fix bug (#4706)

PaddleRec/ctr/wide_deep/train.py

@@ -48,5 +48,5 @@ def train(args, train_data_path):
 if __name__ == "__main__":
     args = args.parse_args()
     logger.info("epoch:{}, batch_size: {}, use_gpu: {}, train_data_path: {}, model_dir: {}, hidden1_units: {}, hidden2_units: {}, hidden3_units: {}".format(
-        args.epoch, args.batch_size, args.use_gpu, args.train_data_path, args.model_dir, args.hidden1_units, args.hidden2_units, args.hidden3_units))
+        args.epochs, args.batch_size, args.use_gpu, args.train_data_path, args.model_dir, args.hidden1_units, args.hidden2_units, args.hidden3_units))
     train(args, args.train_data_path)

PaddleRec/ncf/args.py

@@ -6,6 +6,7 @@ def parse_args():
     parser.add_argument('--dataset', nargs='?', default='ml-1m', help='Choose a dataset.')
     parser.add_argument('--epochs', type=int, default=20, help='Number of epochs.')
     parser.add_argument('--batch_size', type=int, default=256, help='Batch size.')
+    parser.add_argument('--test_epoch', type=str, default='19',help='test_epoch')
     parser.add_argument('--test_batch_size', type=int, default=100, help='Batch size.')
     parser.add_argument('--num_factors', type=int, default=8, help='Embedding size.')
     parser.add_argument('--num_users', type=int, default=6040, help='num_users')

PaddleRec/ncf/evaluate.py

 import math
-import heapq  # for retrieval topK
+import heapq # for retrieval topK
 import multiprocessing
 import numpy as np
 from time import time
@@ -23,36 +23,30 @@ _K = None
 _args = None
 _model_path = None
 
-
 def run_infer(args, model_path, test_data_path):
     test_data_generator = utils.Dataset()
-
+    
     with fluid.scope_guard(fluid.Scope()):
-        test_reader = fluid.io.batch(
-            test_data_generator.test(test_data_path, False),
-            batch_size=args.test_batch_size)
-
+        test_reader = fluid.io.batch(test_data_generator.test(test_data_path, False), batch_size=args.test_batch_size)
+            
         place = fluid.CUDAPlace(0) if args.use_gpu else fluid.CPUPlace()
         exe = fluid.Executor(place)
         exe.run(fluid.default_startup_program())
 
-        infer_program, feed_target_names, fetch_vars = fluid.io.load_inference_model(
-            model_path, exe)
+        infer_program, feed_target_names, fetch_vars = fluid.io.load_inference_model(model_path, exe)
 
         for data in test_reader():
             user_input = np.array([dat[0] for dat in data])
             item_input = np.array([dat[1] for dat in data])
 
-            pred_val = exe.run(
-                infer_program,
-                feed={"user_input": user_input,
-                      "item_input": item_input},
-                fetch_list=fetch_vars,
-                return_numpy=True)
-
+            pred_val = exe.run(infer_program,
+                       feed={"user_input": user_input,
+                            "item_input": item_input},
+                       fetch_list=fetch_vars,
+                        return_numpy=True)
+        
             return pred_val[0].reshape(1, -1).tolist()[0]
 
-
 def evaluate_model(args, testRatings, testNegatives, K, model_path):
     """
     Evaluate the performance (Hit_Ratio, NDCG) of top-K recommendation
@@ -62,23 +56,22 @@ def evaluate_model(args, testRatings, testNegatives, K, model_path):
     global _testRatings
     global _testNegatives
     global _K
-    global _model_path
+    global _model_path 
     global _args
-
+    
     _args = args
-    _model_path = model_path
+    _model_path= model_path
     _testRatings = testRatings
     _testNegatives = testNegatives
     _K = K
-
-    hits, ndcgs = [], []
+        
+    hits, ndcgs = [],[]
     for idx in range(len(_testRatings)):
-        (hr, ndcg) = eval_one_rating(idx)
+        (hr,ndcg) = eval_one_rating(idx)
         hits.append(hr)
-        ndcgs.append(ndcg)
+        ndcgs.append(ndcg)      
     return (hits, ndcgs)
 
