experanto.utils.SessionSpecificSampler

class SessionSpecificSampler(*args, **kwargs)

Bases: Sampler

A batch sampler specific to a single session that efficiently generates batches from the session’s indices.

Methods

__init__(indices, batch_size[, drop_last, ...])	Initialize session-specific sampler.
get_state()	Return the state of the sampler (including RNG state).
set_position(position)	Set the current batch position.
set_state(state)	Restore the state of the sampler (including RNG state).

__init__(indices, batch_size, drop_last=False, shuffle=False, seed=None)

Initialize session-specific sampler.

Parameters:

• indices (list) – Dataset indices belonging to this session.

• batch_size (int) – Number of samples per batch.

• drop_last (bool, optional) – Whether to drop the last batch if smaller than batch_size. Default is False.

• shuffle (bool, optional) – Whether to shuffle indices. Default is False.

• seed (int, optional) – Random seed for reproducibility.

__len__()

Return the number of batches.

set_position(position)

Set the current batch position.

get_state()

Return the state of the sampler (including RNG state).

set_state(state)

Restore the state of the sampler (including RNG state).

__iter__()

Yield batches of indices starting from the current position.