Struct TrainingHyperparams 

pub struct TrainingHyperparams {
    pub epochs: u32,
    pub batch_size: u32,
    pub learning_rate: f64,
    pub warmup_steps: u64,
    pub weight_decay: f64,
    pub lr_scheduler: LrScheduler,
    pub seed: u64,
    pub max_seq_len: usize,
    pub gradient_accumulation_steps: u32,
    pub max_grad_norm: f64,
}

Training hyperparameters.

Fields

epochs: u32

Number of training epochs.

batch_size: u32

Batch size per device.

learning_rate: f64

Initial learning rate.

warmup_steps: u64

Warmup steps for LR scheduler.

weight_decay: f64

Weight decay factor.

lr_scheduler: LrScheduler

Learning rate scheduler type.

seed: u64

Random seed for reproducibility.

max_seq_len: usize

Maximum sequence length (tokens).

gradient_accumulation_steps: u32

Gradient accumulation steps (effective batch = batch_size * grad_accum).

max_grad_norm: f64

Maximum gradient norm for clipping.

Trait Implementations

impl Clone for TrainingHyperparams

fn clone(&self) -> TrainingHyperparams

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TrainingHyperparams

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

Formats the value using the given formatter.

impl Default for TrainingHyperparams

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for TrainingHyperparams

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for TrainingHyperparams

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

where __S: Serializer,

Serialize this value into the given Serde serializer.