update data augment and init method for MobileViTv3-v2

ppcls/arch/backbone/model_zoo/mobilevit_v3.py

@@ -687,7 +687,14 @@ class MobileViTv3(nn.Layer):
 
     def _init_weights(self, m):
         if isinstance(m, nn.Conv2D):
+            fan_in = m.weight.shape[1] * m.weight.shape[2] * m.weight.shape[3]
             fan_out = m.weight.shape[0] * m.weight.shape[2] * m.weight.shape[3]
+            if self.mobilevit_v2_based:
+                bound = 1.0 / fan_in**0.5
+                nn.initializer.Uniform(-bound, bound)(m.weight)
+                if m.bias is not None:
+                    nn.initializer.Uniform(-bound, bound)(m.bias)
+            else:
                 nn.initializer.KaimingNormal(fan_in=fan_out)(m.weight)
                 if m.bias is not None:
                     nn.initializer.Constant(0)(m.bias)
@@ -695,6 +702,9 @@ class MobileViTv3(nn.Layer):
             nn.initializer.Constant(1)(m.weight)
             nn.initializer.Constant(0)(m.bias)
         elif isinstance(m, nn.Linear):
+            if self.mobilevit_v2_based:
+                nn.initializer.XavierUniform()(m.weight)
+            else:
                 nn.initializer.TruncatedNormal(std=.02)(m.weight)
             if m.bias is not None:
                 nn.initializer.Constant(0)(m.bias)

ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x0_5.yaml

@@ -93,15 +93,15 @@ DataLoader:
             r1: 0.3
             attempt: 10
             use_log_aspect: True
-            mode: pixel
+            mode: const
       batch_transform_ops:
         - OpSampler:
             MixupOperator:
               alpha: 0.2
-              prob: 0.5
+              prob: 0.25
             CutmixOperator:
               alpha: 1.0
-              prob: 0.5
+              prob: 0.25
     sampler:
       name: DistributedBatchSampler
       batch_size: 128

ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x0_75.yaml

@@ -93,15 +93,15 @@ DataLoader:
             r1: 0.3
             attempt: 10
             use_log_aspect: True
-            mode: pixel
+            mode: const
       batch_transform_ops:
         - OpSampler:
             MixupOperator:
               alpha: 0.2
-              prob: 0.5
+              prob: 0.25
             CutmixOperator:
               alpha: 1.0
-              prob: 0.5
+              prob: 0.25
     sampler:
       name: DistributedBatchSampler
       batch_size: 128

ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x1_0.yaml

@@ -93,15 +93,15 @@ DataLoader:
             r1: 0.3
             attempt: 10
             use_log_aspect: True
-            mode: pixel
+            mode: const
       batch_transform_ops:
         - OpSampler:
             MixupOperator:
               alpha: 0.2
-              prob: 0.5
+              prob: 0.25
             CutmixOperator:
               alpha: 1.0
-              prob: 0.5
+              prob: 0.25
     sampler:
       name: DistributedBatchSampler
       batch_size: 128