-
 def eval_one_rating(idx):
     rating = _testRatings[idx]
     items = _testNegatives[idx]
@@ -87,9 +80,9 @@ def eval_one_rating(idx):
     items.append(gtItem)
     # Get prediction scores
     map_item_score = {}
-    users = np.full(len(items), u, dtype='int32')
-    users = users.reshape(-1, 1)
-    items_array = np.array(items).reshape(-1, 1)
+    users = np.full(len(items), u, dtype = 'int32')
+    users = users.reshape(-1,1)
+    items_array = np.array(items).reshape(-1,1)
     temp = np.hstack((users, items_array))
     np.savetxt("Data/test.txt", temp, fmt='%d', delimiter=',')
     predictions = run_infer(_args, _model_path, _args.test_data_path)
@@ -98,7 +91,7 @@ def eval_one_rating(idx):
         item = items[i]
         map_item_score[item] = predictions[i]
     items.pop()
-
+    
     # Evaluate top rank list
     ranklist = heapq.nlargest(_K, map_item_score, key=map_item_score.get)
     hr = getHitRatio(ranklist, gtItem)
@@ -106,17 +99,15 @@ def eval_one_rating(idx):
 
     return (hr, ndcg)
 
-
 def getHitRatio(ranklist, gtItem):
     for item in ranklist:
         if item == gtItem:
             return 1
     return 0
 
-
 def getNDCG(ranklist, gtItem):
     for i in range(len(ranklist)):
         item = ranklist[i]
         if item == gtItem:
-            return math.log(2) / math.log(i + 2)
+            return math.log(2) / math.log(i+2)
     return 0

PaddleRec/ncf/get_train_data.py

@@ -32,7 +32,7 @@ def get_train_data(filename, write_file, num_negatives):
         file = open(write_file, 'w') 
         print("writing " + write_file)
         
-        for (u, i) in mat:
+        for (u, i) in mat.keys():
             # positive instance
             user_input = str(u)
             item_input = str(i)

PaddleRec/ncf/infer.py

@@ -23,7 +23,7 @@ if __name__ == "__main__":
     topK = 10
    
     begin = time.time()
-    model_path = args.model_dir + "/epoch_" + str(12)
+    model_path = args.model_dir + "/epoch_" + args.test_epoch
     (hits, ndcgs) = evaluate_model(args, testRatings, testNegatives, topK, model_path)
     hr, ndcg = np.array(hits).mean(), np.array(ndcgs).mean()
     end = time.time()

PaddleRec/rerank/listwise/README.md

@@ -27,20 +27,6 @@
 
  python3.7 
 
-## 数据集说明
-
-本项目构造数据集验证模型的正确性，字段说明如下：
-
-user_slot_name：用户端特征群id
-
-item_slot_name：item段特征群id
-
-lenght：item的长度
-
-label：用户对给定的是否点击item的list
-
-注意：由于构造数据集的限制，本项目只用一个epoch，如果多个epoch，则多个epoch的数据是变化的，没有意义，因此只采用一个epoch。
-
 ## 单机训练
 
 GPU环境

PaddleRec/rerank/listwise/args.py

@@ -22,7 +22,7 @@ import sys
 
 def parse_args():
     parser = argparse.ArgumentParser(description=__doc__)
-    parser.add_argument("--epochs", type=int, default=1, help="epochs")
+    parser.add_argument("--epochs", type=int, default=20, help="epochs")
     parser.add_argument("--batch_size", type=int, default=32, help="batch_size")
     parser.add_argument("--test_epoch", type=int, default=1, help="test_epoch")
     parser.add_argument('--use_gpu', type=int, default=0, help='whether using gpu')

PaddleRec/rerank/listwise/train.py

@@ -51,17 +51,17 @@ def train(args):
     train_reader = fluid.io.batch(train_data_generator.get_train_data(), batch_size=args.batch_size)
     loader = fluid.io.DataLoader.from_generator(feed_list=inputs, capacity=args.batch_size, iterable=True)
     loader.set_sample_list_generator(train_reader, places=place)
-    for epoch in range(args.epochs):
-        for i in range(args.sample_size):
-            for batch_id, data in enumerate(loader()):
-                begin = time.time()
-                loss_val, auc = exe.run(program=fluid.default_main_program(),
-                        feed=data,
-                        fetch_list=[loss.name, auc_val],
-                        return_numpy=True)
-                end = time.time()
-                logger.info("epoch: {}, batch_id: {}, batch_time: {:.5f}s, loss: {:.5f}, auc: {:.5f}".format(
-                    epoch, batch_id, end-begin, float(np.array(loss_val)), float(np.array(auc))))
+
+    for i in range(args.sample_size):
+        for batch_id, data in enumerate(loader()):
+            begin = time.time()
+            loss_val, auc = exe.run(program=fluid.default_main_program(),
+                    feed=data,
+                    fetch_list=[loss.name, auc_val],
+                    return_numpy=True)
+            end = time.time()
+            logger.info("batch_id: {}, batch_time: {:.5f}s, loss: {:.5f}, auc: {:.5f}".format(
+                batch_id, end-begin, float(np.array(loss_val)), float(np.array(auc))))
         
     #save model
     model_dir = os.path.join(args.model_dir, 'epoch_' + str(1), "checkpoint")