o wZh@@sddlmZmZmZmZddlZddlmZddlmZm Z m Z m Z m Z m Z mZmZmZmZmZmZmZddgZGdddeZd d ed e d ed ed e d e_deedeedeedeedeededededededededefddZdeedeedeedeedeededededededededefddZe ed  d#deedeedeedeedeeded!eedededededededefd"dZdS)$)AnycastOptionalUnionN)Tensor) _capturable_doc_default_to_fused_or_foreach_differentiable_doc_disable_dynamo_if_unsupported _foreach_doc!_get_capturable_supported_devices_get_scalar_dtype _maximize_doc _params_doc_use_grad_for_differentiable _view_as_real OptimizerParamsTAdadeltaadadeltacseZdZ     dddddded eeefd ed ed ed eedededeffddZ fddZ de e e fdeedeedeedeedeef ddZed ddZZS)!r??ư>rNF) capturablemaximizedifferentiableparamslrrhoeps weight_decayforeachrrrc st|tr|dkrtdd|kstd|d|kr$dks,ntd|d|ks7td|d|ksBtd|t|||||||| d } t|| dS) NrzTensor lr must be 1-elementgzInvalid learning rate: rzInvalid rho value: zInvalid epsilon value: zInvalid weight_decay value: )rrr r!rrr"r) isinstancerZnumel ValueErrordictsuper__init__) selfrrrr r!r"rrrdefaults __class__C/var/www/auris/lib/python3.10/site-packages/torch/optim/adadelta.pyr's*  zAdadelta.__init__cst||jD]S}|dd|dd|dd|dd|dD]4}|j|g}t|dkr[t|ds[t |d}|drQtj |t |j d ntj |t d |d<q'q dS) Nr"rFrrrrstepdtypedevicer0) r& __setstate__ param_groups setdefaultstategetlentorchZ is_tensorfloattensorrr1)r(r6grouppZp_stateZstep_valr*r,r-r3@s&         zAdadelta.__setstate__r<params_with_gradgrads square_avgs acc_deltas state_stepsc Csd}|dD]n}|jdurq|t|O}|||jjr"td||j|j|} t| dkr_|drAtjdt |j dntjdt d| d <tj |tj d | d <tj |tj d | d <|| d || d || d q|S) NFrz*Adadelta does not support sparse gradientsrrr,r/r2r.)Z memory_format square_avg acc_delta) gradr9 is_complexappendZ is_sparse RuntimeErrorr6r8Zzerosrr1Z zeros_likeZpreserve_format) r(r<r>r?r@rArB has_complexr=r6r,r,r- _init_groupSs2        zAdadelta._init_groupcCs|d}|dur!t |}Wdn1swY|jD]J}g}g}g}g}g}|d|d|d|d|d|d|d|d f\} } } } } }}}|||||||}t|||||| | | | | ||||d q$|S) zPerform a single optimization step. Args: closure (Callable, optional): A closure that reevaluates the model and returns the loss. Nrrr r!r"rrr) rrr r!r"rrrrI)Z _cuda_graph_capture_health_checkr9Z enable_gradr4rJr)r(closureZlossr<r>r?r@rArBrrr r!r"rrrrIr,r,r-r.~sd    z Adadelta.step)rrrrNN)__name__ __module__ __qualname__rrr:rrboolr'r3r%strrlistrJrr. __classcell__r,r,r*r-rsZ      $  +aImplements Adadelta algorithm. .. math:: \begin{aligned} &\rule{110mm}{0.4pt} \\ &\textbf{input} : \gamma \text{ (lr)}, \: \theta_0 \text{ (params)}, \: f(\theta) \text{ (objective)}, \: \rho \text{ (decay)}, \: \lambda \text{ (weight decay)} \\ &\textbf{initialize} : v_0 \leftarrow 0 \: \text{ (square avg)}, \: u_0 \leftarrow 0 \: \text{ (accumulate variables)} \\[-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 v_{t-1} \rho + g^2_t (1 - \rho) \\ &\hspace{5mm}\Delta x_t \leftarrow \frac{\sqrt{u_{t-1} + \epsilon }}{ \sqrt{v_t + \epsilon} }g_t \hspace{21mm} \\ &\hspace{5mm} u_t \leftarrow u_{t-1} \rho + \Delta x^2_t (1 - \rho) \\ &\hspace{5mm}\theta_t \leftarrow \theta_{t-1} - \gamma \Delta x_t \\ &\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 `ADADELTA: An Adaptive Learning Rate Method`_. z Args: ar lr (float, Tensor, optional): coefficient that scale delta before it is applied to the parameters (default: 1.0) rho (float, optional): coefficient used for computing a running average of squared gradients (default: 0.9). A higher value of `rho` will result in a slower average, which can be helpful for preventing oscillations in the learning process. eps (float, optional): term added to the denominator to improve numerical stability (default: 1e-6). weight_decay (float, optional): weight decay (L2 penalty) (default: 0) z zd .. _ADADELTA\: An Adaptive Learning Rate Method: https://arxiv.org/abs/1212.5701 rr?r@rArBrrr r!rrrrIcsLtjs"| r"tddtfddt||Ds"Jddt|||||D]y\} }}}}|d7}| s9|n| }|dkrG|j| |d }t| r[t|}t|}t|}| |j ||d|d || }|| }| r|| }| | || |j ||d|d t| rt|}| j|| d q*dS) NFZ supports_xlac30|]\}}|jj|jjko|jjvVqdSrLr1type.0r=r.Zcapturable_supported_devicesr,r-    z*_single_tensor_adadelta..IIf capturable=True, params and state_steps must be on supported devices: .rralphavalue)r9compiler is_compilingr allzipaddrFZ view_as_realZmul_Zaddcmul_Zsqrt_cloneZdiv_Zview_as_complexZadd_)rr?r@rArBrrr r!rrrrIparamrErCrDr.stddeltar,rZr-_single_tensor_adadeltas>         rlcs.| rJdtjs(| r(tddtfddt||Ds(Jddt|dkr0dSt|||||g} | D]\\}}}}}}t t t |}t t t |}t t t |}t t t |}t t t |}| rst ||||tjs|djrtj|tjd d d d d nt|d | rt|}|dkr| rtj|||d ntj|||d }t||tj|||d |dt||}t|t||}t|t||t||t||tj|||d |d| r t|tj r t|| t||q>tj||| d q>dS)Nz#_foreach ops don't support autogradFrTc3rUrLrVrXrZr,r-r[Ar\z)_multi_tensor_adadelta..r]r^rrcpu)r1r_rra)r9rcrdr rerfr8rZ"_group_tensors_by_device_and_dtypevaluesrrRrrZis_cpuZ _foreach_add_r;Z _foreach_negZ _foreach_addZ _foreach_mul_Z_foreach_addcmul_Z_foreach_sqrt_Z _foreach_div_r#)rr?r@rArBrrr r!rrrrIZgrouped_tensorsZdevice_params_Z device_grads_Zdevice_square_avgs_Zdevice_acc_deltas_Zdevice_state_steps__Z device_paramsZ device_gradsZdevice_square_avgsZdevice_acc_deltasZdevice_state_stepsrjZdeltasr,rZr-_multi_tensor_adadelta*s|                 rp)Zsingle_tensor_fnFr"c Cstjstdd|Dstd|durt||dd\}}|r*tjr*td|r4tjs4t}nt }||||||| | | | | |||d dS) zvFunctional API that performs Adadelta algorithm computation. See :class:`~torch.optim.Adadelta` for details. css|] }t|tjVqdSrL)r#r9r)rYtr,r,r-r[s  zadadelta..zPAPI has changed, `state_steps` argument must contain a list of singleton tensorsNF)Z use_fusedz6torch.jit.script not supported with foreach optimizers)rrr r!rrrrI) r9rcrdrerHr ZjitZ is_scriptingrprl)rr?r@rArBrr"rrIrrr r!rrofuncr,r,r-rs<  )FNFF)typingrrrrr9rZ optimizerrr r r r r rrrrrrr__all__r__doc__rRr:rPrlrprr,r,r,r-s <&  5      6      d