You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Docs: fix unexpected indentation around lists (#1297)
* Docs: fix unexpected indentation around lists
Signed-off-by: Adam J. Stewart <ajstewart426@gmail.com>
* Possible fix for MAnet
Signed-off-by: Adam J. Stewart <ajstewart426@gmail.com>
* Possible fix for DeepLabV3+
Signed-off-by: Adam J. Stewart <ajstewart426@gmail.com>
---------
Signed-off-by: Adam J. Stewart <ajstewart426@gmail.com>
Copy file name to clipboardExpand all lines: segmentation_models_pytorch/decoders/fpn/model.py
+6-5Lines changed: 6 additions & 5 deletions
Original file line number
Diff line number
Diff line change
@@ -40,11 +40,12 @@ class FPN(SegmentationModel):
40
40
upsampling: Final upsampling factor. Default is 4 to preserve input-output spatial shape identity
41
41
aux_params: Dictionary with parameters of the auxiliary output (classification head). Auxiliary output is build
42
42
on top of encoder if **aux_params** is not **None** (default). Supported params:
43
-
- classes (int): A number of classes
44
-
- pooling (str): One of "max", "avg". Default is "avg"
45
-
- dropout (float): Dropout factor in [0, 1)
46
-
- activation (str): An activation function to apply "sigmoid"/"softmax"
47
-
(could be **None** to return logits)
43
+
44
+
- classes (int): A number of classes
45
+
- pooling (str): One of "max", "avg". Default is "avg"
46
+
- dropout (float): Dropout factor in [0, 1)
47
+
- activation (str): An activation function to apply "sigmoid"/"softmax"
48
+
(could be **None** to return logits)
48
49
kwargs: Arguments passed to the encoder class ``__init__()`` function. Applies only to ``timm`` models. Keys with ``None`` values are pruned before passing.
Copy file name to clipboardExpand all lines: segmentation_models_pytorch/decoders/linknet/model.py
+7-5Lines changed: 7 additions & 5 deletions
Original file line number
Diff line number
Diff line change
@@ -34,6 +34,7 @@ class Linknet(SegmentationModel):
34
34
random initialization. The pretrained variant is defined in the encoder name
35
35
decoder_use_norm: Specifies normalization between Conv2D and activation.
36
36
Accepts the following types:
37
+
37
38
- **True**: Defaults to `"batchnorm"`.
38
39
- **False**: No normalization (`nn.Identity`).
39
40
- **str**: Specifies normalization type using default parameters. Available values:
@@ -55,11 +56,12 @@ class Linknet(SegmentationModel):
55
56
**callable** and **None**. Default is **None**.
56
57
aux_params: Dictionary with parameters of the auxiliary output (classification head). Auxiliary output is build
57
58
on top of encoder if **aux_params** is not **None** (default). Supported params:
58
-
- classes (int): A number of classes
59
-
- pooling (str): One of "max", "avg". Default is "avg"
60
-
- dropout (float): Dropout factor in [0, 1)
61
-
- activation (str): An activation function to apply "sigmoid"/"softmax"
62
-
(could be **None** to return logits)
59
+
60
+
- classes (int): A number of classes
61
+
- pooling (str): One of "max", "avg". Default is "avg"
62
+
- dropout (float): Dropout factor in [0, 1)
63
+
- activation (str): An activation function to apply "sigmoid"/"softmax"
64
+
(could be **None** to return logits)
63
65
kwargs: Arguments passed to the encoder class ``__init__()`` function. Applies only to ``timm`` models. Keys with ``None`` values are pruned before passing.
