fix paddle/understand_sentiment_lstm.py (the size of last batch is sa…

…mller than batch_size)
dzhwinter · Dec 12, 2017 · d3b8e07 · d3b8e07
1 parent 9c2b3c5
commit d3b8e07
Show file tree

Hide file tree

Showing 3 changed files with 24 additions and 12 deletions.
diff --git a/paddle/resnet.py b/paddle/resnet.py
@@ -168,9 +168,10 @@ def run_benchmark(model, args):
     iter = 0
     im_num = 0
     for pass_id in range(args.pass_num):
-        accuracy.reset(exe)
         if iter == args.iterations:
             break
+        accuracy.reset(exe)
+
         for batch_id, data in enumerate(train_reader()):
             if iter == args.skip_batch_num:
                 start_time = time.time()

diff --git a/paddle/understand_sentiment_dynamic_lstm.py b/paddle/understand_sentiment_dynamic_lstm.py
@@ -121,10 +121,9 @@ def run_benchmark(model, args):
     exe = fluid.Executor(place)
     exe.run(fluid.default_startup_program())
 
-    for it, pass_id in enumerate(xrange(args.pass_num)):
+    iterator = 0
+    for pass_id in xrange(args.pass_num):
         accuracy.reset(exe)
-        if iter == args.iterations:
-            break
         for data in train_reader():
             tensor_words = to_lodtensor(map(lambda x: x[0], data), place)
 
@@ -140,8 +139,13 @@ def run_benchmark(model, args):
                       "label": tensor_label},
                 fetch_list=[avg_cost] + accuracy.metrics)
             pass_acc = accuracy.eval(exe)
-            print("Iter: %d, loss: %s, acc: %s, pass_acc: %s" %
-                  (it, str(loss), str(acc), str(pass_acc)))
+
+            iterator += 1
+            print("pass_id:%d, Iter: %d, loss: %s, acc: %s, pass_acc: %s" %
+                  (pass_id, iterator, str(loss), str(acc), str(pass_acc)))
+
+            if iterator == args.iterations:
+                return
 
 
 if __name__ == '__main__':

diff --git a/paddle/understand_sentiment_lstm.py b/paddle/understand_sentiment_lstm.py
@@ -119,7 +119,7 @@ def run_benchmark(model, args):
     word_dict = paddle.dataset.imdb.word_dict()
 
     print("load word dict successfully")
-
+    
     dict_dim = len(word_dict)
     data = fluid.layers.data(
         name="words",
@@ -147,13 +147,15 @@ def run_benchmark(model, args):
     place = fluid.CPUPlace() if args.device == 'CPU' else fluid.GPUPlace(0)
     exe = fluid.Executor(place)
     exe.run(fluid.default_startup_program())
-    for it, pass_id in enumerate(xrange(args.pass_num)):
+
+    iterator = 0
+    for pass_id in xrange(args.pass_num):
         accuracy.reset(exe)
-        if iter == args.iterations:
-            break
         for data in train_reader():
             chopped_data = chop_data(
                 data, chop_len=args.seq_len, batch_size=args.batch_size)
+            if len(data) < args.batch_size : continue
+
             tensor_words, tensor_label = prepare_feed_data(chopped_data, place)
 
             loss, acc = exe.run(
@@ -162,8 +164,13 @@ def run_benchmark(model, args):
                       "label": tensor_label},
                 fetch_list=[avg_cost] + accuracy.metrics)
             pass_acc = accuracy.eval(exe)
-            print("Iter: %d, loss: %s, acc: %s, pass_acc: %s" %
-                  (it, str(loss), str(acc), str(pass_acc)))
+
+            iterator += 1
+            print("pass_id:%d, Iter: %d, loss: %s, acc: %s, pass_acc: %s" %
+                  (pass_id, iterator, str(loss), str(acc), str(pass_acc)))
+
+            if iterator == args.iterations:
+                return
 
 
 if __name__ == '__main__':