cerebras.modelzoo.losses.DPOLoss.DPOLoss

class cerebras.modelzoo.losses.DPOLoss.DPOLoss

Bases: torch.nn.Module

DPO Loss :param beta: Temperature parameter for the DPO loss, typically something in the range of 0.1 to 0.5.

We ignore the reference model as beta -> 0.

Parameters

reference_free – If True, we ignore the _provided_ reference model and implicitly use a reference model that assigns equal probability to all responses.

Methods

forward

__init__(beta=0.1, loss_type='sigmoid', reference_free=False)

__call__(*args: Any, **kwargs: Any) → Any

Call self as a function.

static __new__(cls, *args: Any, **kwargs: Any) → Any

previous

cerebras.modelzoo.losses.DPOLoss

next

cerebras.modelzoo.losses.GPTLMHeadModelLoss