i+VddlmZddlmZmZddlmZmZmZmZm Z m Z m Z ddgZ GddeZ dd e d e d ed ed ed ed ze _ d$deedeedeedeedeedeeded zdededed zded zded zdededeezdeezdeezded ed!ed"d f*d#Zy )%)Tensor)Adamadam)_capturable_doc_differentiable_doc _foreach_doc _fused_doc _maximize_doc _params_docParamsTAdamWadamwceZdZ ddddddddedeezdeeezeezfdeded ed ed edzd ed ededzddffdZfdZ xZ S)rFN)maximizeforeach capturabledifferentiablefusedparamslrbetaseps weight_decayamsgradrrrrrreturnc<t |||||||||| | | d y)NT)rrrrrdecoupled_weight_decay)super__init__) selfrrrrrrrrrrr __class__s f/Volumes/fast/ai/experiments/voice-extract-mac/.venv/lib/python3.12/site-packages/torch/optim/adamw.pyr zAdamW.__init__s;       !)#'  cPt|||jD]}d|d< y)NTr)r __setstate__ param_groups)r!stategroupr"s r#r&zAdamW.__setstate__6s+ U#&&E.2E* +'r$)gMbP?)g?g+?g:0yE>g{Gz?F) __name__ __module__ __qualname__r floatrtupleboolr r& __classcell__)r"s@r#rrs"7C" # $!   FN UV^UV^34           d{   B33r$a Implements AdamW algorithm, where weight decay does not accumulate in the momentum nor variance. .. math:: \begin{aligned} &\rule{110mm}{0.4pt} \\ &\textbf{input} : \gamma \text{(lr)}, \: \beta_1, \beta_2 \text{(betas)}, \: \theta_0 \text{(params)}, \: f(\theta) \text{(objective)}, \: \epsilon \text{ (epsilon)} \\ &\hspace{13mm} \lambda \text{(weight decay)}, \: \textit{amsgrad}, \: \textit{maximize} \\ &\textbf{initialize} : m_0 \leftarrow 0 \text{ (first moment)}, v_0 \leftarrow 0 \text{ ( second moment)}, \: v_0^{max}\leftarrow 0 \\[-1.ex] &\rule{110mm}{0.4pt} \\ &\textbf{for} \: t=1 \: \textbf{to} \: \ldots \: \textbf{do} \\ &\hspace{5mm}\textbf{if} \: \textit{maximize}: \\ &\hspace{10mm}g_t \leftarrow -\nabla_{\theta} f_t (\theta_{t-1}) \\ &\hspace{5mm}\textbf{else} \\ &\hspace{10mm}g_t \leftarrow \nabla_{\theta} f_t (\theta_{t-1}) \\ &\hspace{5mm} \theta_t \leftarrow \theta_{t-1} - \gamma \lambda \theta_{t-1} \\ &\hspace{5mm}m_t \leftarrow \beta_1 m_{t-1} + (1 - \beta_1) g_t \\ &\hspace{5mm}v_t \leftarrow \beta_2 v_{t-1} + (1-\beta_2) g^2_t \\ &\hspace{5mm}\widehat{m_t} \leftarrow m_t/\big(1-\beta_1^t \big) \\ &\hspace{5mm}\textbf{if} \: amsgrad \\ &\hspace{10mm} v_t^{max} \leftarrow \mathrm{max}(v_{t-1}^{max},v_t) \\ &\hspace{10mm}\widehat{v_t} \leftarrow v_t^{max}/\big(1-\beta_2^t \big) \\ &\hspace{5mm}\textbf{else} \\ &\hspace{10mm}\widehat{v_t} \leftarrow v_t/\big(1-\beta_2^t \big) \\ &\hspace{5mm}\theta_t \leftarrow \theta_t - \gamma \widehat{m_t}/ \big(\sqrt{\widehat{v_t}} + \epsilon \big) \\ &\rule{110mm}{0.4pt} \\[-1.ex] &\bf{return} \: \theta_t \\[-1.ex] &\rule{110mm}{0.4pt} \\[-1.ex] \end{aligned} For further details regarding the algorithm we refer to `Decoupled Weight Decay Regularization`_. z Args: ao lr (float, Tensor, optional): learning rate (default: 1e-3). A tensor LR is not yet supported for all our implementations. Please use a float LR if you are not also specifying fused=True or capturable=True. betas (tuple[Union[float, Tensor], Union[float, Tensor]], optional): coefficients used for computing running averages of gradient and its square. If a tensor is provided, must be 1-element. (default: (0.9, 0.999)) eps (float, optional): term added to the denominator to improve numerical stability (default: 1e-8) weight_decay (float, optional): weight decay coefficient (default: 1e-2) amsgrad (bool, optional): whether to use the AMSGrad variant of this algorithm from the paper `On the Convergence of Adam and Beyond`_ (default: False) z a8 .. Note:: A prototype implementation of Adam and AdamW for MPS supports `torch.float32` and `torch.float16`. .. _Decoupled Weight Decay Regularization: https://arxiv.org/abs/1711.05101 .. _On the Convergence of Adam and Beyond: https://openreview.net/forum?id=ryQu7f-RZ Nrgradsexp_avgs exp_avg_sqsmax_exp_avg_sqs state_stepsrrrr grad_scale found_inf has_complexrbeta1beta2rrrrrc Bt||||||f|||| | | | | ||||||ddy)zpFunctional API that performs AdamW algorithm computation. See :class:`~torch.optim.AdamW` for details. T)rrrrr6r7r8rr9r:rrrrrN)r)rr1r2r3r4r5rrrrr6r7r8rr9r:rrrrs r#rrsR:   % ! #+r$)NFFNNNF)torchrrr optimizerrrr r r r r __all__r__doc__listr/r-rr$r#rBs G %3D%3R$J         'K@ ^   $#3 L3 <36l3f 3 &\ 3 f 3D[333 $;3 3}33"#3$ 6>%3& 6>'3( )3*+3, -3./30 13r$