2VhOddlZddlZddlZddlmZddlmZddlm Z ddlm Z dZ dZ dZ d Zd Zd Zd Zd Zd;dZdZdZddZdZd?dZd>dZd@dZdZ d;dZ!dZ"d@dZ#d@dZ$d@d Z%d!Z&d"Z' dAd#Z( dAd$Z)d%Z* dBd&Z+ dBd'Z, dBd(Z- dCd)Z.dDd*Z/dDd+Z0d,Z1dEd-Z2dEd.Z3dFd/Z4dGd0Z5d1Z6d2Z7 dHd3Z8 dId4Z9 dJd5Z:d6Z;d7Z dKd:Z?y)LN)backend)#compute_conv_transpose_output_shape)cast)convert_to_tensorc@tjj|SN)tfnnreluxs O/home/dcms/DCMS/lib/python3.12/site-packages/keras/src/backend/tensorflow/nn.pyr r  55::a=c@tjj|Sr)r r relu6r s rrrs 55;;q>rcV|}tjj|}||_|Sr)r r sigmoid _keras_logits)r logitsoutputs rrrs& F UU]]1 F!F Mrc t|}tj|dktjd|jtj|dk\tjd|jd|dzzS)Ndtype??)rr whereconstantrr s rsparse_sigmoidr"sa!A 88 R Cqww' aS8#Q-H rc@tjj|Sr)r r tanhr s rr$r$&rrcF|tjj|z Sr)r mathr$r s r tanh_shrinkr'*s rww||A rc@tjj|Sr)r r&softplusr s rr)r).s 77  A rc@tjj|Sr)r r softsignr s rr+r+2s 55>>! rc tj||kD||z tj|| k||ztj|Sr)r r zeros_liker thresholds r soft_shrinkr06sD 88 I  I  iZY a0@A rc tj|dktj|tj|dkdtj|dzdz|S)Nrrg?)r r r-powr s r sparse_plusr4>sN 88 R a Q"&&Q"22A6 rc@tjj|Sr)r r silur s rr6r6Frrct|}t||j}|tjtj||zz}|dz S)Nrr2)rrr sqrtsquare)r bys r squareplusr<JsE!A!177+A BGGBIIaL1$ %%A q5Lrc@tjj|Sr)r r& log_sigmoidr s rr>r>Qs 77  q !!rcDtjj||S)N)alpha)r r leaky_relu)r negative_slopes rrArAUs 55  A^  44rct|}t|tjd|jztjd|jz S)Ng@g@)rrr r!rr s r hard_sigmoidrDYs?!A R[[agg.. /"++c1772K KKrc|t|zSr)rDr s r hard_silurF^s |A rctjj|}|dk(r|Stj|dkD|||zS)Nrr)r r elur )r r@ress rrHrHbs: %%))A,C z xxAsECK00rc@tjj|Sr)r r selur s rrKrKjrrcZt|}tjj||S)N) approximate)rr r gelu)r rMs rrNrNns#!A 55::a[: 11rctj|d|tjjtj|d|z zzS)Nr)r maximumr&expm1minimum)r r@s rcelurSssC ::a   1cU")! rc|j|dzdk7rtd|jd|tj|d|\}}|tj|zS)Nr2rz4axis size must be divisible by 2. Received: x.shape=z with axis=)num_or_size_splitsaxis)shape ValueErrorr splitr)r rVx1x2s rglur\yshwwt}qA !!" TF <  XXaAD 9FB  2 rc2tj|ddS)Ngr)clip_value_minclip_value_max)r clip_by_valuer s r hard_tanhras  Ad3 GGrctjtj||kD|tj|Sr)r r absr-r.s r hard_shrinkrds+ 88BFF1I )1bmmA.> ??rc6tj||kD||Sr)r r )r r/ default_values rr/r/s 88A M1m 44rc"|}|btj|dg}tjj|d}tj|tj|}n!tjj||}||_|SNrrV)r reshaper softmaxrWr)r rVrrs rrkrksq F |At$vB/FBHHQK0qt,!F Mrc |atj|dg}tjj|d}tj|tj|Stjj||Srh)r rjr log_softmaxrW)r rVrs rrmrmsg |At$""6"3zz&"((1+.. 