Copy file name to clipboardExpand all lines: segmentation_models_pytorch/decoders/manet/model.py
+11-8Lines changed: 11 additions & 8 deletions
Original file line number
Diff line number
Diff line change
@@ -15,9 +15,10 @@
15
15
classMAnet(SegmentationModel):
16
16
"""MAnet_ : Multi-scale Attention Net. The MA-Net can capture rich contextual dependencies based on
17
17
the attention mechanism, using two blocks:
18
-
- Position-wise Attention Block (PAB), which captures the spatial dependencies between pixels in a global view
19
-
- Multi-scale Fusion Attention Block (MFAB), which captures the channel dependencies between any feature map by
20
-
multi-scale semantic feature fusion
18
+
19
+
- Position-wise Attention Block (PAB), which captures the spatial dependencies between pixels in a global view
20
+
- Multi-scale Fusion Attention Block (MFAB), which captures the channel dependencies between any feature map by
21
+
multi-scale semantic feature fusion
21
22
22
23
Args:
23
24
encoder_name: Name of the classification model that will be used as an encoder (a.k.a backbone)
@@ -34,6 +35,7 @@ class MAnet(SegmentationModel):
34
35
Length of the list should be the same as **encoder_depth**
35
36
decoder_use_norm: Specifies normalization between Conv2D and activation.
36
37
Accepts the following types:
38
+
37
39
- **True**: Defaults to `"batchnorm"`.
38
40
- **False**: No normalization (`nn.Identity`).
39
41
- **str**: Specifies normalization type using default parameters. Available values:
@@ -59,11 +61,12 @@ class MAnet(SegmentationModel):
59
61
**callable** and **None**. Default is **None**.
60
62
aux_params: Dictionary with parameters of the auxiliary output (classification head). Auxiliary output is build
61
63
on top of encoder if **aux_params** is not **None** (default). Supported params:
62
-
- classes (int): A number of classes
63
-
- pooling (str): One of "max", "avg". Default is "avg"
64
-
- dropout (float): Dropout factor in [0, 1)
65
-
- activation (str): An activation function to apply "sigmoid"/"softmax"
66
-
(could be **None** to return logits)
64
+
65
+
- classes (int): A number of classes
66
+
- pooling (str): One of "max", "avg". Default is "avg"
67
+
- dropout (float): Dropout factor in [0, 1)
68
+
- activation (str): An activation function to apply "sigmoid"/"softmax"
69
+
(could be **None** to return logits)
67
70
kwargs: Arguments passed to the encoder class ``__init__()`` function. Applies only to ``timm`` models. Keys with ``None`` values are pruned before passing.
Copy file name to clipboardExpand all lines: segmentation_models_pytorch/decoders/pan/model.py
+6-5Lines changed: 6 additions & 5 deletions
Original file line number
Diff line number
Diff line change
@@ -43,11 +43,12 @@ class PAN(SegmentationModel):
43
43
upsampling: Final upsampling factor. Default is 4 to preserve input-output spatial shape identity
44
44
aux_params: Dictionary with parameters of the auxiliary output (classification head). Auxiliary output is build
45
45
on top of encoder if **aux_params** is not **None** (default). Supported params:
46
-
- classes (int): A number of classes
47
-
- pooling (str): One of "max", "avg". Default is "avg"
48
-
- dropout (float): Dropout factor in [0, 1)
49
-
- activation (str): An activation function to apply "sigmoid"/"softmax"
50
-
(could be **None** to return logits)
46
+
47
+
- classes (int): A number of classes
48
+
- pooling (str): One of "max", "avg". Default is "avg"
49
+
- dropout (float): Dropout factor in [0, 1)
50
+
- activation (str): An activation function to apply "sigmoid"/"softmax"
51
+
(could be **None** to return logits)
51
52
kwargs: Arguments passed to the encoder class ``__init__()`` function. Applies only to ``timm`` models. Keys with ``None`` values are pruned before passing.
Copy file name to clipboardExpand all lines: segmentation_models_pytorch/decoders/pspnet/model.py
+7-5Lines changed: 7 additions & 5 deletions
Original file line number
Diff line number
Diff line change
@@ -33,6 +33,7 @@ class PSPNet(SegmentationModel):
33
33
psp_out_channels: A number of filters in Spatial Pyramid
34
34
decoder_use_norm: Specifies normalization between Conv2D and activation.
