update ce (#4171)

PaddleCV/PaddleGAN/trainer/SPADE.py

@@ -401,7 +401,7 @@ class SPADE(object):
                                   0], batch_time))
 
                 sys.stdout.flush()
-
+                batch_id += 1
             if self.cfg.run_test:
                 test_program = gen_trainer.infer_program
                 image_name = fluid.data(

PaddleCV/face_detection/reader.py

@@ -97,7 +97,10 @@ def preprocess(img, bbox_labels, mode, settings, image_path):
 
         # sampling
         batch_sampler = []
-
+        # used for continuous evaluation
+        if 'ce_mode' in os.environ:
+           random.seed(0)
+           np.random.seed(0)
         prob = np.random.uniform(0., 1.)
         if prob > settings.data_anchor_sampling_prob:
             scale_array = np.array([16, 32, 64, 128, 256, 512])
@@ -229,7 +232,7 @@ def expand_bboxes(bboxes,
 
 def train_generator(settings, file_list, batch_size, shuffle=True):
     def reader():
-        if shuffle:
+        if shuffle and 'ce_mode' not in os.environ:
             np.random.shuffle(file_list)
         batch_out = []
         for item in file_list:

PaddleCV/face_detection/train.py

@@ -150,6 +150,7 @@ def train(args, config, train_params, train_file_list):
 
     #only for ce
     if args.enable_ce:
+        is_shuffle = False
         SEED = 102
         startup_prog.random_seed = SEED
         train_prog.random_seed = SEED

PaddleNLP/emotion_detection/utils.py

@@ -119,7 +119,7 @@ def data_reader(file_path, word_dict, num_examples, phrase, epoch, max_seq_len):
         Reader function
         """
         for idx in range(epoch):
-            if phrase == "train":
+            if phrase == "train" and 'ce_mode' not in os.environ:
                 random.shuffle(all_data)
             for wids, label, seq_len in all_data:
                 yield wids, label, seq_len