55  QT  **rc|t|}tj|d|}tj||}tjdtj ||dz|j }dgt|j z}d||<tj||}||dz |z z dkD}tj||d}tjtj||j |d }tj||d dz |z } tj|| z d} | S) N DESCENDING) directionrVrirrrrrTrVkeepdims) rr sortcumsumrangerWrlenrjr reduce_sumrrP) rrV logits_sorted logits_cumsumrr_shapesupportlogits_cumsum_safektaurs r sparsemaxrs v &FGGFlFMIIm$7M BHHV$T*Q.fllCAcC %%GGDM 1gA}q0A559G'=#> bgggv||44$OA ==+$ F Ja OC ZZ c *F Mrct|jdz }|dk(rtj|d}|S|dk(rtj|d}|S|dk(rtj|d}|St d|jd) Nr2rrr2r)rr2rr)rr2rrzePooling inputs's shape must be 3, 4 or 5, corresponding to 1D, 2D and 3D inputs. But received shape: .)rvrWr transposerX)inputsnum_spatial_dimss r_transpose_spatial_inputsrs6<<(1,1fi0 M Q fl3 M Q fo6 M  228,,q B  rct|jdz }|dk(rtj|d}|S|dk(rtj|d}|S|dk(rtj|d}|S)Nr2rr)rrrr2r)rrrr2r)rvrWr r)outputsrs r_transpose_spatial_outputsrsu7==)A-1,,w 2 N Q ,,w 5 N Q ,,w8 Nrc tj|}||n|}|j}tdt |j }|dk(r t |}tjj|||||}|dk(r t|}|SN channels_lastchannels_first) rstandardize_data_formatupper_convert_data_formatrvrWrr r max_poolrr pool_sizestridespadding data_formattf_data_formatrs rrr11+>K"?iGmmoG)/3v||;LMN&&+62eenn G&&,W5 Nrc tj|}||n|}|j}tdt |j }|dk(r t |}tjj|||||}|dk(r t|}|Sr) rrrrrvrWrr r avg_poolrrs r average_poolrrrc|dk(r!|dk(ry|dk(ry|dk(rytd|d |d k(r!|dk(ry |dk(ry |dk(ry td|d td|d)NrrNWCrNHWCNDHWCzInput rank not supported: z. Expected values are [3, 4, 5]rNCWNCHWNCDHWzInvalid data_format: z9. Expected values are ["channels_first", "channels_last"])rX)rndims rrrso% 19 QY QY,TF300  ( ( 19 QY QY,TF300  #K=1F F  rc\ fd tjd fd}dk(xrtjdk(}t j dk(rjd}njd }|xs|jd k7}|r|S S) Ncttj}tjj j |S)Nr dilations)rrvrWr r convolutionr)rr dilation_raterkernelrrs r_convzconv.._conv7sL-k3v||;LMuu     MMO&# !  rT) jit_compilecSr)rsr _conv_xlazconv.._conv_xlaDs wrrrrrr)r functionrvrWrr) rrrrrrr needs_xlachannelsrs `````` @rconvr/s   [[T"#//JC 4E4JI11+>Ko%<<#<<?9Xb)99I{wrctj|}t|jdz }|dkDrt d|j dt |d}|j}t|tr|f|z}t|tr|f|z}|dk(r|dk(rd|dzzdz}d}n d|dzz}d}tj||}tj|d }|dnd|z}tjj|||||| } tj| |gS|dk(r d|zdz}d}nd|z}d}tjj|||||| S) Nr2z<`inputs` rank must be 3 (1D conv) or 4 (2D conv). Received: rrrrrrrrrir)rrrvrWrXrrr isinstanceintr expand_dimsr depthwise_conv2dsqueeze) rrrrrrrrspatial_start_dimrs rdepthwise_convrWs11+>K6<<(1,! J{{m1   *+q9NmmoG'3*//-%&(+;; 1 / )Wq[(4/G ! w{*G ! (9:Q/ - 54-;O %%((    &# ) zz'$5#677o%.4'7" 55 ! !" " rc tj|}t|jdz }|dkDrt d|dt |d}|j }t|tr|f|z}t|tr|f|z}|dk(r|dk(rd|dzzdz}d} n d|dzz}d} tj|| }tj|d }tj|d }|dnd|z}tjj||||||| } tj| | gS|dk(r d|zdz}nd|z}tjj||||||| S) Nr2z>`num_spatial_dims` must be 1 or 2. Received: num_spatial_dims=rrrrrrrrir)rrrvrWrXrrrrr rr separable_conv2dr) rdepthwise_kernelpointwise_kernelrrrrrrrrs rseparable_convrs11+>K6<<(1,! 