35
35
Accepts the following types:
36
+
36
37
- **True**: Defaults to `"batchnorm"`.
37
38
- **False**: No normalization (`nn.Identity`).
38
39
- **str**: Specifies normalization type using default parameters. Available values:
@@ -56,11 +57,12 @@ class PSPNet(SegmentationModel):
56
57
upsampling: Final upsampling factor. Default is 8 to preserve input-output spatial shape identity
57
58
aux_params: Dictionary with parameters of the auxiliary output (classification head). Auxiliary output is build
58
59
on top of encoder if **aux_params** is not **None** (default). Supported params:
59
-
- classes (int): A number of classes
60
-
- pooling (str): One of "max", "avg". Default is "avg"
61
-
- dropout (float): Dropout factor in [0, 1)
62
-
- activation (str): An activation function to apply "sigmoid"/"softmax"
63
-
(could be **None** to return logits)
60
+
61
+
- classes (int): A number of classes
62
+
- pooling (str): One of "max", "avg". Default is "avg"
63
+
- dropout (float): Dropout factor in [0, 1)
64
+
- activation (str): An activation function to apply "sigmoid"/"softmax"
65
+
(could be **None** to return logits)
64
66
kwargs: Arguments passed to the encoder class ``__init__()`` function. Applies only to ``timm`` models. Keys with ``None`` values are pruned before passing.
Copy file name to clipboardExpand all lines: segmentation_models_pytorch/decoders/segformer/model.py
+6-5Lines changed: 6 additions & 5 deletions
Original file line number
Diff line number
Diff line change
@@ -34,11 +34,12 @@ class Segformer(SegmentationModel):
34
34
upsampling: A number to upsample the output of the model, default is 4 (same size as input)
35
35
aux_params: Dictionary with parameters of the auxiliary output (classification head). Auxiliary output is build
36
36
on top of encoder if **aux_params** is not **None** (default). Supported params:
37
-
- classes (int): A number of classes
38
-
- pooling (str): One of "max", "avg". Default is "avg"
39
-
- dropout (float): Dropout factor in [0, 1)
40
-
- activation (str): An activation function to apply "sigmoid"/"softmax"
41
-
(could be **None** to return logits)
37
+
38
+
- classes (int): A number of classes
39
+
- pooling (str): One of "max", "avg". Default is "avg"
40
+
- dropout (float): Dropout factor in [0, 1)
41
+
- activation (str): An activation function to apply "sigmoid"/"softmax"
42
+
(could be **None** to return logits)
42
43
kwargs: Arguments passed to the encoder class ``__init__()`` function. Applies only to ``timm`` models. Keys with ``None`` values are pruned before passing.
Copy file name to clipboardExpand all lines: segmentation_models_pytorch/decoders/unet/model.py
+7-5Lines changed: 7 additions & 5 deletions
Original file line number
Diff line number
Diff line change
@@ -44,6 +44,7 @@ class Unet(SegmentationModel):
44
44
Length of the list should be the same as **encoder_depth**
45
45
decoder_use_norm: Specifies normalization between Conv2D and activation.
46
46
Accepts the following types:
47
+
47
48
- **True**: Defaults to `"batchnorm"`.
48
49
- **False**: No normalization (`nn.Identity`).
49
50
- **str**: Specifies normalization type using default parameters. Available values:
@@ -69,11 +70,12 @@ class Unet(SegmentationModel):
69
70
**callable** and **None**. Default is **None**.
70
71
aux_params: Dictionary with parameters of the auxiliary output (classification head). Auxiliary output is build
71
72
on top of encoder if **aux_params** is not **None** (default). Supported params:
72
-
- classes (int): A number of classes
73
-
- pooling (str): One of "max", "avg". Default is "avg"
74
-
- dropout (float): Dropout factor in [0, 1)
75
-
- activation (str): An activation function to apply "sigmoid"/"softmax"
76
-
(could be **None** to return logits)
73
+
74
+
- classes (int): A number of classes
75
+
- pooling (str): One of "max", "avg". Default is "avg"
76
+
- dropout (float): Dropout factor in [0, 1)
77
+
- activation (str): An activation function to apply "sigmoid"/"softmax"
78
+
(could be **None** to return logits)
77
79
kwargs: Arguments passed to the encoder class ``__init__()`` function. Applies only to ``timm`` models. Keys with ``None`` values are pruned before passing.
