6bi "SrSSKJs Jr SSKJr SSKJr SSK J r \"5\ "SSS/S 9"S S \R555r \ RRS \R5\ lg) zAdam optimizer implementation.N) optimizer)register_keras_serializable) keras_exportzkeras.optimizers.Adamz"keras.optimizers.experimental.Adamz*keras.dtensor.experimental.optimizers.Adam)v1cj^\rSrSrSrSU4SjjrU4SjrSrU4SjrSr U=r $) Adama Optimizer that implements the Adam algorithm. Adam optimization is a stochastic gradient descent method that is based on adaptive estimation of first-order and second-order moments. According to [Kingma et al., 2014](http://arxiv.org/abs/1412.6980), the method is "*computationally efficient, has little memory requirement, invariant to diagonal rescaling of gradients, and is well suited for problems that are large in terms of data/parameters*". Args: learning_rate: A `tf.Tensor`, floating point value, a schedule that is a `tf.keras.optimizers.schedules.LearningRateSchedule`, or a callable that takes no arguments and returns the actual value to use. The learning rate. Defaults to `0.001`. beta_1: A float value or a constant float tensor, or a callable that takes no arguments and returns the actual value to use. The exponential decay rate for the 1st moment estimates. Defaults to `0.9`. beta_2: A float value or a constant float tensor, or a callable that takes no arguments and returns the actual value to use. The exponential decay rate for the 2nd moment estimates. Defaults to `0.999`. epsilon: A small constant for numerical stability. This epsilon is "epsilon hat" in the Kingma and Ba paper (in the formula just before Section 2.1), not the epsilon in Algorithm 1 of the paper. Defaults to `1e-7`. amsgrad: Boolean. Whether to apply AMSGrad variant of this algorithm from the paper "On the Convergence of Adam and beyond". Defaults to `False`. {{base_optimizer_keyword_args}} Reference: - [Kingma et al., 2014](http://arxiv.org/abs/1412.6980) - [Reddi et al., 2018]( https://openreview.net/pdf?id=ryQu7f-RZ) for `amsgrad`. Notes: The default value of 1e-7 for epsilon might not be a good default in general. For example, when training an Inception network on ImageNet a current good choice is 1.0 or 0.1. Note that since Adam uses the formulation just before Section 2.1 of the Kingma and Ba paper rather than the formulation in Algorithm 1, the "epsilon" referred to here is "epsilon hat" in the paper. The sparse implementation of this algorithm (used when the gradient is an IndexedSlices object, typically because of `tf.gather` or an embedding lookup in the forward pass) does apply momentum to variable slices even if they were not used in the forward pass (meaning they have a gradient equal to zero). Momentum decay (beta1) is also applied to the entire momentum accumulator. This means that the sparse behavior is equivalent to the dense behavior (in contrast to some momentum implementations which ignore momentum unless a variable slice was actually used). c >[TU]"SUUUUU U U U U S. UD6 URU5UlX lX0lX@lXPlg)N) name weight_decayclipnorm clipvalueglobal_clipnormuse_ema ema_momentumema_overwrite_frequency jit_compile)super__init___build_learning_rate_learning_ratebeta_1beta_2epsilonamsgrad)self learning_raterrrrr r rrrrrrr kwargs __class__s V/home/james-whalen/.local/lib/python3.13/site-packages/tf_keras/src/optimizers/adam.pyr Adam.__init__\se$  %+%$;#  #77 F    c>[TU]U5 [US5(aUR(agSUl/Ul/UlUHUnURR URUSS95 UR R URUSS95 MW UR(a:/Ul UH,nURR URUSS95 M. gg)zInitialize optimizer variables. Adam optimizer has 3 types of variables: momentums, velocities and velocity_hat (only set when amsgrad is applied), Args: var_list: list of model variables to build Adam variables on. _builtNTm)model_variable variable_namevvhat) rbuildhasattrr% _momentums _velocitiesappendadd_variable_from_referencer_velocity_hats)rvar_listvarr s r!r+ Adam.builds  h 4 " "t{{  C OO " "00#&c1     # #00#&c1   <<"$D ##**44'*&5  r#c[R"URUR5n[R"URS-UR5n[R "[R"UR UR5U5n[R "[R"URUR5U5nURU5nURURUnURURUn U[R"SU- 5-SU- - n [U[R5(GaxURU*SUR - -5 UR![R"UR"SUR - -UR$55 U RU *SUR- -5 U R![R"[R&"UR"5SUR- -UR$55 UR((aCUR*URUn U R-[R."X55 U n UR1X-[R"U 5UR2-- 5 gURX- SUR - -5 U R[R&"U5U - SUR- -5 UR((aCUR*URUn U R-[R."X55 U n UR1X-[R"U 5UR2-- 5 g)z=Update step given gradient and the associated model variable.N)tfcastrdtype iterationspowrr_var_keyr- _index_dictr.sqrt isinstance IndexedSlices assign_add scatter_addvaluesindicessquarerr1assignmaximum assign_subr) rgradientvariablelr local_step beta_1_power beta_2_powervar_keyr&r)alphav_hats r! update_stepAdam.update_steps WWT'' 8WWT__q0(..A vvbggdkk8>>BJO vvbggdkk8>>BJO --) OOD,,W5 6   T--g6 7RWWQ-..!l2BC h 0 0 1 1 LL!q4;;/ 0 MM  OOq4;;79I9I  LL!q4;;/ 0 MM  IIhoo.!dkk/B$$  ||++D,<,[TU]5nURURUR5UR UR URURS.5 U$)N)rrrrr) r get_configupdate_serialize_hyperparameterrrrrr)rconfigr s r!rUAdam.get_configs]#% !%!?!?''"++++<<<<   r#) r%rr-r.r1rrrr)gMbP?g?g+?gHz>FNNNNFgGz?NTr) __name__ __module__ __qualname____firstlineno____doc__rr+rRrU__static_attributes__ __classcell__)r s@r!rrsR8x $ "H!F)KVr#rz{{base_optimizer_keyword_args}})r^tensorflow.compat.v2compatv2r7tf_keras.src.optimizersr'tf_keras.src.saving.object_registrationr tensorflow.python.util.tf_exportr Optimizerrreplacebase_optimizer_keyword_argsrr#r!rjs%!!-O: (0  {9  { {|||##%y'L'L r#