polish code

fluid/PaddleCV/image_classification/models/fast_resnet.py

@@ -130,41 +130,3 @@ class FastResNet():
         short = self.shortcut(input, num_filters * 4, stride)
 
         return fluid.layers.elementwise_add(x=short, y=conv2, act='relu')
-
-
-def _model_reader_dshape_classdim(args, is_train, val_bs=None, sz=224, trn_dir="", min_scale=0.08, rect_val=False):
-    reader = None
-    if args.data_set == "imagenet":
-        class_dim = 1000
-        if args.data_format == 'NCHW':
-            dshape = [3, sz, sz]
-        else:
-            dshape = [sz, sz, 3]
-        if is_train:
-            reader = torchvision_reader.train(
-                traindir="/data/imagenet/%strain" % trn_dir, sz=sz, min_scale=min_scale)
-        else:
-            reader = torchvision_reader.test(
-                valdir="/data/imagenet/%svalidation" % trn_dir, bs=val_bs, sz=sz, rect_val=rect_val)
-    else:
-        raise ValueError("only support imagenet dataset.")
-
-    return None, reader, dshape, class_dim
-
-def lr_decay(lrs, epochs, bs, total_image):
-    boundaries = []
-    values = []
-    import math
-    for idx, epoch in enumerate(epochs):
-        step = math.ceil(total_image * 1.0 / (bs[idx] * 8))
-        ratio = (lrs[idx][1] - lrs[idx][0]) / (epoch[1] - epoch[0])
-        lr_base = lrs[idx][0]
-        for s in xrange(epoch[0], epoch[1]):
-            if boundaries:
-                boundaries.append(boundaries[-1] + step)
-            else:
-                boundaries = [step]
-            values.append(lr_base + ratio * (s - epoch[0]))
-    values.append(lrs[-1])
-    return boundaries, values
-