U ʹj7@s4ddlZddlmZddlmZmZmZmZmZmZddl m Z m Z ddgZ GdddeZ d d ed ed ed e _de ee ee ee ee ee eeeeeeeeedddZe ee ee ee ee eeeeeeeeed ddZe ee ee ee ee eeeeeeeeed ddZdS)N)Tensor) Optimizer_default_to_fused_or_foreach_use_grad_for_differentiable_differentiable_doc _foreach_doc _maximize_doc)ListOptionalRMSproprmspropc sNeZdZdeeeedfdd Zfd d Zd d ZedddZ Z S)r {Gz?Gz?:0yE>rFN)foreachmaximizedifferentiablec sd|kstd|d|ks,td|d|ksBtd|d|ksXtd|d|ksntd|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__R/var/www/html/bwave/python/venv/lib/python3.8/site-packages/torch/optim/rmsprop.pyr s,  zRMSprop.__init__csXt||jD]@}|dd|dd|dd|dd|ddqdS)NrrrFrrr)r __setstate__ param_groups setdefault)rstategroupr!r#r$r%0s      zRMSprop.__setstate__c Cs$|dD]}|jdkrq|||jjr4td||j|j|}t|dkrd|d<tj|tjd|d<|ddkrtj|tjd|d<|d rtj|tjd|d <||d|ddkr||d|d r||d |d rt |dt rtd |dd 7<qdS)Nrz)RMSprop does not support sparse gradientsrstep)Z memory_format square_avgrZmomentum_bufferrgrad_avgrz`step` can't be a tensorr) gradappendZ is_sparse RuntimeErrorr(lentorchZ zeros_likeZpreserve_format isinstancer) rr)params_with_gradgrads square_avgsmomentum_buffer_list grad_avgspr(r#r#r$ _init_group9s@          zRMSprop._init_groupc Csd}|dk r&t |}W5QRX|jD]t}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) rrrrrrrrr)r1Z enable_gradr&r9r ) rclosureZlossr)r3r4r5r7r6r#r#r$r*_s8  z RMSprop.step) rrrrrFNFF)N) __name__ __module__ __qualname__r boolrr%r9rr* __classcell__r#r#r!r$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 F)rr4r5r7r6rrrrrrrrrcCsn|dkrt||dd\}}|r0tjr0td|rDtjsDt}nt}|||||||| | | | | ||d dS)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)rrrrrrrr)rr1ZjitZ is_scriptingr/_multi_tensor_rmsprop_single_tensor_rmsprop)rr4r5r7r6rrrrrrrrr_funcr#r#r$r s,) rr4r5r7r6rrrrrrrrcCsLt|D]<\} }|| }| s"|n| }|| }|dkrF|j||d}t|}|rrt|}t|}t|}||j||d|d| r|| }|rt|}||d||j||dd }n| }| r||}n | |}| dkr4|| }|rt|}||  |||j || dq|j ||| dqdS)Nrrrvalue) enumerateaddr1 is_complex view_as_realZmul_Zaddcmul_Zlerp_ZaddcmulZsqrt_sqrtZadd_Zaddcdiv_)rr4r5r7r6rrrrrrrriparamr-r+Zis_complex_paramr,avgbufr#r#r$rAs:         rAcCst|dkrdS| rtdt|||||g} | D]D\\}}}}}}| rZt|}|dkr| rxtj|||dntj|||d}dd}||}||}||}t ||tj |||d|d| r||}t ||d|tj |||dd}t |t||nt|}t||| dkrj||}t || t|||tj||| dq8tj|||| dq8dS) Nrz#_foreach ops don't support autogradrDcSsdd|DS)NcSs$g|]}t|rt|n|qSr#)r1rJrK).0tr#r#r$ VszH_multi_tensor_rmsprop.._view_complex_as_real..r#)Z tensor_listr#r#r$_view_complex_as_realUsz4_multi_tensor_rmsprop.._view_complex_as_realrrErG)r0AssertionErrorrZ"_group_tensors_by_device_and_dtypevaluesr1Z _foreach_negZ _foreach_add_Z _foreach_addZ _foreach_mul_Z_foreach_addcmul_Z_foreach_lerp_Z_foreach_addcmulZ_foreach_sqrt_Z _foreach_sqrtZ_foreach_addcdiv_)rr4r5r7r6rrrrrrrrZgrouped_tensorsZgrouped_paramsZ grouped_gradsZgrouped_square_avgsZgrouped_grad_avgsZgrouped_momentum_buffer_listrBrTrOr#r#r$r@2sD         r@)NFF)r1rZ optimizerrrrrrr typingr r __all__r __doc__r>floatr rAr@r#r#r#r$sz  *  E 4 :