[Fix] Align HSigmoid with pytorch official implementation(#1622)

* [Fix] align hsigmoid with pytorch official

* [Fix] add warnings for Hsigmoid

* [Fix] fix format

* [Fix] add unittest

* [Fix] fix docstring

mmcv/cnn/bricks/hsigmoid.py

@@ -1,4 +1,6 @@
 # Copyright (c) OpenMMLab. All rights reserved.
+import warnings
+
 import torch.nn as nn
 
 from .registry import ACTIVATION_LAYERS
@@ -8,11 +10,15 @@ from .registry import ACTIVATION_LAYERS
 class HSigmoid(nn.Module):
     """Hard Sigmoid Module. Apply the hard sigmoid function:
     Hsigmoid(x) = min(max((x + bias) / divisor, min_value), max_value)
-    Default: Hsigmoid(x) = min(max((x + 1) / 2, 0), 1)
+    Default: Hsigmoid(x) = min(max((x + 3) / 6, 0), 1)
+
+    Note:
+        In MMCV v1.4.4, we modified the default value of args to align with
+        PyTorch official.
 
     Args:
-        bias (float): Bias of the input feature map. Default: 1.0.
-        divisor (float): Divisor of the input feature map. Default: 2.0.
+        bias (float): Bias of the input feature map. Default: 3.0.
+        divisor (float): Divisor of the input feature map. Default: 6.0.
         min_value (float): Lower bound value. Default: 0.0.
         max_value (float): Upper bound value. Default: 1.0.
 
@@ -20,8 +26,14 @@ class HSigmoid(nn.Module):
         Tensor: The output tensor.
     """
 
-    def __init__(self, bias=1.0, divisor=2.0, min_value=0.0, max_value=1.0):
+    def __init__(self, bias=3.0, divisor=6.0, min_value=0.0, max_value=1.0):
         super(HSigmoid, self).__init__()
+        warnings.warn(
+            'In MMCV v1.4.4, we modified the default value of args to align '
+            'with PyTorch official. Previous Implementation: '
+            'Hsigmoid(x) = min(max((x + 1) / 2, 0), 1). '
+            'Current Implementation: '
+            'Hsigmoid(x) = min(max((x + 3) / 6, 0), 1).')
         self.bias = bias
         self.divisor = divisor
         assert self.divisor != 0

requirements/test.txt

@@ -6,4 +6,4 @@ onnxruntime>=1.8.0
 pytest
 PyTurboJPEG
 scipy
-tiffile
+tifffile

tests/test_cnn/test_hsigmoid.py

@@ -15,7 +15,7 @@ def test_hsigmoid():
     input = torch.randn(input_shape)
     output = act(input)
     expected_output = torch.min(
-        torch.max((input + 1) / 2, torch.zeros(input_shape)),
+        torch.max((input + 3) / 6, torch.zeros(input_shape)),
         torch.ones(input_shape))
     # test output shape
     assert output.shape == expected_output.shape
@@ -23,12 +23,12 @@ def test_hsigmoid():
     assert torch.equal(output, expected_output)
 
     # test with designated parameters
-    act = HSigmoid(3, 6, 0, 1)
+    act = HSigmoid(1, 2, 0, 1)
     input_shape = torch.Size([1, 3, 64, 64])
     input = torch.randn(input_shape)
     output = act(input)
     expected_output = torch.min(
-        torch.max((input + 3) / 6, torch.zeros(input_shape)),
+        torch.max((input + 1) / 2, torch.zeros(input_shape)),
         torch.ones(input_shape))
     # test output shape
     assert output.shape == expected_output.shape