a hWg@sdZddlmZmZmZddlZddlmZddlmZm Z m Z m Z m Z m Z mZmZmZmZmZmZmZmZmZmZddgZGd ddeZd d ed e d ed ed e d e_eeeeeeeeeeeeeeeeeeeeeeedddZeeeeeeeeeeeeeeeeeeeeeeedddZe eddeeeeeeeeeeeeeeeeeeeeeeeeedddZdS)z'Implementation for the NAdam algorithm.)castOptionalUnionN)Tensor)_capturable_doc_default_to_fused_or_foreach_differentiable_doc_disable_dynamo_if_unsupported _foreach_doc!_get_capturable_supported_devices_get_scalar_dtype _get_value _maximize_doc _params_doc_stack_if_compiling _to_scalar_use_grad_for_differentiable _view_as_real OptimizerParamsTNAdamnadamcszeZdZddddddeeeefeeefeeeee eeeed fd d Z fd d Z ddZ e dddZZS)rMb`?g?g+?:0yE>rMbp?FN)foreachmaximize capturabledifferentiable) paramslrbetaseps weight_decaymomentum_decaydecoupled_weight_decayrrrr c st|tr|dkrtdd|ks4td|d|ksJtd|d|dkrbdksvntd|dd|dkrdksntd |dd|kstd |d|kstd |t||||||| || | d } t|| dS) NrzTensor lr must be 1-elementzInvalid learning rate: zInvalid epsilon value: r?z#Invalid beta parameter at index 0: z#Invalid beta parameter at index 1: zInvalid weight_decay value: zInvalid momentum_decay value: ) r"r#r$r%r&r'rrrr ) isinstancerZnumel ValueErrordictsuper__init__) selfr!r"r#r$r%r&r'rrrr defaults __class__?/var/www/auris/lib/python3.9/site-packages/torch/optim/nadam.pyr.!s6 zNAdam.__init__cst||jD]}|dd|dd|dd|dd|dd|dD]}|j|g}t|dkrZt|d st |d }|drtj |t |j d ntj |t d |d <t|d sZ|d }|drtj |t |j d ntj |t d |d <qZqdS) NrFrrr r'r!rstepdtypedevicer7 mu_product) r- __setstate__ param_groups setdefaultstategetlentorchZ is_tensorfloattensorr r8)r/r>grouppZp_stateZstep_valZ mu_prod_valr1r3r4r;Ls4            zNAdam.__setstate__c Cs,d}|dD]} | jdur |t| O}|| | jjrDtd|| j|j| } t| dkr|drtjdt | j dntj dt d | d <|drtj dt | j dntj d t d | d <tj | tjd | d<tj | tjd | d<|| d|| d|| d || d q |S)NFr!z'NAdam does not support sparse gradientsrrr3r6r(r9r5r)r:)Z memory_formatexp_avg exp_avg_sq)gradrA is_complexappendZ is_sparse RuntimeErrorr>r@zerosr r8rCZonesZ zeros_likeZpreserve_format) r/rDparams_with_gradgradsexp_avgs exp_avg_sqs mu_products state_steps has_complexrEr>r3r3r4 _init_groupjs:        zNAdam._init_groupc Cs|d}|durBt|}Wdn1s80Y|jD]}g}g}g}g}g}g} ttttf|d\} } |||||||| } t|||||| | | |d|d|d|d|d|d|d |d |d | d qH|S) zPerform a single optimization step. Args: closure (Callable, optional): A closure that reevaluates the model and returns the loss. Nr#r"r%r&r$rr'rrr ) beta1beta2r"r%r&r$rr'rrr rS) Z _cuda_graph_capture_health_checkrAZ enable_gradr<rtuplerBrTr) r/closureZlossrDrMrNrOrPrQrRrUrVrSr3r3r4r5sV $  z NAdam.step)rrrrrF)N)__name__ __module__ __qualname__rrrBrrWboolrr.r;rTrr5 __classcell__r3r3r1r4r s8   + 2a Implements NAdam algorithm. .. math:: \begin{aligned} &\rule{110mm}{0.4pt} \\ &\textbf{input} : \gamma_t \text{ (lr)}, \: \beta_1,\beta_2 \text{ (betas)}, \: \theta_0 \text{ (params)}, \: f(\theta) \text{ (objective)} \\ &\hspace{13mm} \: \lambda \text{ (weight decay)}, \:\psi \text{ (momentum decay)} \\ &\hspace{13mm} \: \textit{decoupled\_weight\_decay}, \:\textit{maximize} \\ &\textbf{initialize} : m_0 \leftarrow 0 \text{ ( first moment)}, v_0 \leftarrow 0 \text{ ( second moment)} \\[-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} \\ &\hspace{5mm} \textbf{if} \: \lambda \neq 0 \\ &\hspace{10mm}\textbf{if} \: \textit{decoupled\_weight\_decay} \\ &\hspace{15mm} \theta_t \leftarrow \theta_{t-1} - \gamma \lambda \theta_{t-1} \\ &\hspace{10mm}\textbf{else} \\ &\hspace{15mm} g_t \leftarrow g_t + \lambda \theta_{t-1} \\ &\hspace{5mm} \mu_t \leftarrow \beta_1 \big(1 - \frac{1}{2} 0.96^{t \psi} \big) \\ &\hspace{5mm} \mu_{t+1} \leftarrow \beta_1 \big(1 - \frac{1}{2} 0.96^{(t+1)\psi}\big)\\ &\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 \mu_{t+1} m_t/(1-\prod_{i=1}^{t+1}\mu_i)\\[-1.ex] & \hspace{11mm} + (1-\mu_t) g_t /(1-\prod_{i=1}^{t} \mu_{i}) \\ &\hspace{5mm}\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 `Incorporating Nesterov Momentum into Adam`_. z Args: a lr (float, Tensor, optional): learning rate (default: 2e-3) betas (Tuple[float, float], optional): coefficients used for computing running averages of gradient and its square (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 (L2 penalty) (default: 0) momentum_decay (float, optional): momentum momentum_decay (default: 4e-3) decoupled_weight_decay (bool, optional): whether to decouple the weight decay as in AdamW to obtain NAdamW. If True, the algorithm does not accumulate weight decay in the momentum nor variance. (default: False) z z .. _Incorporating Nesterov Momentum into Adam: https://openreview.net/forum?id=OM0jvwB8jIp57ZJjtNEZ .. _Decoupled Weight Decay Regularization: https://arxiv.org/abs/1711.05101 )r!rNrOrPrQrRrUrVr"r%r&r$r'rrr rSc Csttjst|}t|D]R\}}| s0||n|| }||}||}||}||}t|rt|}t|}t|}t|}tjs|rt }|j j |j j kr|j j krnn |j j |vsJd|d|d7}|r|}nt |}d||}| dkr8| r*| d|| n|j|| d}|ddd|| }|ddd|d| }||9}||d|| |j||d|d ||}|s|r|| }||}|| d|d|}|| |d|}||||||qt ||}|| |j||| d|dt |d |j||| |d|d qdS) NzVIf capturable=True, params, mu_products and state_steps must be on supported devices: .rralphar)?Q?)value)rAjit is_scriptingr enumeraterIZ view_as_realcompiler is_compilingr