01 4  *+q9NmmoG'3*//-%&(+;; 1 / )Wq[(4/G ! w{*G ! (9:>>*:C>>*:C - 54-;O %%((     &#) zz'$5#677o%.4'7" 55 ! !" " rc tj|}t|t|j}|jdd}|jd} t |j} t j|} t| D]\} } |  | | | | <t| || |||||}t jj|||||j||S)Nr)rrr) rrrrvrWlistr enumeraterr conv_transposer)rrrroutput_paddingrrr kernel_sizefilters input_shapesymbolic_shapeie output_shapes rrrs11+>K)+s6<<7HIN,,s#Kll2Gv||$KXXf%N+&/1 9+A.KN/7 L 55   "  rct|d}|d}ntj|}|r|dkr|t|jzdz}t j |j}tj||f}tjtj|d|}|jDcgc]}tj|}}tj|ddi}|j|tj|d|D cgc]} tj| |df}} |D cgc]&} tj| tj(}} tj |d}t#|j} | j||tj$||| S|d k(rd nd \} } tj&||| | || Scc}wcc} wcc} w) Nint64rfloat32rrindexingijribool)TF)NN)on_value off_valuerVr)rrstandardize_dtypervrWr&prodr rjr greater_equalrumeshgridinsertrPrconcatr SparseTensorone_hot) r num_classesrVrsparse values_countvaluesdimindicesarWrrs rrrs!7+A }))%0  !8#agg,&*Dyy) A /))&!4EB,-GG4S288C=44++w66tRZZ1-.=DE2::a,!23EE189A2771bhh'99))G!,QWW  T;'w66+0F?- Hi ::    5F9s:G%G*<+G/c<t|jdkDrdnd}tj|dk(r|rgt |||dd}t j j||d}|j}t j||}|j||St ||||}t j|| S|r2t ||||d}t j j||dSt ||||}t j|| S) Nrrrint8T)rVrr)rVoutput_is_sparse)rVrri) rvrWrrrr r reduce_maxr set_shape reduce_any)r rrVrrreduction_axisr outputs_shapes r multi_hotr sagg,*QN  '61 ;TGii**nt+G$MMMgggu-G   m ,Na4uEG==~> > ;TtG99''nt( a4uEG==~> >rc|}|}t|d}|r|j}d}t|dxrPt|tjj tj f xr|jj|k(xr| }|r?t|jjdk(sJ|jjd}d}|r"|s|rtjd|d|dd ||fS) zCRetrieves logits tensor from maybe-softmax or maybe-sigmoid tensor.rToprrz"`zK` received `from_logits=True`, but the `output` argument was produced by a zB activation and thus does not represent logits. Was this intended?r2) stacklevel) hasattrrrr __internal__ EagerTensorVariablertypervrwarningswarn)r from_logitsop_typefn_nameoutput_ from_logits_has_keras_logitsfrom_expected_op_types r _get_logitsrBsGLv7&&   &6BOO$?$?#MN N & IINNg %   699##$)))))""1% (,A  77>i@! !   L  rctj|}tj|}t|jdkr%t d|jd|jt|jt|jk7r%t d|jd|jt |j|jD]5\}}| | ||k7st d|jd|jt ||dd\}}|r"tjj|||S|tj||d z }tj|tjd tjz }tj|tjj|z| S) a Categorical crossentropy between an output tensor and a target tensor. Args: target: A tensor of the same shape as `output`. output: A tensor resulting from a softmax (unless `from_logits` is `True`, in which case `output` is expected to be the logits). from_logits: Boolean, whether `output` is the result of a softmax, or is a tensor of logits. axis: Int specifying the channels axis. `axis=-1` corresponds to data format `channels_last`, and `axis=1` corresponds to data format `channels_first`. Returns: Output tensor. Example: >>> a = tf.constant([1., 0., 0., 0., 1., 0., 0., 0., 1.], shape=[3,3]) >>> print(a) tf.Tensor( [[1. 0. 0.] [0. 1. 0.] [0. 0. 1.]], shape=(3, 3), dtype=float32) >>> b = tf.constant([.9, .05, .05, .05, .89, .06, .05, .01, .94], ... shape=[3, 3]) >>> print(b) tf.Tensor( [[0.9 0.05 0.05] [0.05 0.89 0.06] [0.05 0.01 0.94]], shape=(3, 3), dtype=float32) >>> loss = categorical_crossentropy(a, b) >>> print(np.around(loss, 5)) [0.10536 0.11653 0.06188] >>> loss = categorical_crossentropy(a, a) >>> print(np.around(loss, 5)) [0. 