o *ѹgX8@sddlZddlmZddlmZmZmZmZmZmZm Z ddl m Z m Z ddgZ GdddeZd d ed ed ed e_  d"de ede ede ede ede ede edededededededededefddZde ede ede ede ede edededededededededefddZde ede ede ede ede edededededededededefd d!ZdS)#N)Tensor) Optimizer_default_to_fused_or_foreach_use_grad_for_differentiable_differentiable_doc _foreach_doc _maximize_doc _view_as_real)ListOptionalRMSproprmspropcsdeZdZ         ddeeded effd d Zfd d ZddZedddZ Z S)r {Gz?Gz?:0yE>rFNforeachmaximizedifferentiablec sd|ks td|d|kstd|d|ks!td|d|ks,td|d|ks7td|t|||||||| | d } t|| dS)NgzInvalid learning rate: zInvalid epsilon value: zInvalid momentum value: zInvalid weight_decay value: zInvalid alpha value: ) lrmomentumalphaepscentered weight_decayrrr) ValueErrordictsuper__init__) selfparamsrrrrrrrrrdefaults __class__W/Users/admin/.pyenv/versions/3.10.0/lib/python3.10/site-packages/torch/optim/rmsprop.pyr s,  zRMSprop.__init__csXt||jD] }|dd|dd|dd|dd|ddq dS)NrrrFrrr)r __setstate__ param_groups setdefault)rstategroupr"r$r%r&0s      zRMSprop.__setstate__c Cs0d}|dD]}|jdurq|t|O}|||jjr"td||j|j|} t| dkr_d| d<tj|tj d| d<|ddkrQtj|tj d| d <|d r_tj|tj d| d <|| d|ddkrs|| d |d r~|| d |d rt | dt rtd | dd7<q|S)NFr z)RMSprop does not support sparse gradientsrstep)Z memory_format square_avgrZmomentum_bufferrgrad_avgrz`step` can't be a tensorr) gradtorch is_complexappendZ is_sparse RuntimeErrorr)lenZ zeros_likeZpreserve_format isinstancer) rr*params_with_gradgrads square_avgsmomentum_buffer_list grad_avgs has_complexpr)r$r$r% _init_group9s@          zRMSprop._init_groupc Csd}|durt |}Wdn1swY|jD];}g}g}g}g}g}|||||||} t||||||d|d|d|d|d|d|d|d |d | d q |S) zPerforms a single optimization step. Args: closure (Callable, optional): A closure that reevaluates the model and returns the loss. Nrrrrrrrrr) rrrrrrrrrr:)r/Z enable_gradr'r<r) rclosureZlossr*r5r6r7r9r8r:r$r$r%r+bs<  z RMSprop.step) rrrrrFNFF)N) __name__ __module__ __qualname__r boolrr&r<rr+ __classcell__r$r$r"r%r s*    % )a 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\\ &\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. a Args: params (iterable): iterable of parameters to optimize or dicts defining parameter groups lr (float, optional): learning rate (default: 1e-2) momentum (float, optional): momentum factor (default: 0) alpha (float, optional): smoothing constant (default: 0.99) eps (float, optional): term added to the denominator to improve numerical stability (default: 1e-8) centered (bool, optional) : if ``True``, compute the centered RMSProp, the gradient is normalized by an estimation of its variance weight_decay (float, optional): weight decay (L2 penalty) (default: 0) z z Fr r6r7r9r8rrrr:rrrrrrc Csp|dur t||dd\}}|rtjrtd|r"tjs"t}nt}||||||| | | | | ||||ddS)zsFunctional API that performs rmsprop algorithm computation. See :class:`~torch.optim.RMSProp` for details. NF)Z use_fusedz6torch.jit.script not supported with foreach optimizers) rrrrrrrrr:)rr/ZjitZ is_scriptingr2_multi_tensor_rmsprop_single_tensor_rmsprop)r r6r7r9r8rrrr:rrrrrr_funcr$r$r%rs. c CsFt|D]\}}||}| s|n| }||}|dkr"|j||d}t|}|r8t|}t|}t|}||j||d|d| re||}|rRt|}||d||j||dd }n| }| rq||}n| |}| dkr||}|rt|}||  |||j || dq|j ||| dqdS)Nrrrvalue) enumerateaddr/r0Z view_as_realZmul_Zaddcmul_Zlerp_ZaddcmulZsqrt_sqrtZadd_Zaddcdiv_)r r6r7r9r8rrrrrrrrr:iparamr.r,Zis_complex_paramr-avgbufr$r$r%rDs<        rDc Cst|dkrdS| rJdt|||||g}|D]\\}}}}}}| r,t|}|dkrC| r;tj|||dntj|||d}t|}| re||g}| dkrV| || r]| |t |g|Rt ||tj |||d|d| rt ||d|tj|||dd}t|t||n t|}t||| dkrt || t|||tj||| dqtj|||| dqdS)Nrz#_foreach ops don't support autogradrGrrHrJ)r3rZ"_group_tensors_by_device_and_dtypevaluesr/Z _foreach_negZ _foreach_add_Z _foreach_addlistr1r Z _foreach_mul_Z_foreach_addcmul_Z_foreach_lerp_Z_foreach_addcmulZ_foreach_sqrt_Z _foreach_sqrtZ_foreach_addcdiv_)r r6r7r9r8rrrrrrrrr:Zgrouped_tensorsZgrouped_paramsZ grouped_gradsZgrouped_square_avgsZgrouped_grad_avgsZgrouped_momentum_buffer_listrEZstate_and_gradsrPr$r$r%rC9sH          rC)NFFF)r/rZ optimizerrrrrrr r typingr r __all__r __doc__rAfloatrrDrCr$r$r$r%s $*  E     5      :