[Fluid Clean] remove paddle.fluid.dygraph.nn.conv2D (PaddlePaddle#1504)

* [Fluid Clean] remove paddle.fluid.dygraph.nn.conv2D

* remove layers_old in ofa

demo/darts/model.py

@@ -20,7 +20,8 @@
 import paddle.fluid as fluid
 from paddle.fluid.param_attr import ParamAttr
 from paddle.fluid.initializer import ConstantInitializer, MSRAInitializer
-from paddle.fluid.dygraph.nn import Conv2D, Pool2D, BatchNorm, Linear
+from paddle.nn import Conv2D
+from paddle.fluid.dygraph.nn import Pool2D, BatchNorm, Linear
 from paddle.fluid.dygraph.base import to_variable
 from genotypes import PRIMITIVES
 from genotypes import Genotype

demo/darts/operations.py

@@ -13,7 +13,8 @@
 # limitations under the License.
 
 import paddle.fluid as fluid
-from paddle.fluid.dygraph.nn import Conv2D, Pool2D, BatchNorm
+from paddle.nn import Conv2D
+from paddle.fluid.dygraph.nn import Pool2D, BatchNorm
 from paddle.fluid.param_attr import ParamAttr
 from paddle.fluid.initializer import ConstantInitializer, MSRAInitializer
 
@@ -58,10 +59,8 @@
 
 
 def bn_param_config(affine=False):
-    gama = ParamAttr(
-        initializer=ConstantInitializer(value=1), trainable=affine)
-    beta = ParamAttr(
-        initializer=ConstantInitializer(value=0), trainable=affine)
+    gama = ParamAttr(initializer=ConstantInitializer(value=1), trainable=affine)
+    beta = ParamAttr(initializer=ConstantInitializer(value=0), trainable=affine)
     return gama, beta
 
 
@@ -107,8 +106,7 @@ def __init__(self, c_in, c_out, affine=True):
             param_attr=fluid.ParamAttr(initializer=MSRAInitializer()),
             bias_attr=False)
         gama, beta = bn_param_config(affine)
-        self.bn = BatchNorm(
-            num_channels=c_out, param_attr=gama, bias_attr=beta)
+        self.bn = BatchNorm(num_channels=c_out, param_attr=gama, bias_attr=beta)
 
     def forward(self, x):
         x = fluid.layers.relu(x)
@@ -140,8 +138,7 @@ def __init__(self, c_in, c_out, kernel_size, stride, padding, affine=True):
             param_attr=fluid.ParamAttr(initializer=MSRAInitializer()),
             bias_attr=False)
         gama, beta = bn_param_config(affine)
-        self.bn1 = BatchNorm(
-            num_channels=c_in, param_attr=gama, bias_attr=beta)
+        self.bn1 = BatchNorm(num_channels=c_in, param_attr=gama, bias_attr=beta)
         self.conv3 = Conv2D(
             num_channels=c_in,
             num_filters=c_in,
@@ -257,8 +254,7 @@ def __init__(self, c_in, c_out, kernel_size, stride, padding, affine=True):
             param_attr=fluid.ParamAttr(initializer=MSRAInitializer()),
             bias_attr=False)
         gama, beta = bn_param_config(affine)
-        self.bn = BatchNorm(
-            num_channels=c_out, param_attr=gama, bias_attr=beta)
+        self.bn = BatchNorm(num_channels=c_out, param_attr=gama, bias_attr=beta)
 
     def forward(self, x):
         x = fluid.layers.relu(x)

demo/one_shot/train.py

@@ -21,7 +21,8 @@
 import paddle
 import paddle.fluid as fluid
 from paddle.fluid.optimizer import AdamOptimizer
-from paddle.fluid.dygraph.nn import Conv2D, Pool2D, Linear
+from paddle.nn import Conv2D
+from paddle.fluid.dygraph.nn import Pool2D, Linear
 from paddle.fluid.dygraph.base import to_variable
 
 from paddleslim.nas.one_shot import SuperMnasnet
@@ -142,8 +143,7 @@ def train_mnist(args, model, tokens=None):
     epoch_num = args.epoch
     BATCH_SIZE = 64
 
-    adam = AdamOptimizer(
-        learning_rate=0.001, parameter_list=model.parameters())
+    adam = AdamOptimizer(learning_rate=0.001, parameter_list=model.parameters())
 
     train_reader = paddle.fluid.io.batch(
         paddle.dataset.mnist.train(), batch_size=BATCH_SIZE, drop_last=True)
@@ -187,8 +187,7 @@ def train_mnist(args, model, tokens=None):
         print("Loss at epoch {} , acc is: {}".format(epoch, test_acc))
 
     save_parameters = (not args.use_data_parallel) or (
-        args.use_data_parallel and
-        fluid.dygraph.parallel.Env().local_rank == 0)
+        args.use_data_parallel and fluid.dygraph.parallel.Env().local_rank == 0)
     if save_parameters:
         fluid.save_dygraph(model.state_dict(), "save_temp")
         print("checkpoint saved")

paddleslim/models/dygraph/mobilenet.py

@@ -24,7 +24,8 @@
 from paddle.fluid.initializer import MSRA
 from paddle.fluid.param_attr import ParamAttr
 from paddle.fluid.layer_helper import LayerHelper
-from paddle.fluid.dygraph.nn import Conv2D, Pool2D, BatchNorm, Linear
+from paddle.nn import Conv2D
+from paddle.fluid.dygraph.nn import Pool2D, BatchNorm, Linear
 from paddle.fluid.dygraph.base import to_variable
 from paddle.fluid import framework
 

