x-j5ddlmZddlmZmZddlZddlZddlm Z ddl m Z m Z m Z er ddlmZddlmZ d&d'd%ZdS)() annotations) TYPE_CHECKINGLiteralN) strong_wolfe)_value_and_gradient&check_initial_inverse_hessian_estimatecheck_input_type)Callable)Tensord2:0yE>r?float32objective_funcCallable[[Tensor], Tensor]initial_positionr history_sizeint max_iterstolerance_gradfloattolerance_change initial_inverse_hessian_estimate Tensor | Noneline_search_fnLiteral['strong_wolfe']max_line_search_itersinitial_step_lengthdtypeLiteral['float32', 'float64']name str | Nonereturn(tuple[bool, int, Tensor, Tensor, Tensor]c D  dvrtd dd} t|d| |"tj|jd n"t|d | t ||tj|} t| \}}tj d gdd }tj d gd d }tj d gd d }tj d gd d }tj gd tj d gd d }tj d gdd }|jd}tj d z|f }tj d z|f }tj d zd f }tj d zd f fd}  f d}tj j |||||||| ||||||g ||| ||fS)a Minimizes a differentiable function `func` using the L-BFGS method. The L-BFGS is a quasi-Newton method for solving an unconstrained optimization problem over a differentiable function. Closely related is the Newton method for minimization. Consider the iterate update formula: .. math:: x_{k+1} = x_{k} + H_k \nabla{f_k} If :math:`H_k` is the inverse Hessian of :math:`f` at :math:`x_k`, then it's the Newton method. If :math:`H_k` is symmetric and positive definite, used as an approximation of the inverse Hessian, then it's a quasi-Newton. In practice, the approximated Hessians are obtained by only using the gradients, over either whole or part of the search history, the former is BFGS, the latter is L-BFGS. Reference: Jorge Nocedal, Stephen J. Wright, Numerical Optimization, Second Edition, 2006. pp179: Algorithm 7.5 (L-BFGS). Args: objective_func: the objective function to minimize. ``objective_func`` accepts a 1D Tensor and returns a scalar. initial_position (Tensor): the starting point of the iterates, has the same shape with the input of ``objective_func`` . history_size (Scalar): the number of stored vector pairs {si,yi}. Default value: 100. max_iters (int, optional): the maximum number of minimization iterations. Default value: 50. tolerance_grad (float, optional): terminates if the gradient norm is smaller than this. Currently gradient norm uses inf norm. Default value: 1e-7. tolerance_change (float, optional): terminates if the change of function value/position/parameter between two iterations is smaller than this value. Default value: 1e-9. initial_inverse_hessian_estimate (Tensor, optional): the initial inverse hessian approximation at initial_position. It must be symmetric and positive definite. If not given, will use an identity matrix of order N, which is size of ``initial_position`` . Default value: None. line_search_fn (str, optional): indicate which line search method to use, only support 'strong wolfe' right now. May support 'Hager Zhang' in the future. Default value: 'strong wolfe'. max_line_search_iters (int, optional): the maximum number of line search iterations. Default value: 50. initial_step_length (float, optional): step length used in first iteration of line search. different initial_step_length may cause different optimal result. For methods like Newton and quasi-Newton the initial trial step length should always be 1.0. Default value: 1.0. dtype ('float32' | 'float64', optional): data type used in the algorithm, the data type of the input parameter must be consistent with the dtype. Default value: 'float32'. name (str, optional): Name for the operation. For more information, please refer to :ref:`api_guide_Name`. Default value: None. Returns: output(tuple): - is_converge (bool): Indicates whether found the minimum within tolerance. - num_func_calls (int): number of objective function called. - position (Tensor): the position of the last iteration. If the search converged, this value is the argmin