update ce

fluid/PaddleCV/human_pose_estimation/train.py

@@ -147,9 +147,6 @@ def train(args):
 
     # Dataloader
     train_reader = paddle.batch(reader.train(), batch_size=args.batch_size)
-    if args.enable_ce:
-        import lib.coco_reader_ce as reader_ce
-        train_reader = paddle.batch(reader_ce.train_ce(), batch_size=args.batch_size)
     feeder = fluid.DataFeeder(place=place, feed_list=[image, target, target_weight])
 
     train_exe = fluid.ParallelExecutor(

fluid/PaddleCV/metric_learning/imgtool.py

@@ -9,6 +9,9 @@ import math
 import random
 import functools
 import numpy as np
+import os
+
+is_ce = int(os.environ.get('is_ce', 0))
 
 #random.seed(0)
 
@@ -17,6 +20,8 @@ def rotate_image(img):
     (h, w) = img.shape[:2]
     center = (w // 2, h // 2)
     angle = random.randint(-10, 10)
+    if is_ce:
+        aggle = 0
     M = cv2.getRotationMatrix2D(center, angle, 1.0)
     rotated = cv2.warpAffine(img, M, (w, h))
     return rotated
@@ -27,6 +32,8 @@ def random_crop(img, size, scale=None, ratio=None):
     ratio = [3. / 4., 4. / 3.] if ratio is None else ratio
 
     aspect_ratio = math.sqrt(random.uniform(*ratio))
+    if is_ce:
+        aspect_ratio = math.sqrt(1.)
     w = 1. * aspect_ratio
     h = 1. / aspect_ratio
 
@@ -35,14 +42,18 @@ def random_crop(img, size, scale=None, ratio=None):
     scale_max = min(scale[1], bound)
     scale_min = min(scale[0], bound)
 
-    target_area = img.shape[0] * img.shape[1] * random.uniform(scale_min,
-                                                             scale_max)
+    target_area = img.shape[0] * img.shape[1] * random.uniform(scale_min, scale_max)
+    if is_ce:
+        target_area = img.shape[0] * img.shape[1] * (scale_min + scale_max) / 2.
     target_size = math.sqrt(target_area)
     w = int(target_size * w)
     h = int(target_size * h)
 
     i = random.randint(0, img.shape[0] - h)
     j = random.randint(0, img.shape[1] - w)
+    if is_ce:
+        i = int(img.shape[0] - h) // 2
+        j = int(img.shape[1] - w) // 2
 
     img = img[i:i+h, j:j+w, :]
     resized = cv2.resize(img, (size, size), interpolation=cv2.INTER_LANCZOS4)
@@ -69,6 +80,9 @@ def crop_image(img, target_size, center):
     else:
         w_start = random.randint(0, width - size)
         h_start = random.randint(0, height - size)
+        if is_ce:
+            w_start = (width - size) // 2
+            h_start = (height - size) // 2
     w_end = w_start + size
     h_end = h_start + size
     img = img[h_start:h_end, w_start:w_end, :]
@@ -93,6 +107,8 @@ def process_image(sample, mode, color_jitter, rotate,
             img = distort_color(img)
         if random.randint(0, 1) == 1:
             img = img[:, ::-1, :]
+        if is_ce:
+            img = img[:, ::-1, :]
     else:
         if crop_size > 0:
             img = resize_short(img, crop_size)

fluid/PaddleCV/metric_learning/reader.py

@@ -10,6 +10,8 @@ import numpy as np
 import paddle
 from imgtool import process_image
 
+is_ce = int(os.environ.get('is_ce', 0))
+
 random.seed(0)
 
 DATA_DIR = "./data/Stanford_Online_Products/"
@@ -32,7 +34,8 @@ def init_sop(mode):
             if label not in train_data:
                 train_data[label] = []
             train_data[label].append(path)
-        random.shuffle(train_image_list)
+        if not is_ce:
+            random.shuffle(train_image_list)
         print("{} dataset size: {}".format(mode, len(train_data)))
         return train_data, train_image_list
     else:
@@ -67,13 +70,15 @@ def common_iterator(data, settings):
         lab_num = len(labs)
         ind = list(range(0, lab_num))
         while True:
-            random.shuffle(ind)
+            if not is_ce:
+                random.shuffle(ind)
             ind_sample = ind[:class_num]
             for ind_i in ind_sample:
                 lab = labs[ind_i]
                 data_list = data[lab]
                 data_ind = list(range(0, len(data_list)))
-                random.shuffle(data_ind)
+                if not is_ce:
+                    random.shuffle(data_ind)
                 anchor_ind = data_ind[:samples_each_class]
 
                 for anchor_ind_i in anchor_ind:
@@ -90,17 +95,21 @@ def triplet_iterator(data, settings):
         lab_num = len(labs)
         ind = list(range(0, lab_num))
         while True:
-            random.shuffle(ind)
+            if not is_ce:
+                random.shuffle(ind)
             ind_pos, ind_neg = ind[:2]
             lab_pos = labs[ind_pos]
             pos_data_list = data[lab_pos]
             data_ind = list(range(0, len(pos_data_list)))
-            random.shuffle(data_ind)
+            if not is_ce:
+                random.shuffle(data_ind)
             anchor_ind, pos_ind = data_ind[:2]
 
             lab_neg = labs[ind_neg]
             neg_data_list = data[lab_neg]
             neg_ind = random.randint(0, len(neg_data_list) - 1)
+            if is_ce:
+                neg_ind = 1
             
             anchor_path = DATA_DIR + pos_data_list[anchor_ind]
             yield anchor_path, lab_pos
@@ -158,6 +167,8 @@ def createreader(settings, mode):
     assert(image_shape[1] == image_shape[2])
     image_size = int(image_shape[2])
     keep_order = False if mode != 'train' or settings.loss_name in ['softmax', 'arcmargin'] else True
+    if is_ce:
+        keep_order = True
     image_mapper = functools.partial(process_image,
             mode=mode, color_jitter=False, rotate=False, crop_size=image_size)
     reader = paddle.reader.xmap_readers(

fluid/PaddleCV/metric_learning/train_elem.py

@@ -194,10 +194,6 @@ def train_async(args):
     
     train_reader = paddle.batch(reader.train(args), batch_size=train_batch_size, drop_last=True)
     test_reader = paddle.batch(reader.test(args), batch_size=test_batch_size, drop_last=False)
-    if args.enable_ce:
-        import reader_ce
-        train_reader = paddle.batch(reader_ce.train(args), batch_size=train_batch_size, drop_last=False)
-        test_reader = paddle.batch(reader_ce.test(args), batch_size=test_batch_size, drop_last=False)
     test_feeder = fluid.DataFeeder(place=place, feed_list=[image, label])
     train_py_reader.decorate_paddle_reader(train_reader)