;i=RSSKJr "SS\5r"SS\5r"SS\5rg) )Optimizer1Statec>^\rSrSrSU4SjjrSrU=r$)SGDc \>US:Xa [S5e[T U] SUUX44SUUUU U U 5 g)a Base SGD optimizer. Arguments: params (`torch.tensor`): The input parameters to optimize. lr (`float`): The learning rate. momentum (`float`, defaults to 0): The momentum value speeds up the optimizer by taking bigger steps. dampening (`float`, defaults to 0): The dampening value reduces the momentum of the optimizer. weight_decay (`float`, defaults to 0.0): The weight decay value for the optimizer. nesterov (`bool`, defaults to `False`): Whether to use Nesterov momentum. optim_bits (`int`, defaults to 32): The number of bits of the optimizer state. args (`object`, defaults to `None`): An object with additional arguments. min_8bit_size (`int`, defaults to 4096): The minimum number of elements of the parameter tensors for 8-bit optimization. percentile_clipping (`int`, defaults to 100): Adapts clipping threshold automatically by tracking the last 100 gradient norms and clipping the gradient at a certain percentile to improve stability. block_wise (`bool`, defaults to `True`): Whether to independently quantize each block of tensors to reduce outlier effects and improve stability. r&SGD without momentum is not supported!momentumNNotImplementedErrorsuper__init__) selfparamslrr dampening weight_decaynesterov optim_bitsargs min_8bit_sizepercentile_clipping block_wise __class__s f/home/dmtnaga/Documents/work/airagagent/rag_env/lib/python3.13/site-packages/bitsandbytes/optim/sgd.pyr SGD.__init__ sJR q=%&NO O      !        ) rrrF NdT__name__ __module__ __qualname____firstlineno__r__static_attributes__ __classcell__rs@rrrs*  7 7 rrc<^\rSrSrSU4SjjrSrU=r$)SGD8bitCc \>US:Xa [S5e[T U] SUUX44SUSUUU U 5 g)as 8-bit SGD optimizer. Arguments: params (`torch.tensor`): The input parameters to optimize. lr (`float`): The learning rate. momentum (`float`, defaults to 0): The momentum value speeds up the optimizer by taking bigger steps. dampening (`float`, defaults to 0): The dampening value reduces the momentum of the optimizer. weight_decay (`float`, defaults to 0.0): The weight decay value for the optimizer. nesterov (`bool`, defaults to `False`): Whether to use Nesterov momentum. args (`object`, defaults to `None`): An object with additional arguments. min_8bit_size (`int`, defaults to 4096): The minimum number of elements of the parameter tensors for 8-bit optimization. percentile_clipping (`int`, defaults to 100): Adapts clipping threshold automatically by tracking the last 100 gradient norms and clipping the gradient at a certain percentile to improve stability. block_wise (`bool`, defaults to `True`): Whether to independently quantize each block of tensors to reduce outlier effects and improve stability. rrr r rNr rrrr rrrrrrrrs rrSGD8bit.__init__DsJL q=%&NO O      !       rrrrrFNr r!Tr"r)s@rr+r+C'  4 4 rr+c<^\rSrSrSU4SjjrSrU=r$)SGD32bit{c \>US:Xa [S5e[T U] SUUX44SUSUUU U 5 g)at 32-bit SGD optimizer. Arguments: params (`torch.tensor`): The input parameters to optimize. lr (`float`): The learning rate. momentum (`float`, defaults to 0): The momentum value speeds up the optimizer by taking bigger steps. dampening (`float`, defaults to 0): The dampening value reduces the momentum of the optimizer. weight_decay (`float`, defaults to 0.0): The weight decay value for the optimizer. nesterov (`bool`, defaults to `False`): Whether to use Nesterov momentum. args (`object`, defaults to `None`): An object with additional arguments. min_8bit_size (`int`, defaults to 4096): The minimum number of elements of the parameter tensors for 8-bit optimization. percentile_clipping (`int`, defaults to 100): Adapts clipping threshold automatically by tracking the last 100 gradient norms and clipping the gradient at a certain percentile to improve stability. block_wise (`bool`, defaults to `True`): Whether to independently quantize each block of tensors to reduce outlier effects and improve stability. rrr r rNr r.s rrSGD32bit.__init__|sJL q=%&NO O      !        rrr0r"r)s@rr3r3{r1rr3N)bitsandbytes.optim.optimizerrrr+r3rrrr8s0 98 /8 v5 o5 p5 5 r