o <&iE-@sdddlZddlmZddlmZmZmZmZmZmZm Z ddl m Z m Z ddgZ GdddeZd d ed ed ed e_  dde ede ede ede ede edededededededefddZde ede ede ede edededededededefddZde ede ede ede edededededededefddZdS) N)Tensor) Optimizer_use_grad_for_differentiable_default_to_fused_or_foreach_differentiable_doc _foreach_doc _maximize_doc _view_as_real)ListOptionalAdadeltaadadeltacsdeZdZ     dddddeed ed effd d Zfd dZddZedddZ Z S)r ??ư>rNF)maximizedifferentiableforeachrrc sd|ks td|d|krdksntd|d|ks(td|d|ks3td|t|||||||d} t|| dS)NgzInvalid learning rate: rzInvalid rho value: zInvalid epsilon value: zInvalid weight_decay value: )lrrhoeps weight_decayrrr) ValueErrordictsuper__init__) selfparamsrrrrrrrdefaults __class__?C:\wamp64\www\opt\env\Lib\site-packages\torch/optim/adadelta.pyr s$  zAdadelta.__init__cs@t||jD]}|dd|dd|ddq dS)NrrFr)r __setstate__ param_groups setdefault)rstategroupr r"r#r$,s    zAdadelta.__setstate__c Csd}|dD][}|jdurq|t|O}|||jjr"td||j|j|}t|dkrKd|d<tj|tj d|d<tj|tj d|d<||d||d|dd 7<q|S) NFrz*Adadelta does not support sparse gradientsrstep)Z memory_format square_avg acc_deltar) gradtorch is_complexappendZ is_sparse RuntimeErrorr'lenZ zeros_likeZpreserve_format) rr(params_with_gradgrads square_avgs acc_deltas has_complexpr'r"r"r# _init_group3s,        zAdadelta._init_groupcCsd}|durt |}Wdn1swY|jD]A}g}g}g}g}|d|d|d|d|d|d|df\}} } } } } }||||||}t|||||| | | | | ||d q |S) zPerform a single optimization step. Args: closure (Callable, optional): A closure that reevaluates the model and returns the loss. Nrrrrrrr)rrrrrrrr6)r-Z enable_gradr%r8r)rclosureZlossr(r2r3r4r5rrrrrrrr6r"r"r#r)PsD   z Adadelta.step)rrrrN)N) __name__ __module__ __qualname__r boolrr$r8rr) __classcell__r"r"r r#r s(    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`_. a Args: params (iterable): iterable of parameters to optimize or dicts defining parameter groups 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). lr (float, optional): coefficient that scale delta before it is applied to the parameters (default: 1.0) weight_decay (float, optional): weight decay (L2 penalty) (default: 0) z zd .. _ADADELTA\: An Adaptive Learning Rate Method: https://arxiv.org/abs/1212.5701 Frr3r4r5rrr6rrrrrc Csj|dur t||dd\} }|rtjrtd|r"tjs"t} nt} | ||||||| | | ||d dS)zvFunctional API that performs Adadelta algorithm computation. See :class:`~torch.optim.Adadelta` for details. NF)Z use_fusedz6torch.jit.script not supported with foreach optimizers)rrrrrrr6)rr-ZjitZ is_scriptingr0_multi_tensor_adadelta_single_tensor_adadelta)rr3r4r5rrr6rrrrr_funcr"r"r#rs( cCst||||D]t\} } } }|s| n| } |dkr| j| |d} t| r3t| } t|}t| } | |j| | d|d| |}||}| rT|}| || ||j||d|dt| rst |}| j || dqdS)Nralpharvalue) zipaddr-r.Z view_as_realZmul_Zaddcmul_Zsqrt_cloneZdiv_Zview_as_complexZadd_)rr3r4r5rrrrrrr6paramr,r*r+stddeltar"r"r#r@s*      r@cCs4| rJdt|dkrdSt||||g} | D]|\\} } }}}|r*t| } | r3t| | |||dkrJ|rBtj| | |dntj| | |d} t ||tj || | d|dt||}t |t||}t |t ||t || tj| || dt ||tj |||d|dqdS)Nz#_foreach ops don't support autogradrrCrrE) r1rZ"_group_tensors_by_device_and_dtypevaluesr-Z _foreach_negr Z _foreach_add_Z _foreach_addZ _foreach_mul_Z_foreach_addcmul_Z_foreach_sqrt_Z _foreach_div_)rr3r4r5rrrrrrr6Zgrouped_tensorsZ device_paramsZ device_gradsZdevice_square_avgsZdevice_acc_deltasrArKZdeltasr"r"r#r? s2           r?)NFF)r-rZ optimizerrrrrrr r typingr r __all__r __doc__r=floatrr@r?r"r"r"r#s $u 9     0    )