Copy file name to clipboardExpand all lines: segmentation_models_pytorch/decoders/unetplusplus/model.py
+7-5Lines changed: 7 additions & 5 deletions
Original file line number
Diff line number
Diff line change
@@ -33,6 +33,7 @@ class UnetPlusPlus(SegmentationModel):
33
33
Length of the list should be the same as **encoder_depth**
34
34
decoder_use_norm: Specifies normalization between Conv2D and activation.
35
35
Accepts the following types:
36
+
36
37
- **True**: Defaults to `"batchnorm"`.
37
38
- **False**: No normalization (`nn.Identity`).
38
39
- **str**: Specifies normalization type using default parameters. Available values:
@@ -58,11 +59,12 @@ class UnetPlusPlus(SegmentationModel):
58
59
**callable** and **None**. Default is **None**.
59
60
aux_params: Dictionary with parameters of the auxiliary output (classification head). Auxiliary output is build
60
61
on top of encoder if **aux_params** is not **None** (default). Supported params:
61
-
- classes (int): A number of classes
62
-
- pooling (str): One of "max", "avg". Default is "avg"
63
-
- dropout (float): Dropout factor in [0, 1)
64
-
- activation (str): An activation function to apply "sigmoid"/"softmax"
65
-
(could be **None** to return logits)
62
+
63
+
- classes (int): A number of classes
64
+
- pooling (str): One of "max", "avg". Default is "avg"
65
+
- dropout (float): Dropout factor in [0, 1)
66
+
- activation (str): An activation function to apply "sigmoid"/"softmax"
67
+
(could be **None** to return logits)
66
68
kwargs: Arguments passed to the encoder class ``__init__()`` function. Applies only to ``timm`` models. Keys with ``None`` values are pruned before passing.
Copy file name to clipboardExpand all lines: segmentation_models_pytorch/decoders/upernet/model.py
+7-5Lines changed: 7 additions & 5 deletions
Original file line number
Diff line number
Diff line change
@@ -29,6 +29,7 @@ class UPerNet(SegmentationModel):
29
29
decoder_segmentation_channels: A number of convolution filters in segmentation blocks, default is 64
30
30
decoder_use_norm: Specifies normalization between Conv2D and activation.
31
31
Accepts the following types:
32
+
32
33
- **True**: Defaults to `"batchnorm"`.
33
34
- **False**: No normalization (`nn.Identity`).
34
35
- **str**: Specifies normalization type using default parameters. Available values:
@@ -50,11 +51,12 @@ class UPerNet(SegmentationModel):
50
51
**callable** and **None**. Default is **None**.
51
52
aux_params: Dictionary with parameters of the auxiliary output (classification head). Auxiliary output is build
52
53
on top of encoder if **aux_params** is not **None** (default). Supported params:
53
-
- classes (int): A number of classes
54
-
- pooling (str): One of "max", "avg". Default is "avg"
55
-
- dropout (float): Dropout factor in [0, 1)
56
-
- activation (str): An activation function to apply "sigmoid"/"softmax"
57
-
(could be **None** to return logits)
54
+
55
+
- classes (int): A number of classes
56
+
- pooling (str): One of "max", "avg". Default is "avg"
57
+
- dropout (float): Dropout factor in [0, 1)
58
+
- activation (str): An activation function to apply "sigmoid"/"softmax"
59
+
(could be **None** to return logits)
58
60
kwargs: Arguments passed to the encoder class ``__init__()`` function. Applies only to ``timm`` models. Keys with ``None`` values are pruned before passing.
0 commit comments