Enum LayerCheckpoint 

pub enum LayerCheckpoint {
    Linear {
        in_features: usize,
        out_features: usize,
        weight: TensorSnapshot,
        bias: TensorSnapshot,
    },
    ReLU,
    LeakyReLU {
        negative_slope: f32,
    },
    Sigmoid,
    Tanh,
    Dropout {
        rate: f32,
    },
    Conv2d {
        in_channels: usize,
        out_channels: usize,
        kernel_h: usize,
        kernel_w: usize,
        stride_h: usize,
        stride_w: usize,
        weight: TensorSnapshot,
        bias: Option<TensorSnapshot>,
    },
    BatchNorm2d {
        num_features: usize,
        epsilon: f32,
        gamma: TensorSnapshot,
        beta: TensorSnapshot,
        running_mean: TensorSnapshot,
        running_var: TensorSnapshot,
    },
    MaxPool2d {
        kernel_h: usize,
        kernel_w: usize,
        stride_h: usize,
        stride_w: usize,
    },
    AvgPool2d {
        kernel_h: usize,
        kernel_w: usize,
        stride_h: usize,
        stride_w: usize,
    },
    Flatten,
    GlobalAvgPool2d,
    Softmax,
    Embedding {
        num_embeddings: usize,
        embedding_dim: usize,
        weight: TensorSnapshot,
    },
    LayerNorm {
        normalized_shape: usize,
        eps: f32,
        gamma: TensorSnapshot,
        beta: TensorSnapshot,
    },
    GroupNorm {
        num_groups: usize,
        num_channels: usize,
        eps: f32,
        gamma: TensorSnapshot,
        beta: TensorSnapshot,
    },
    DepthwiseConv2d {
        channels: usize,
        kernel_h: usize,
        kernel_w: usize,
        stride_h: usize,
        stride_w: usize,
        weight: TensorSnapshot,
        bias: Option<TensorSnapshot>,
    },
    SeparableConv2d {
        in_channels: usize,
        out_channels: usize,
        kernel_h: usize,
        kernel_w: usize,
        stride_h: usize,
        stride_w: usize,
        depthwise_weight: TensorSnapshot,
        pointwise_weight: TensorSnapshot,
        bias: Option<TensorSnapshot>,
    },
}

Serializable layer checkpoint payload.

Variants

Linear

Fields

in_features: usize

out_features: usize

weight: TensorSnapshot

bias: TensorSnapshot

ReLU

LeakyReLU

Fields

negative_slope: f32

Sigmoid

Tanh

Dropout

Fields

rate: f32

Conv2d

Fields

in_channels: usize

out_channels: usize

kernel_h: usize

kernel_w: usize

stride_h: usize

stride_w: usize

weight: TensorSnapshot

bias: Option<TensorSnapshot>

BatchNorm2d

Fields

num_features: usize

epsilon: f32

gamma: TensorSnapshot

beta: TensorSnapshot

running_mean: TensorSnapshot

running_var: TensorSnapshot

MaxPool2d

Fields

kernel_h: usize

kernel_w: usize

stride_h: usize

stride_w: usize

AvgPool2d

Fields

kernel_h: usize

kernel_w: usize

stride_h: usize

stride_w: usize

Flatten

GlobalAvgPool2d

Softmax

Embedding

Fields

num_embeddings: usize

embedding_dim: usize

weight: TensorSnapshot

LayerNorm

Fields

normalized_shape: usize

eps: f32

gamma: TensorSnapshot

beta: TensorSnapshot

GroupNorm

Fields

num_groups: usize

num_channels: usize

eps: f32

gamma: TensorSnapshot

beta: TensorSnapshot

DepthwiseConv2d

Fields

channels: usize

kernel_h: usize

kernel_w: usize

stride_h: usize

stride_w: usize

weight: TensorSnapshot

bias: Option<TensorSnapshot>

SeparableConv2d

Fields

in_channels: usize

out_channels: usize

kernel_h: usize

kernel_w: usize

stride_h: usize

stride_w: usize

depthwise_weight: TensorSnapshot

pointwise_weight: TensorSnapshot

bias: Option<TensorSnapshot>

Trait Implementations

impl Clone for LayerCheckpoint

fn clone(&self) -> LayerCheckpoint

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for LayerCheckpoint

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for LayerCheckpoint

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for LayerCheckpoint

fn eq(&self, other: &LayerCheckpoint) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for LayerCheckpoint

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for LayerCheckpoint