pub struct ConvShape {
pub batch: usize,
pub c_in: usize,
pub c_out: usize,
pub l_in: usize,
pub k: usize,
pub stride: usize,
pub dilation: usize,
pub pad_left: usize,
pub pad_right: usize,
pub groups: usize,
}Expand description
Geometry of a ternary 1-D convolution Y[B, C_out, L_out] = scale ⊙ conv1d(X, W) — the codec’s
conv op (ADR 0030). The weight is packed 2-D [C_out, (C_in/groups)·K] (the per-output-channel
ternary reshape), so k_g() is the matmul contraction and n_g() the per-group output count.
Fields§
§batch: usizeBatch size.
c_in: usizeInput channels (divisible by groups).
c_out: usizeOutput channels (divisible by groups).
l_in: usizeInput length.
k: usizeKernel size.
stride: usizeStride (≥ 1).
dilation: usizeDilation (≥ 1).
pad_left: usizeLeft zero-padding.
pad_right: usizeRight zero-padding.
groups: usizeConvolution groups (≥ 1).
Implementations§
Trait Implementations§
impl Copy for ConvShape
impl Eq for ConvShape
impl StructuralPartialEq for ConvShape
Auto Trait Implementations§
impl Freeze for ConvShape
impl RefUnwindSafe for ConvShape
impl Send for ConvShape
impl Sync for ConvShape
impl Unpin for ConvShape
impl UnsafeUnpin for ConvShape
impl UnwindSafe for ConvShape
Blanket Implementations§
Source§impl<T> BorrowMut<T> for Twhere
T: ?Sized,
impl<T> BorrowMut<T> for Twhere
T: ?Sized,
Source§fn borrow_mut(&mut self) -> &mut T
fn borrow_mut(&mut self) -> &mut T
Mutably borrows from an owned value. Read more