paddleslim/models/dygraph/resnet.py

@@ -15,7 +15,8 @@
 import paddle
 import paddle.fluid as fluid
 from paddle.fluid.layer_helper import LayerHelper
-from paddle.fluid.dygraph.nn import Conv2D, Pool2D, BatchNorm, Linear
+from paddle.nn import Conv2D
+from paddle.fluid.dygraph.nn import Pool2D, BatchNorm, Linear
 
 
 class ConvBNLayer(fluid.dygraph.Layer):
@@ -114,11 +115,7 @@ def __init__(self, layers=50, class_dim=100):
         num_filters = [64, 128, 256, 512]
 
         self.conv = ConvBNLayer(
-            num_channels=3,
-            num_filters=64,
-            filter_size=7,
-            stride=1,
-            act='relu')
+            num_channels=3, num_filters=64, filter_size=7, stride=1, act='relu')
         self.pool2d_max = Pool2D(
             pool_size=3, pool_stride=2, pool_padding=1, pool_type='max')
 

paddleslim/nas/darts/search_space/conv_bert/model/bert.py

@@ -23,8 +23,10 @@
 import numpy as np
 import paddle
 import paddle.fluid as fluid
-from paddle.fluid.dygraph import Embedding, LayerNorm, Linear, to_variable, Layer, guard
-from paddle.fluid.dygraph.nn import Conv2D, Pool2D, BatchNorm, Linear
+from paddle.nn import Conv2D
+from paddle.fluid.dygraph import Embedding, LayerNorm, Linear, Layer
+from paddle.fluid.dygraph import Pool2D, BatchNorm, Linear
+from paddle.fluid.dygraph import to_variable, guard
 from paddle.fluid import ParamAttr
 from paddle.fluid.initializer import MSRA
 from .transformer_encoder import EncoderLayer

paddleslim/nas/darts/search_space/conv_bert/model/transformer_encoder.py

@@ -22,8 +22,9 @@
 
 import paddle
 import paddle.fluid as fluid
+from paddle.nn import Conv2D
 from paddle.fluid.dygraph import Embedding, LayerNorm, Linear
-from paddle.fluid.dygraph import Conv2D, BatchNorm, Pool2D
+from paddle.fluid.dygraph import BatchNorm, Pool2D
 from paddle.fluid.dygraph import Layer
 from paddle.fluid.dygraph import to_variable
 from paddle.fluid.initializer import NormalInitializer

paddleslim/nas/ofa/__init__.py

@@ -16,10 +16,4 @@
 from .convert_super import supernet
 from .utils.special_config import *
 from .get_sub_model import *
-
-from .utils.utils import get_paddle_version
-pd_ver = get_paddle_version()
-if pd_ver == 185:
-    from .layers_old import *
-else:
-    from .layers import *
+from .layers import *

paddleslim/nas/ofa/convert_super.py

@@ -18,24 +18,15 @@
 import numbers
 import paddle
 from ...common import get_logger
+import paddle.nn as nn
+from paddle.nn import Conv2D, Conv2DTranspose, Linear, LayerNorm, Embedding, SyncBatchNorm
+from paddle import ParamAttr
 from .utils.utils import get_paddle_version
 pd_ver = get_paddle_version()
-if pd_ver == 185:
-    import paddle.fluid.dygraph.nn as nn
-    from paddle.fluid.dygraph.nn import Conv2D, Conv2DTranspose, Linear, LayerNorm, Embedding
-    from paddle.fluid import ParamAttr
-    from .layers_old import *
-    from . import layers_old as layers
-    Layer = paddle.fluid.dygraph.Layer
-else:
-    import paddle.nn as nn
-    from paddle.nn import Conv2D, Conv2DTranspose, Linear, LayerNorm, Embedding, SyncBatchNorm
-    from paddle import ParamAttr
-    from .layers import *
-    from . import layers
-    Layer = paddle.nn.Layer
+from .layers import *
+from . import layers
+from paddle.nn import Layer
 from .layers_base import Block
-from . import layers_old
 _logger = get_logger(__name__, level=logging.INFO)
 
 __all__ = ['supernet', 'Convert']

paddleslim/nas/ofa/layers.py

@@ -994,19 +994,19 @@ def forward(self, input):
         if in_dygraph_mode():
             if feature_dim != self._mean.shape[0]:
                 batch_norm_out, t1, t2, t3, t4, _ = _C_ops.batch_norm(
-                    input, weight, bias, mean, variance, self._momentum,
-                    self._epsilon, self._data_format, not self.training,
-                    self._use_global_stats, trainable_statistics, False, False)
+                    input, mean, variance, weight, bias, not self.training,
+                    self._momentum, self._epsilon, self._data_format,
+                    self._use_global_stats, trainable_statistics)
                 self._mean[:feature_dim].set_value(mean)
                 self._variance[:feature_dim].set_value(variance)
                 mean_out[:feature_dim].set_value(mean_out_tmp)
                 variance_out[:feature_dim].set_value(variance_out_tmp)
                 return batch_norm_out
             else:
                 batch_norm_out, t1, t2, t3, t4, _ = _C_ops.batch_norm(
-                    input, weight, bias, mean, variance, self._momentum,
-                    self._epsilon, self._data_format, not self.training,
-                    self._use_global_stats, trainable_statistics, False)
+                    input, mean, variance, weight, bias, not self.training,
+                    self._momentum, self._epsilon, self._data_format,
+                    self._use_global_stats, trainable_statistics)
                 return batch_norm_out
 
         elif _in_legacy_dygraph():