!20621 fix the bug with slightly lower accuracy of warpctc.

Merge pull request !20621 from 郑彬/warpctc_0720

model_zoo/official/cv/warpctc/src/warpctc_for_train.py

@@ -22,7 +22,6 @@ from mindspore.nn.cell import Cell
 from mindspore.nn.wrap.grad_reducer import DistributedGradReducer
 import mindspore.nn as nn
 from mindspore.common.tensor import Tensor
-from src.model_utils.device_adapter import get_device_num
 
 compute_norm = C.MultitypeFuncGraph("compute_norm")
 
@@ -68,7 +67,7 @@ class TrainOneStepCellWithGradClip(Cell):
         Tensor, a scalar Tensor with shape :math:`()`.
     """
 
-    def __init__(self, network, optimizer, sens=1.0):
+    def __init__(self, network, optimizer, device_num, sens=1.0):
         super(TrainOneStepCellWithGradClip, self).__init__(auto_prefix=False)
         self.network = network
         self.network.set_grad()
@@ -87,7 +86,7 @@ class TrainOneStepCellWithGradClip(Cell):
         self.cast = P.Cast()
         self.concat = P.Concat(axis=0)
         self.ten = Tensor(np.array([10.0]).astype(np.float32))
-        if get_device_num() > 1:
+        if device_num > 1:
             self.reducer_flag = True
         if self.reducer_flag:
             self.grad_reducer = DistributedGradReducer(optimizer.parameters)

model_zoo/official/cv/warpctc/train.py

@@ -145,7 +145,7 @@ def train():
     opt = nn.SGD(params=net.trainable_params(), learning_rate=lr, momentum=config.momentum)
 
     net = WithLossCell(net, loss)
-    net = TrainOneStepCellWithGradClip(net, opt).set_train()
+    net = TrainOneStepCellWithGradClip(net, opt, device_num).set_train()
     # define model
     model = Model(net)
     # define callbacks