darknet 0.4.0

A Rust wrapper for Darknet, an open source neural network framework written in C and CUDA.
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

use darknet_sys as sys;
use num_derive::FromPrimitive;

/// Layer types.
#[repr(usize)]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, FromPrimitive)]
pub enum LayerType {
    Gru = sys::LAYER_TYPE_GRU as usize,
    Rnn = sys::LAYER_TYPE_RNN as usize,
    Sam = sys::LAYER_TYPE_SAM as usize,
    Cost = sys::LAYER_TYPE_COST as usize,
    Crnn = sys::LAYER_TYPE_CRNN as usize,
    Crop = sys::LAYER_TYPE_CROP as usize,
    Iseg = sys::LAYER_TYPE_ISEG as usize,
    Lstm = sys::LAYER_TYPE_LSTM as usize,
    Xnor = sys::LAYER_TYPE_XNOR as usize,
    Yolo = sys::LAYER_TYPE_YOLO as usize,
    Blank = sys::LAYER_TYPE_BLANK as usize,
    Empty = sys::LAYER_TYPE_EMPTY as usize,
    Local = sys::LAYER_TYPE_LOCAL as usize,
    Reorg = sys::LAYER_TYPE_REORG as usize,
    Route = sys::LAYER_TYPE_ROUTE as usize,
    Active = sys::LAYER_TYPE_ACTIVE as usize,
    L2Norm = sys::LAYER_TYPE_L2NORM as usize,
    Region = sys::LAYER_TYPE_REGION as usize,
    Avgpool = sys::LAYER_TYPE_AVGPOOL as usize,
    Dropout = sys::LAYER_TYPE_DROPOUT as usize,
    Logxent = sys::LAYER_TYPE_LOGXENT as usize,
    Maxpool = sys::LAYER_TYPE_MAXPOOL as usize,
    Network = sys::LAYER_TYPE_NETWORK as usize,
    Softmax = sys::LAYER_TYPE_SOFTMAX as usize,
    Shortcut = sys::LAYER_TYPE_SHORTCUT as usize,
    Upsample = sys::LAYER_TYPE_UPSAMPLE as usize,
    Batchnorm = sys::LAYER_TYPE_BATCHNORM as usize,
    Connected = sys::LAYER_TYPE_CONNECTED as usize,
    ConvLstm = sys::LAYER_TYPE_CONV_LSTM as usize,
    Detection = sys::LAYER_TYPE_DETECTION as usize,
    ReorgOld = sys::LAYER_TYPE_REORG_OLD as usize,
    Convolutional = sys::LAYER_TYPE_CONVOLUTIONAL as usize,
    GaussianYolo = sys::LAYER_TYPE_GAUSSIAN_YOLO as usize,
    LocalAvgpool = sys::LAYER_TYPE_LOCAL_AVGPOOL as usize,
    Normalization = sys::LAYER_TYPE_NORMALIZATION as usize,
    ScaleChannels = sys::LAYER_TYPE_SCALE_CHANNELS as usize,
    Deconvolutional = sys::LAYER_TYPE_DECONVOLUTIONAL as usize,
}

/// Activation types.
#[repr(usize)]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, FromPrimitive)]
pub enum Activation {
    Elu = sys::ACTIVATION_ELU as usize,
    Gelu = sys::ACTIVATION_GELU as usize,
    Mish = sys::ACTIVATION_MISH as usize,
    Plse = sys::ACTIVATION_PLSE as usize,
    Ramp = sys::ACTIVATION_RAMP as usize,
    Relu = sys::ACTIVATION_RELU as usize,
    Selu = sys::ACTIVATION_SELU as usize,
    Tanh = sys::ACTIVATION_TANH as usize,
    Leaky = sys::ACTIVATION_LEAKY as usize,
    Lhtan = sys::ACTIVATION_LHTAN as usize,
    Loggy = sys::ACTIVATION_LOGGY as usize,
    Relie = sys::ACTIVATION_RELIE as usize,
    Relu6 = sys::ACTIVATION_RELU6 as usize,
    Stair = sys::ACTIVATION_STAIR as usize,
    Swish = sys::ACTIVATION_SWISH as usize,
    Linear = sys::ACTIVATION_LINEAR as usize,
    Hardtan = sys::ACTIVATION_HARDTAN as usize,
    Logistic = sys::ACTIVATION_LOGISTIC as usize,
    NormChan = sys::ACTIVATION_NORM_CHAN as usize,
    NormChanSoftmax = sys::ACTIVATION_NORM_CHAN_SOFTMAX as usize,
    NormChanSoftmaxMaxval = sys::ACTIVATION_NORM_CHAN_SOFTMAX_MAXVAL as usize,
}

/// Binary activation types.
#[repr(usize)]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, FromPrimitive)]
pub enum BinaryActivation {
    Add = sys::BINARY_ACTIVATION_ADD as usize,
    Div = sys::BINARY_ACTIVATION_DIV as usize,
    Sub = sys::BINARY_ACTIVATION_SUB as usize,
    Mult = sys::BINARY_ACTIVATION_MULT as usize,
}

/// Cost, or namely, loss types.
#[repr(usize)]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, FromPrimitive)]
pub enum CostType {
    L1 = sys::COST_TYPE_L1 as usize,
    Seg = sys::COST_TYPE_SEG as usize,
    Sse = sys::COST_TYPE_SSE as usize,
    Wgan = sys::COST_TYPE_WGAN as usize,
    Masked = sys::COST_TYPE_MASKED as usize,
    Smooth = sys::COST_TYPE_SMOOTH as usize,
}

/// Weights format types.
#[repr(usize)]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, FromPrimitive)]
pub enum WeightsType {
    NoWeights = sys::WEIGHTS_TYPE_T_NO_WEIGHTS as usize,
    PerChannel = sys::WEIGHTS_TYPE_T_PER_CHANNEL as usize,
    PerFeature = sys::WEIGHTS_TYPE_T_PER_FEATURE as usize,
}

/// Weights normalization types.
#[repr(usize)]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, FromPrimitive)]
pub enum WeightsNormalizion {
    None = sys::WEIGHTS_NORMALIZATION_T_NO_NORMALIZATION as usize,
    Relu = sys::WEIGHTS_NORMALIZATION_T_RELU_NORMALIZATION as usize,
    Softmax = sys::WEIGHTS_NORMALIZATION_T_SOFTMAX_NORMALIZATION as usize,
}

/// Non-maximum suppression (NMS) types.
#[repr(usize)]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, FromPrimitive)]
pub enum NmsKind {
    Diou = sys::NMS_KIND_DIOU_NMS as usize,
    Greedy = sys::NMS_KIND_GREEDY_NMS as usize,
    Corners = sys::NMS_KIND_CORNERS_NMS as usize,
    Default = sys::NMS_KIND_DEFAULT_NMS as usize,
}

/// IoU loss types.
#[repr(usize)]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, FromPrimitive)]
pub enum IoULoss {
    IoU = sys::IOU_LOSS_IOU as usize,
    Mse = sys::IOU_LOSS_MSE as usize,
    CIoU = sys::IOU_LOSS_CIOU as usize,
    DIoU = sys::IOU_LOSS_DIOU as usize,
    GIoU = sys::IOU_LOSS_GIOU as usize,
}

/// YOLO point types.
#[repr(usize)]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, FromPrimitive)]
pub enum YoloPoint {
    Center = sys::YOLO_POINT_YOLO_CENTER as usize,
    LeftTop = sys::YOLO_POINT_YOLO_LEFT_TOP as usize,
    RightBottom = sys::YOLO_POINT_YOLO_RIGHT_BOTTOM as usize,
}