o 3wi@s@ddlmZGdddeZGdddeZGdddeZdS) )Optimizer2Statecs8eZdZ           dfd d ZZS)LAMBMbP?Tg?g+?:0yE>rF Nd?cs(tjd|||||| | | | | dd dS)a Base LAMB optimizer. Arguments: params (`torch.tensor`): The input parameters to optimize. lr (`float`, defaults to 1e-3): The learning rate. bias_correction (`bool`, defaults to `True`): Whether to apply bias correction to the first and second-order moments. betas (`tuple(float, float)`, defaults to (0.9, 0.999)): The beta values are the decay rates of the first and second-order moment of the optimizer. eps (`float`, defaults to 1e-8): The epsilon value prevents division by zero in the optimizer. weight_decay (`float`, defaults to 1e-2): The weight decay value for the optimizer. amsgrad (`bool`, defaults to `False`): Whether to use the [AMSGrad](https://hf.co/papers/1904.09237) variant of Adam that uses the maximum of past squared gradients instead. adam_w_mode (`bool`, defaults to `True`): Whether to use the AdamW variant. 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. max_unorm (`float`, defaults to 1.0): The maximum gradient norm. lambr  max_unormNsuper__init__)selfparamslrbias_correctionbetaseps weight_decayamsgrad adam_w_mode optim_bitsargs min_8bit_sizepercentile_clipping block_wiser  __class__T/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/bitsandbytes/optim/lamb.pyr s2 z LAMB.__init__) rTrrrFTrNrr Fr __name__ __module__ __qualname__r __classcell__r!r!rr"rsrc6eZdZ           d fd d ZZS)LAMB8bitrTrrrFNrr r c(tjd|||||d| | | | dd dS)a( 8-bit LAMB optimizer. Arguments: params (`torch.tensor`): The input parameters to optimize. lr (`float`, defaults to 1e-3): The learning rate. bias_correction (`bool`, defaults to `True`): Whether to apply bias correction to the first and second-order moments. betas (`tuple(float, float)`, defaults to (0.9, 0.999)): The beta values are the decay rates of the first and second-order moment of the optimizer. eps (`float`, defaults to 1e-8): The epsilon value prevents division by zero in the optimizer. weight_decay (`float`, defaults to 1e-2): The weight decay value for the optimizer. amsgrad (`bool`, defaults to `False`): Whether to use the [AMSGrad](https://hf.co/papers/1904.09237) variant of Adam that uses the maximum of past squared gradients instead. adam_w_mode (`bool`, defaults to `True`): Whether to use the AdamW variant. 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. max_unorm (`float`, defaults to 1.0): The maximum gradient norm. r r r Nrrrrrrrrrrrrrrr rr!r"rL/ zLAMB8bit.__init__ rTrrrFTNrr Fr r#r!r!rr"r)Kr)cr() LAMB32bitrTrrrFNrr r cr*)a) 32-bit LAMB optimizer. Arguments: params (`torch.tensor`): The input parameters to optimize. lr (`float`, defaults to 1e-3): The learning rate. bias_correction (`bool`, defaults to `True`): Whether to apply bias correction to the first and second-order moments. betas (`tuple(float, float)`, defaults to (0.9, 0.999)): The beta values are the decay rates of the first and second-order moment of the optimizer. eps (`float`, defaults to 1e-8): The epsilon value prevents division by zero in the optimizer. weight_decay (`float`, defaults to 1e-2): The weight decay value for the optimizer. amsgrad (`bool`, defaults to `False`): Whether to use the [AMSGrad](https://hf.co/papers/1904.09237) variant of Adam that uses the maximum of past squared gradients instead. adam_w_mode (`bool`, defaults to `True`): Whether to use the AdamW variant. 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. max_unorm (`float`, defaults to 1.0): The maximum gradient norm. r rr r Nrr,rr!r"rr-zLAMB32bit.__init__r.r#r!r!rr"r0r/r0N)bitsandbytes.optim.optimizerrrr)r0r!r!r!r"s C@