some fix about CE (#1242)

* ce fix

fluid/image_classification/.run_ce.sh

@@ -2,10 +2,11 @@
 
 # This file is only used for continuous evaluation.
 export FLAGS_cudnn_deterministic=True
+BATCH_SIZE=56
 cudaid=${object_detection_cudaid:=0}
 export CUDA_VISIBLE_DEVICES=$cudaid
-python train.py --batch_size=64 --num_epochs=5 --enable_ce=True --lr_strategy=cosine_decay | python _ce.py
+python train.py --batch_size=${BATCH_SIZE} --num_epochs=5 --enable_ce=True --lr_strategy=cosine_decay | python _ce.py
 
 cudaid=${object_detection_cudaid_m:=0, 1, 2, 3}
 export CUDA_VISIBLE_DEVICES=$cudaid
-python train.py --batch_size=64 --num_epochs=5 --enable_ce=True --lr_strategy=cosine_decay | python _ce.py
+python train.py --batch_size=${BATCH_SIZE} --num_epochs=5 --enable_ce=True --lr_strategy=cosine_decay | python _ce.py

fluid/object_detection/reader.py

@@ -292,24 +292,24 @@ def train(settings,
           shuffle=True,
           use_multiprocessing=True,
           num_workers=8,
-          max_queue=24):
+          max_queue=24,
+          enable_ce=False):
     file_list = os.path.join(settings.data_dir, file_list)
 
-    def infinite_reader(gen):
-        while True:
-            for data in gen():
-                yield data
-
     if 'coco' in settings.dataset:
         generator = coco(settings, file_list, "train", batch_size, shuffle)
     else:
         generator = pascalvoc(settings, file_list, "train", batch_size, shuffle)
 
+    def infinite_reader():
+        while True:
+            for data in generator():
+                yield data
+
     def reader():
         try:
             enqueuer = GeneratorEnqueuer(
-                infinite_reader(generator),
-                use_multiprocessing=use_multiprocessing)
+                infinite_reader(), use_multiprocessing=use_multiprocessing)
             enqueuer.start(max_queue_size=max_queue, workers=num_workers)
             generator_output = None
             while True:
@@ -325,7 +325,10 @@ def train(settings,
             if enqueuer is not None:
                 enqueuer.stop()
 
-    return reader
+    if enable_ce:
+        return infinite_reader
+    else:
+        return reader
 
 
 def test(settings, file_list, batch_size):

fluid/object_detection/train.py

@@ -32,7 +32,7 @@ add_arg('enable_ce',     bool,  False, "Whether use CE to evaluate the model")
 
 train_parameters = {
     "pascalvoc": {
-        "train_images": 19200,
+        "train_images": 16551,
         "image_shape": [3, 300, 300],
         "class_num": 21,
         "batch_size": 64,
@@ -143,7 +143,6 @@ def train(args,
         startup_prog.random_seed = 111
         train_prog.random_seed = 111
         test_prog.random_seed = 111
-        num_workers = 1
 
     train_py_reader, loss = build_program(
         main_prog=train_prog,
@@ -170,14 +169,14 @@ def train(args,
     if parallel:
         train_exe = fluid.ParallelExecutor(main_program=train_prog,
             use_cuda=use_gpu, loss_name=loss.name)
-
     train_reader = reader.train(data_args,
                                 train_file_list,
                                 batch_size_per_device,
                                 shuffle=is_shuffle,
                                 use_multiprocessing=True,
                                 num_workers=num_workers,
-                                max_queue=24)
+                                max_queue=24,
+                                enable_ce=enable_ce)
     test_reader = reader.test(data_args, val_file_list, batch_size)
     train_py_reader.decorate_paddle_reader(train_reader)
     test_py_reader.decorate_paddle_reader(test_reader)