of the objective function regarding to the initial position. - objective_value (Tensor): objective function value at the `position`. - objective_gradient (Tensor): objective function gradient at the `position`. Examples: .. code-block:: python :name: code-example1 >>> # Example1: 1D Grid Parameters >>> import paddle >>> # Randomly simulate a batch of input data >>> inputs = paddle. normal(shape=(100, 1)) >>> labels = inputs * 2.0 >>> # define the loss function >>> def loss(w): ... y = w * inputs ... return paddle.nn.functional.square_error_cost(y, labels).mean() >>> # Initialize weight parameters >>> w = paddle.normal(shape=(1,)) >>> # Call the bfgs method to solve the weight that makes the loss the smallest, and update the parameters >>> for epoch in range(0, 10): ... # Call the bfgs method to optimize the loss, note that the third parameter returned represents the weight ... w_update = paddle.incubate.optimizer.functional.minimize_bfgs(loss, w)[2] ... # Use paddle.assign to update parameters in place ... paddle.assign(w_update, w) .. code-block:: python :name: code-example2 >>> # Example2: Multidimensional Grid Parameters >>> import paddle >>> def flatten(x): ... return x. flatten() >>> def unflatten(x): ... return x.reshape((2,2)) >>> # Assume the network parameters are more than one dimension >>> def net(x): ... assert len(x.shape) > 1 ... return x.square().mean() >>> # function to be optimized >>> def bfgs_f(flatten_x): ... return net(unflatten(flatten_x)) >>> x = paddle.rand([2,2]) >>> for i in range(0, 10): ... # Flatten x before using minimize_bfgs ... x_update = paddle.incubate.optimizer.functional.minimize_bfgs(bfgs_f, flatten(x))[2] ... # unflatten x_update, then update parameters ... paddle.assign(unflatten(x_update), x) )rfloat64z?The dtype must be 'float32' or 'float64', but the specified is .minimize_lbfgsrNrr!rrint64shape fill_valuer!Fboolc | k|zSN) kdone is_convergenum_func_callsvaluexkg1sk_vecyk_vecrhok_vecheadtailrs j/var/www/html/banglarbhumi/venv/lib/python3.11/site-packages/paddle/incubate/optimizer/functional/lbfgs.pycondzminimize_lbfgs..condsI $&&c   tj|} tjg dz d} fd} fd}tjj||| | gtj| }tjg dzd} fd} fd}tjj||| |g| }!d krt#||"  \}}}}ntd !d tj||z|||z}||z }tj ||tjj d kfdfd}tj r| <| <| <nctj  |tj  |tj | dzz dtjj k fdd||z}|}|dz }tj|t j}tj|t j}tj||%kz|$kz|tj||tj||d kz|||||||| g S)Nrr,r-c|kSr2r3)iqai_vecr?s r@rAz*minimize_lbfgs..body..cond 9 rBc btjr(|tj||z||<nCtj|||tj||z}||||zz }|dz }|||fSNr)paddlein_dynamic_modedotstaticsetitemmod)rErFrGrr=r;r<s r@bodyz*minimize_lbfgs..body..bodys%'' $QK&*VAY*B*BBq ..Ax{VZq 1-E-EEF1Iq ))AQ L))Aa< rBrArQ loop_varsc|kSr2r3)rErr>s r@rAz*minimize_lbfgs..body..condrHrBc|tj||z}||||z zz}|dz}||fSrJ)rKrMrP)rErUbetarGrr=r;r<s r@rQz*minimize_lbfgs..body..bodys[A;F1Iq!9!99DF1IT!122AQ L))Aa4KrBr)fr9pkrr r!zNCurrently only support line_search_fn = 'strong_wolfe', but the specified is ''gc4tjdgdS)Nrg@@r-)rKfullr+sr@z.minimize_lbfgs..body..sFKqcfEJJJrBcdz S)Nrr3)rhok_invsr@r]z.minimize_lbfgs..body..s C(NrBc6tj|dz|dSrJ)rKassign)r?s r@true_fnz-minimize_lbfgs..body..true_fn s M$(D ) ) ) ) )rBcSr2r3)r?rbsr@r]z.minimize_lbfgs..body..$sGGDMMrB)p)rKrar\rPrNnn while_loopmatmulrNotImplementedErrorrMrArLrOlinalgnormnpinf)&r4r5r6r7r8r9r:r;r<r=r>r?rFrErArQrUrYalphag2 ls_func_callsskykrhokgnormpk_normr_rbH0rGr!rr rrrrrs& ````` @@r@rQzminimize_lbfgs..bodys" M"   K$( !=!=W           ##DQ6N $    M"a  KbTAXW E E E                ##DQF#KKKR ^ + +.: /$7 /// +E5"mm&raorrr   n}4nEEERZ "W:b"%%}$$ O J J J J " " " "    ! # # CF4LF4L!HTNN]**64<r?r.r;r<r=rArQrurGs` ```` ```` @@r@r*r*$sH *** Ve V V V   G%'97CCC'/ Z(.q1 ? ? ? , .    //OPPP - '..00 1 1B#NB77IE2 1#!7;;;A ;aSU& A A AD+QCEHHHK[sqHHHN;RLPPPL ;aSQg > > >D ;aSQg > > >D  "1 %E \rs#""""")))))))) %%%%%% ((((((  "6:.