o wZh0O$@sdZddlmZmZmZddlZddlmZddlmZm Z m Z m Z m Z m Z mZmZmZmZmZmZmZddgZGd ddeZd d ed ed e d ed e d e_deedeedeedeedeedeededededededededededef dd Zdeedeedeedeedeedeededededededededededef d!d"Ze ed#  $ $ $ $d'deedeedeedeedeedeed%eededededededededededef"d&dZdS)(z)Implementation for the RMSprop algorithm.)castOptionalUnionN)Tensor) _capturable_doc_default_to_fused_or_foreach_differentiable_doc_disable_dynamo_if_unsupported _foreach_doc!_get_capturable_supported_devices_get_scalar_dtype _maximize_doc _params_doc_use_grad_for_differentiable _view_as_real OptimizerParamsTRMSproprmspropcseZdZ          ddedeeefd ed ed ed ed ededeededeffdd Z fddZ ddZ e dddZ ZS)r{Gz?Gz?:0yE>rFNparamslralphaeps weight_decaymomentumcentered capturableforeachmaximizedifferentiablec st|tr|dkrtdd|kstd|d|ks%td|d|ks0td|d|ks;td|d|ksFtd|t|||||||| | | d } t|| dS) NrzTensor lr must be 1-elementgzInvalid learning rate: zInvalid epsilon value: zInvalid momentum value: zInvalid weight_decay value: zInvalid alpha value: ) rrrrrrr r!r"r#) isinstancerZnumel ValueErrordictsuper__init__) selfrrrrrrrr r!r"r#defaults __class__B/var/www/auris/lib/python3.10/site-packages/torch/optim/rmsprop.pyr(s2 zRMSprop.__init__cst||jD]_}|dd|dd|dd|dd|dd|dd|d D]4}|j|g}t|dkrgt|d sgt |d }|dr]tj |t |j d ntj |t d |d <q3q dS) NrrrFr!r"r#r rstepdtypedevicer1) r' __setstate__ param_groups setdefaultstategetlentorchZ is_tensorfloattensorr r2)r)r7grouppZp_stateZstep_valr+r-r.r4Fs*           zRMSprop.__setstate__c Cs8d}|dD]} | jdurq|t| O}|| | jjr"td|| j|j| } t| dkrs|drAtjdt | j dntjdt d| d <tj | tj d | d <|d dkretj | tj d | d <|drstj | tj d | d<|| d || d |d dkr|| d |dr|| dq|S)NFrz)RMSprop does not support sparse gradientsrr r-r0r3r/)Z memory_format square_avgrZmomentum_bufferrgrad_avg) gradr: is_complexappendZ is_sparse RuntimeErrorr7r9Zzerosr r2Z zeros_likeZpreserve_format) r)r=params_with_gradgrads square_avgsmomentum_buffer_list grad_avgs state_steps has_complexr>r7r-r-r. _init_group[sD          zRMSprop._init_groupc Cs|d}|dur!t |}Wdn1swY|jD]B}g}g}g}g}g}g} |||||||| } t|||||| |d|d|d|d|d|d|d|d |d |d | d q$|S) zPerform a single optimization step. Args: closure (Callable, optional): A closure that reevaluates the model and returns the loss. Nrrrrrrr!r"r#r ) rrrrrrr!r"r#r rK)Z _cuda_graph_capture_health_checkr:Z enable_gradr5rLr) r)closureZlossr=rErFrGrIrHrJrKr-r-r.r/sT   z RMSprop.step) rrrrrFFNFFN)__name__ __module__ __qualname__rrr;rboolrr(r4rLrr/ __classcell__r-r-r+r.rsL      )3aj Implements RMSprop algorithm. .. math:: \begin{aligned} &\rule{110mm}{0.4pt} \\ &\textbf{input} : \alpha \text{ (alpha)}, \: \gamma \text{ (lr)}, \: \theta_0 \text{ (params)}, \: f(\theta) \text{ (objective)} \\ &\hspace{13mm} \lambda \text{ (weight decay)},\: \mu \text{ (momentum)}, \: centered, \: \epsilon \text{ (epsilon)} \\ &\textbf{initialize} : v_0 \leftarrow 0 \text{ (square average)}, \: \textbf{b}_0 \leftarrow 0 \text{ (buffer)}, \: g^{ave}_0 \leftarrow 0 \\[-1.ex] &\rule{110mm}{0.4pt} \\ &\textbf{for} \: t=1 \: \textbf{to} \: \ldots \: \textbf{do} \\ &\hspace{5mm}g_t \leftarrow \nabla_{\theta} f_t (\theta_{t-1}) \\ &\hspace{5mm}if \: \lambda \neq 0 \\ &\hspace{10mm} g_t \leftarrow g_t + \lambda \theta_{t-1} \\ &\hspace{5mm}v_t \leftarrow \alpha v_{t-1} + (1 - \alpha) g^2_t \hspace{8mm} \\ &\hspace{5mm} \tilde{v_t} \leftarrow v_t \\ &\hspace{5mm}if \: centered \\ &\hspace{10mm} g^{ave}_t \leftarrow g^{ave}_{t-1} \alpha + (1-\alpha) g_t \\ &\hspace{10mm} \tilde{v_t} \leftarrow \tilde{v_t} - \big(g^{ave}_{t} \big)^2 \\ &\hspace{5mm}if \: \mu > 0 \\ &\hspace{10mm} \textbf{b}_t\leftarrow \mu \textbf{b}_{t-1} + g_t/ \big(\sqrt{\tilde{v_t}} + \epsilon \big) \\ &\hspace{10mm} \theta_t \leftarrow \theta_{t-1} - \gamma \textbf{b}_t \\ &\hspace{5mm} else \\ &\hspace{10mm}\theta_t \leftarrow \theta_{t-1} - \gamma g_t/ \big(\sqrt{\tilde{v_t}} + \epsilon \big) \hspace{3mm} \\ &\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 `lecture notes `_ by G. Hinton. and centered version `Generating Sequences With Recurrent Neural Networks `_. The implementation here takes the square root of the gradient average before adding epsilon (note that TensorFlow interchanges these two operations). The effective learning rate is thus :math:`\gamma/(\sqrt{v} + \epsilon)` where :math:`\gamma` is the scheduled learning rate and :math:`v` is the weighted moving average of the squared gradient. z Args: a0 lr (float, Tensor, optional): learning rate (default: 1e-2) alpha (float, optional): smoothing constant (default: 0.99) 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 (float, optional): momentum factor (default: 0) centered (bool, optional) : if ``True``, compute the centered RMSProp, the gradient is normalized by an estimation of its variance z z rrFrGrIrHrJrrrrrrr"r#r rKc Cst|D]\}}||}tjs,|r,t}|jj|jjkr$|jj|vs,Jd|d||}| s4|n| }||}|d7}| dkrJ|j|| d}t|}|r`t |}t |}t |}| |j ||d|d| r||}|rzt |}| |d||j ||dd}n|}| r||}n||}| dkr||}|rt |}| | |||j|| dq|j||| dqdS)NIIf capturable=True, params and state_steps must be on supported devices: .rrrvalue) enumerater:compiler is_compilingr r2typeaddrBZ view_as_realZmul_Zaddcmul_Zlerp_ZaddcmulZsqrt_sqrtZadd_Zaddcdiv_)rrFrGrIrHrJrrrrrrr"r#r rKiparamr/capturable_supported_devicesrAr?Zis_complex_paramr@avgbufr-r-r._single_tensor_rmspropsL         rec !st|dkrdS| rJdtjs.|r.ttfddt||Ds.Jddt||||||g}| D]0\\}}}}}}}t t t |}t t t |}t t t |}t t t |}|r||g}| dkrzt t t |}| || rt t t |}| |t|g|R| rt|}tjs|djrtj|tjddd dd nt|d | dkr| rtj||| d ntj||| d }t||tj|||d |d | rt t t |}t||d |tj|||d d }t|t||n t|}t||| dkrKt t t |}t|| t||||rAt|tj rAt|| } t|| q=tj||| d q=|rdt|tj rdt|| t|||q=tj|||| d q=dS)Nrz#_foreach ops don't support autogradc3s0|]\}}|jj|jjko|jjvVqdSrN)r2r]).0r>r/rbr-r. is   z(_multi_tensor_rmsprop..rTrUg?cpu)r2rVrrWrY)r9r:r[r\r allziprZ"_group_tensors_by_device_and_dtypevaluesrlistrrCrZ _foreach_negZis_cpuZ _foreach_add_r<Z _foreach_addZ _foreach_mul_Z_foreach_addcmul_Z_foreach_lerp_Z_foreach_addcmulZ_foreach_sqrt_Z _foreach_sqrtZ_foreach_addcdiv_r$Z _foreach_mulZ _foreach_div_)!rrFrGrIrHrJrrrrrrr"r#r rKZgrouped_tensorsZgrouped_params_Zgrouped_grads_Zgrouped_square_avgs_Zgrouped_grad_avgs_Zgrouped_momentum_buffer_list_Zgrouped_state_steps__Zgrouped_paramsZ grouped_gradsZgrouped_square_avgsZgrouped_state_stepsZstate_and_gradsZgrouped_momentum_buffer_listZgrouped_grad_avgsrcZ momentum_lrr-rgr._multi_tensor_rmspropNs                ro)Zsingle_tensor_fnFr!c Cstjstdd|Dstd|durt||dd\}}|r*tjr*td|r4tjs4t}nt }|||||||| | | ||||| || ddS) ztFunctional API that performs rmsprop algorithm computation. See :class:`~torch.optim.RMSProp` for details. css|] }t|tjVqdSrN)r$r:r)rftr-r-r.rhs  zrmsprop..zPAPI has changed, `state_steps` argument must contain a list of singleton tensorsNF)Z use_fusedz6torch.jit.script not supported with foreach optimizers) rrrrrrr"r r#rK) r:r[r\rjrDrZjitZ is_scriptingrore)rrFrGrIrHrJr!r"r#r rKrrrrrrrnfuncr-r-r.rsB  )NFFFF)__doc__typingrrrr:rZ optimizerrrr r r r r rrrrrr__all__rrmr;rRrerorr-r-r-r.s <,,   A      G