r8typerZmul_addZlerp_Zaddcmul_divsqrtZaddcdiv_Zadd_)r!rNrOrPrQrRrUrVr"r%r&r$r'rrr rSiparamrHrFrGr:Zstep_tcapturable_supported_devicesr5Zbias_correction2mumu_nextdenomZmu_product_nextr3r3r4_single_tensor_nadamsn                 rsc (st|dkrdS|rJdtjsb|rbtddtfddt|||DsbJddtt ||||||g}| D]b\\}}}}}}}t t t |}t t t |}t t t |}t t t |}t t t |}t t t |}|rt||||| rt|}tjsJ|djrJtj|tjd d d d d n t|d | dkr| r|t|d | n(| rtj||| d ntj||| d }t||d t|t|||d t|}|rt|} td| }!t|!dt|!d t|!t| td| }"t|"dt|"d t|"~ t|}#t|#d t|#t|#n:fdd|D}#fdd|D}!fdd|D}"t||!t||#t|| ~#|rt|!d t|!t|d }$t|$t|!|$|!}%~$t||"}$t|"t|$d t|"|$|"}&~$t|%|}'t|'|&|t||'|qt fddt||!D}%t fddt||"D}&t||||%t||||&qdS)Nrz#_foreach ops don't support autogradF)Z supports_xlac3sD|]<\}}}|jj|jjko*|jjkno:|jjvVqdSN)r8ri).0rEmpr5)ror3r4 s$z&_multi_tensor_nadam..zWIf capturable=True, params, mu_products, and state_steps must be on supported devices: r^r)cpu)r8r_rrbgcs g|]}dt|dqS)rrarrur5)rVr3r4 sz'_multi_tensor_nadam..cs(g|] }dddt|qS)r)rarbryrzrUr&r3r4r{scs,g|]$}dddt|dqS)r)rarbrryrzr|r3r4r{scs0g|](\}}td|dt|dqSr)ry)rur:rpr"r3r4r{6scs0g|](\}}t|dt||dqSr}ry)rur:rqrr3r4r{<s)!r@rArgrhr allziprrZ"_group_tensors_by_device_and_dtypevaluesrlistrrZ _foreach_negZis_cpuZ _foreach_add_rCZ _foreach_mul_Z _foreach_addZ_foreach_lerp_Z_foreach_addcmul_Z _foreach_sqrtZ _foreach_mulZ _foreach_powZ _foreach_sub_Z _foreach_neg_Z_foreach_sqrt_Z _foreach_div_Z _foreach_subZ_foreach_addcdiv_r)(r!rNrOrPrQrRrUrVr"r%r&r$r'rrr rSZgrouped_tensorsZgrouped_params_Zgrouped_grads_Zgrouped_exp_avgs_Zgrouped_exp_avg_sqs_Zgrouped_mu_products_Zgrouped_state_steps__Zgrouped_paramsZ grouped_gradsZgrouped_exp_avgsZgrouped_exp_avg_sqsZgrouped_mu_productsZgrouped_state_stepsZexp_avg_sq_sqrtexponentZmusZmu_nextsZbias_correction_sqrtrrZstep_size_gradsZstep_size_expavg numeratorr3)rUrVror"r&r4_multi_tensor_nadamzs                                              r)Zsingle_tensor_fnF)r!rNrOrPrQrRr'rrr rSrrUrVr"r%r&r$c Cstdd|Dstdtdd|Ds4td|durNt|| dd\}}|rdtjrdtd |rxtjsxt}nt}|||||||| | |||| |||| | d dS) zpFunctional API that performs NAdam algorithm computation. See :class:`~torch.optim.NAdam` for details. css|]}t|tjVqdSrtr*rArrutr3r3r4rwqznadam..zPAPI has changed, `state_steps` argument must contain a list of singleton tensorscss|]}t|tjVqdSrtrrr3r3r4rwvrzPAPI has changed, `mu_products` argument must contain a list of singleton tensorsNF)Z use_fusedz6torch.jit.script not supported with foreach optimizers) rUrVr"r%r&rr'r$rr rS)rrKrrArdrerrs)r!rNrOrPrQrRr'rrr rSrrUrVr"r%r&r$rfuncr3r3r4rUsH )FNFFFF) __doc__typingrrrrArZ optimizerrrr r r r r rrrrrrrrr__all__rrrBr\rsrrr3r3r3r4s H8' D b \