rl4lms.core_components package

Submodules

rl4lms.core_components.sampler module

class rl4lms.core_components.sampler.PrioritySampler(max_size: int | None = None, priority_scale: float = 0.0)[source]

Bases: object

__init__(max_size: int | None = None, priority_scale: float = 0.0)[source]

Creates a priority sampler

Parameters:

max_size (int) – maximum size of the queue
priority_scale (float) – 0.0 is a pure uniform sampling, 1.0 is completely priority sampling

add(item: Any, priority: float)[source]

sample(size: int) → List[Any][source]

update(item: Any, priority: float)[source]

get_all_samples() → List[Any][source]

rl4lms.core_components.sweep module

rl4lms.core_components.sweep.get_dict_obj(keys: List, values: List) → Dict[source]

rl4lms.core_components.sweep.find_products(splits_by_keys: Dict) → List[Dict][source]

rl4lms.core_components.sweep.to_expand(obj: Any) → bool[source]

rl4lms.core_components.sweep.split_config(obj: Dict) → List[Dict][source]: Recursively splits the given object

rl4lms.core_components.sweep.dict_hash(dictionary: Dict[str, Any]) → str[source]: MD5 hash of a dictionary.

Module contents