Struct candle_transformers::models::stable_diffusion::ddim::DDIMSchedulerConfig

pub struct DDIMSchedulerConfig {
    pub beta_start: f64,
    pub beta_end: f64,
    pub beta_schedule: BetaSchedule,
    pub eta: f64,
    pub steps_offset: usize,
    pub prediction_type: PredictionType,
    pub train_timesteps: usize,
    pub timestep_spacing: TimestepSpacing,
}

The configuration for the DDIM scheduler.

Fields

beta_start: f64

The value of beta at the beginning of training.

beta_end: f64

The value of beta at the end of training.

beta_schedule: BetaSchedule

How beta evolved during training.

eta: f64

The amount of noise to be added at each step.

steps_offset: usize

Adjust the indexes of the inference schedule by this value.

prediction_type: PredictionType

prediction type of the scheduler function, one of epsilon (predicting the noise of the diffusion process), sample (directly predicting the noisy sample) or v_prediction` (see section 2.4 https://imagen.research.google/video/paper.pdf)

train_timesteps: usize

number of diffusion steps used to train the model

timestep_spacing: TimestepSpacing

time step spacing for the diffusion process

Trait Implementations

impl Clone for DDIMSchedulerConfig

fn clone(&self) -> DDIMSchedulerConfig

Returns a copy of the value.

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

Performs copy-assignment from source.

impl Debug for DDIMSchedulerConfig

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

Formats the value using the given formatter.

impl Default for DDIMSchedulerConfig

fn default() -> Self

Returns the “default value” for a type.

impl SchedulerConfig for DDIMSchedulerConfig

fn build(&self, inference_steps: usize) -> Result<Box<dyn Scheduler>>

impl Copy for DDIMSchedulerConfig