 
 
 def parse_log(log):

PaddleRec/din/infer.py → PaddleRec/ctr/din/infer.py

@@ -79,16 +79,18 @@ def infer():
      fetch_targets] = fluid.io.load_inference_model(model_path, exe)
 
     loader = fluid.io.DataLoader.from_generator(
-        feed_list=[inference_program.block(0).var(e) for e in feed_target_names], capacity=10000, iterable=True)
+        feed_list=[
+            inference_program.block(0).var(e) for e in feed_target_names
+        ],
+        capacity=10000,
+        iterable=True)
     loader.set_sample_list_generator(data_reader, places=place)
 
     loss_sum = 0.0
     score = []
     count = 0
     for data in loader():
-        res = exe.run(inference_program,
-                      feed=data,
-                      fetch_list=fetch_targets)
+        res = exe.run(inference_program, feed=data, fetch_list=fetch_targets)
         loss_sum += res[0]
         label_data = list(np.array(data[0]["label"]))
         for i in range(len(label_data)):

PaddleRec/din/network.py → PaddleRec/ctr/din/network.py

@@ -14,6 +14,7 @@
 
 import paddle.fluid as fluid
 
+
 def din_attention(hist, target_expand, mask):
     """activation weight"""
 
@@ -61,14 +62,10 @@ def network(item_count, cat_count):
         name="hist_item_seq", shape=[None, seq_len], dtype="int64")
     hist_cat_seq = fluid.data(
         name="hist_cat_seq", shape=[None, seq_len], dtype="int64")
-    target_item = fluid.data(
-        name="target_item", shape=[None], dtype="int64")
-    target_cat = fluid.data(
-        name="target_cat", shape=[None], dtype="int64")
-    label = fluid.data(
-        name="label", shape=[None, 1], dtype="float32")
-    mask = fluid.data(
-        name="mask", shape=[None, seq_len, 1], dtype="float32")
+    target_item = fluid.data(name="target_item", shape=[None], dtype="int64")
+    target_cat = fluid.data(name="target_cat", shape=[None], dtype="int64")
+    label = fluid.data(name="label", shape=[None, 1], dtype="float32")
+    mask = fluid.data(name="mask", shape=[None, seq_len, 1], dtype="float32")
     target_item_seq = fluid.data(
         name="target_item_seq", shape=[None, seq_len], dtype="int64")
     target_cat_seq = fluid.data(

PaddleRec/din/reader.py → PaddleRec/ctr/din/reader.py

@@ -18,6 +18,7 @@ import numpy as np
 import paddle
 import pickle
 
+
 def pad_batch_data(input, max_len):
     res = np.array([x + [0] * (max_len - len(x)) for x in input])
     res = res.astype("int64").reshape([-1, max_len])
@@ -33,11 +34,9 @@ def make_data(b):
         [[0] * x + [-1e9] * (max_len - x) for x in len_array]).reshape(
             [-1, max_len, 1])
     target_item_seq = np.array(
-        [[x[2]] * max_len for x in b]).astype("int64").reshape(
-            [-1, max_len])
+        [[x[2]] * max_len for x in b]).astype("int64").reshape([-1, max_len])
     target_cat_seq = np.array(
-        [[x[3]] * max_len for x in b]).astype("int64").reshape(
-            [-1, max_len])
+        [[x[3]] * max_len for x in b]).astype("int64").reshape([-1, max_len])
     res = []
     for i in range(len(b)):
         res.append([

PaddleRec/din/train.py → PaddleRec/ctr/din/train.py

@@ -33,11 +33,20 @@ logger.setLevel(logging.INFO)
 def parse_args():
     parser = argparse.ArgumentParser("din")
     parser.add_argument(
-        '--config_path', type=str, default='data/config.txt', help='dir of config')
+        '--config_path',
+        type=str,
+        default='data/config.txt',
+        help='dir of config')
     parser.add_argument(
-        '--train_dir', type=str, default='data/paddle_train.txt', help='dir of train file')
+        '--train_dir',
+        type=str,
+        default='data/paddle_train.txt',
+        help='dir of train file')
     parser.add_argument(
-        '--model_dir', type=str, default='din_amazon', help='dir of saved model')
+        '--model_dir',
+        type=str,
+        default='din_amazon',
+        help='dir of saved model')
     parser.add_argument(
         '--batch_size', type=int, default=16, help='number of batch size')
     parser.add_argument(
@@ -45,15 +54,19 @@ def parse_args():
     parser.add_argument(
         '--use_cuda', type=int, default=0, help='whether to use gpu')
     parser.add_argument(
-        '--parallel', type=int, default=0, help='whether to use parallel executor')
+        '--parallel',
+        type=int,
+        default=0,
+        help='whether to use parallel executor')
     parser.add_argument(
         '--base_lr', type=float, default=0.85, help='based learning rate')
     parser.add_argument(
         '--num_devices', type=int, default=1, help='Number of GPU devices')
     parser.add_argument(
-        '--enable_ce', action='store_true', help='If set, run the task with continuous evaluation logs.')
-    parser.add_argument(
-        '--batch_num', type=int, help="batch num for ce")
+        '--enable_ce',
+        action='store_true',
+        help='If set, run the task with continuous evaluation logs.')
+    parser.add_argument('--batch_num', type=int, help="batch num for ce")
     args = parser.parse_args()
     return args
 
@@ -133,8 +146,8 @@ def train():
 
                 if (global_step > 400000 and global_step % PRINT_STEP == 0) or (
                         global_step <= 400000 and global_step % 50000 == 0):
-                    save_dir = os.path.join(args.model_dir, "global_step_" + str(
-                        global_step))
+                    save_dir = os.path.join(args.model_dir,
+                                            "global_step_" + str(global_step))
                     feed_var_name = [
                         "hist_item_seq", "hist_cat_seq", "target_item",
                         "target_cat", "label", "mask", "target_item_seq",
@@ -156,10 +169,8 @@ def train():
             ce_time = total_time[-1]
         except:
             print("ce info error")
-        print("kpis\teach_pass_duration_card%s\t%s" %
-                    (gpu_num, ce_time))
-        print("kpis\ttrain_loss_card%s\t%s" %
-                    (gpu_num, ce_loss))
+        print("kpis\teach_pass_duration_card%s\t%s" % (gpu_num, ce_time))
+        print("kpis\ttrain_loss_card%s\t%s" % (gpu_num, ce_loss))
 
 
 def get_cards(args):