cached_eval

hybrid_learning.experimentation.fuzzy_exp.fuzzy_exp_helpers.cached_eval(descriptors, input_batch, model, cache=None, device=None)

Given descriptors load model outputs either from cache or generate them from input_batch.

Parameters

• descriptors (List[str]) –

• input_batch (Any) –

• model (Callable[[Any], Union[Tensor, Dict[str, Tensor]]]) –

• cache (Optional[Cache]) –

• device (Optional[Union[str, device]]) –