pub struct MultiGateResidualConfig {
pub d_model: usize,
pub n_stream: usize,
pub init_bias: f64,
}Expand description
Configuration for a single MultiGateResidual.
Fields§
§d_model: usizeModel width d.
n_stream: usizeNumber of parallel residual streams n.
init_bias: f64Initial value for every entry of the gate bias b⁽ᵝ⁾ (see module header).
Implementations§
Source§impl MultiGateResidualConfig
impl MultiGateResidualConfig
Sourcepub fn with_init_bias(self, init_bias: f64) -> Self
pub fn with_init_bias(self, init_bias: f64) -> Self
Sets the value for the field init_bias.
Initial value for every entry of the gate bias b⁽ᵝ⁾ (see module header).
- Defaults to
0.0
Source§impl MultiGateResidualConfig
impl MultiGateResidualConfig
Sourcepub fn init(&self, device: &Device) -> MultiGateResidual
pub fn init(&self, device: &Device) -> MultiGateResidual
Allocate one layer’s MGR parameters (w⁽ᵝ⁾, w⁽ᵅ⁾ zero; b⁽ᵝ⁾ constant).
Trait Implementations§
Source§impl Clone for MultiGateResidualConfig
impl Clone for MultiGateResidualConfig
Source§impl Config for MultiGateResidualConfig
impl Config for MultiGateResidualConfig
§fn load_binary(data: &[u8]) -> Result<Self, ConfigError>
fn load_binary(data: &[u8]) -> Result<Self, ConfigError>
Loads the configuration from a binary buffer. Read more
Source§impl Debug for MultiGateResidualConfig
impl Debug for MultiGateResidualConfig
Source§impl<'de> Deserialize<'de> for MultiGateResidualConfig
impl<'de> Deserialize<'de> for MultiGateResidualConfig
Source§fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>where
D: Deserializer<'de>,
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>where
D: Deserializer<'de>,
Deserialize this value from the given Serde deserializer. Read more
Source§impl Display for MultiGateResidualConfig
impl Display for MultiGateResidualConfig
Auto Trait Implementations§
impl Freeze for MultiGateResidualConfig
impl RefUnwindSafe for MultiGateResidualConfig
impl Send for MultiGateResidualConfig
impl Sync for MultiGateResidualConfig
impl Unpin for MultiGateResidualConfig
impl UnsafeUnpin for MultiGateResidualConfig
impl UnwindSafe for MultiGateResidualConfig
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