neurai.finetune.tuning.lorax package# Submodules# class neurai.finetune.tuning.lorax.transform.LoraWeight(w, a, b, alpha=1.0, shape=None, dtype=None)# Bases: ImplicitArray The weight of lora node. Parameters: w (ArrayValue) – The param of node. a (ArrayValue) – The low-rank matrix. b (ArrayValue) – Another low-rank matrix alpha (float) – The scaling factor of low-rank matrix shape (Optional[Sequence[Union[int, Any]]]) – dtype (Union[Any, str, dtype, SupportsDType, None]) –
