megenginelite-sys 1.8.2

A safe megenginelite wrapper in Rust
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142

# MegEngine is Licensed under the Apache License, Version 2.0 (the "License")
#
# Copyright (c) 2014-2021 Megvii Inc. All rights reserved.
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT ARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
import numpy as np

import megengine.functional as F
from megengine import Parameter

from .init import ones_, zeros_
from .module import Module


class GroupNorm(Module):
    """Simple implementation of GroupNorm. Only support 4d tensor now.
    Reference: https://arxiv.org/pdf/1803.08494.pdf.
    """

    def __init__(self, num_groups, num_channels, eps=1e-5, affine=True, **kwargs):
        super().__init__(**kwargs)
        assert num_channels % num_groups == 0
        self.num_groups = num_groups
        self.num_channels = num_channels
        self.eps = eps
        self.affine = affine
        if self.affine:
            self.weight = Parameter(np.ones(num_channels, dtype=np.float32))
            self.bias = Parameter(np.zeros(num_channels, dtype=np.float32))
        else:
            self.weight = None
            self.bias = None
        self.reset_parameters()

    def reset_parameters(self):
        if self.affine:
            ones_(self.weight)
            zeros_(self.bias)

    def forward(self, x):
        N, C, H, W = x.shape
        assert C == self.num_channels

        x = x.reshape(N, self.num_groups, -1)
        mean = x.mean(axis=2, keepdims=True)
        var = (x * x).mean(axis=2, keepdims=True) - mean * mean

        x = (x - mean) / F.sqrt(var + self.eps)
        x = x.reshape(N, C, H, W)
        if self.affine:
            x = self.weight.reshape(1, -1, 1, 1) * x + self.bias.reshape(1, -1, 1, 1)

        return x

    def _module_info_string(self) -> str:
        s = (
            "groups={num_groups}, channels={num_channels}, "
            "eps={eps}, affine={affine}"
        )
        return s.format(**self.__dict__)


class InstanceNorm(Module):
    """Simple implementation of InstanceNorm. Only support 4d tensor now.
    Reference: https://arxiv.org/abs/1607.08022.
    Note that InstanceNorm equals using GroupNome with num_groups=num_channels.
    """

    def __init__(self, num_channels, eps=1e-05, affine=True, **kwargs):
        super().__init__(**kwargs)
        self.num_channels = num_channels
        self.eps = eps
        self.affine = affine
        if self.affine:
            self.weight = Parameter(np.ones(num_channels, dtype="float32"))
            self.bias = Parameter(np.zeros(num_channels, dtype="float32"))
        else:
            self.weight = None
            self.bias = None
        self.reset_parameters()

    def reset_parameters(self):
        if self.affine:
            ones_(self.weight)
            zeros_(self.bias)

    def forward(self, x):
        N, C, H, W = x.shape
        assert C == self.num_channels
        x = x.reshape(N, C, -1)
        mean = x.mean(axis=2, keepdims=True)
        var = (x ** 2).mean(axis=2, keepdims=True) - mean * mean

        x = (x - mean) / F.sqrt(var + self.eps)
        x = x.reshape(N, C, H, W)
        if self.affine:
            x = self.weight.reshape(1, -1, 1, 1) * x + self.bias.reshape(1, -1, 1, 1)

        return x

    def _module_info_string(self) -> str:
        s = "channels={num_channels}, eps={eps}, affine={affine}"
        return s.format(**self.__dict__)


class LayerNorm(Module):
    """Simple implementation of LayerNorm. Support tensor of any shape as input.
    Reference: https://arxiv.org/pdf/1803.08494.pdf.
    """

    def __init__(self, normalized_shape, eps=1e-05, affine=True, **kwargs):
        super().__init__(**kwargs)
        if isinstance(normalized_shape, int):
            normalized_shape = (normalized_shape,)
        self.normalized_shape = tuple(normalized_shape)
        self.eps = eps
        self.affine = affine
        if self.affine:
            self.weight = Parameter(np.ones(self.normalized_shape, dtype="float32"))
            self.bias = Parameter(np.zeros(self.normalized_shape, dtype="float32"))
        else:
            self.weight = None
            self.bias = None

        self.reset_parameters()

    def reset_parameters(self):
        if self.affine:
            ones_(self.weight)
            zeros_(self.bias)

    def forward(self, x):
        x = F.nn.layer_norm(
            x, self.normalized_shape, self.affine, self.weight, self.bias, self.eps
        )
        return x

    def _module_info_string(self) -> str:
        s = "normalized_shape={normalized_shape}, eps={eps}, affine={affine}"
        return s.format(**self.__dict__)