Struct TrainingCheckpoint 

pub struct TrainingCheckpoint {
    pub training_state: TrainingState,
    pub optimizer_state: Option<OptimizerCheckpoint>,
    pub grad_scaler_state: Option<GradScalerCheckpoint>,
}

Everything besides the adapter weights needed to resume a run.

Fields

training_state: TrainingState

Progress and settings.

optimizer_state: Option<OptimizerCheckpoint>

Optimizer moments keyed by parameter name.

grad_scaler_state: Option<GradScalerCheckpoint>

Loss-scaler state, if a scaler was used.

Implementations

impl TrainingCheckpoint

pub fn save<P: AsRef<Path>>(&self, path: P) -> Result<(), Box<dyn Error>>

Write as JSON (via a temporary file, so a crash never leaves a truncated checkpoint).

pub fn load<P: AsRef<Path>>(path: P) -> Result<Self, Box<dyn Error>>

Read a checkpoint written by save.

Trait Implementations

impl Clone for TrainingCheckpoint

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TrainingCheckpoint

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for TrainingCheckpoint

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for TrainingCheckpoint

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

where __S: Serializer,

Serialize this value into the given Serde serializer.