0. 0.] rzPArguments `target` and `output` must be at least rank 1. Received: target.shape=, output.shape=WArguments `target` and `output` must have the same rank (ndim). Received: target.shape=QArguments `target` and `output` must have the same shape. Received: target.shape=Softmaxcategorical_crossentropy)labelsrrVTrrr)r rrvrWrXziprr !softmax_cross_entropy_with_logitsrwr`repsilonr&log)targetrrrVe1e2s rr r esN ! !& )F  ! !& )F 6<<1 "LL> H   6<<C -- "LL> H  fllFLL1B >bnr & ~_V\\NL & Y(BFKuu66&t7  bmmFD4@ @F  !3):#:F MM&277;;v#66 = ==rcj|dk7r)|t|jdz k7rtd|t||dd\}}t j |}t j |d}t j |}t|jt|jk(r)|jddk(rt j|d}t|jdkrtd |jt|jt|jd dk7r%td |jd |jt|j|jd dD]5\}}| | ||k7std |jd |j|s]t j|tjdtjz }tjj|}tjj||}|S)aNCategorical crossentropy with integer targets. Args: target: An integer tensor. output: A tensor resulting from a softmax (unless `from_logits` is True, in which case `output` is expected to be the logits). from_logits: Boolean, whether `output` is the result of a softmax, or is a tensor of logits. axis: Int specifying the channels axis. `axis=-1` corresponds to data format `channels_last`, and `axis=1` corresponds to data format `channels_first`. Returns: Output tensor. rrz4Only axis=-1 is currently supported. Received: axis=r sparse_categorical_crossentropyrrrizBArgument `output` must be at least rank 1. Received: output.shape=NzRArgument `output` must have rank (ndim) `target.ndim - 1`. Received: target.shape=rzcArguments `target` and `output` must have the same shape up until the last dimension: target.shape=r r)rvrWrXrr rrrrr`rrr&rr (sparse_softmax_cross_entropy_with_logits)rrrrVrrresults rrrs " rzdc&,,/!33B4& I  & Y(IFK ! !& )F WWV7 +F  ! !& )F 6<<C --&,,r2Ba2GF, 6<<1 "LL> +   6<<C Sb 122 "LL> H  fllFLL"$56B >bnr & ~_V\\NL  !! GOO%q7??+<'< V$ UU ; ;f<F Mrc*tj|}tj|}t|jt|jk7r%t d|jd|jt |j|jD]5\}}| | ||k7st d|jd|jt ||dd\}}|r!tjj||Stj|tjdtjz }|tjj|z}|d|z tjjd|z zz }| S) apBinary crossentropy between an output tensor and a target tensor. Args: target: A tensor with the same shape as `output`. output: A tensor. from_logits: Whether `output` is expected to be a logits tensor. By default, we consider that `output` encodes a probability distribution. Returns: A tensor. rrr Sigmoidbinary_crossentropyrrr)r rrvrWrXrrr !sigmoid_cross_entropy_with_logitsr`rrr&r)rrrrrbces rrrsx ! !& )F  ! !& )F 6<<C -- "LL> H  fllFLL1B >bnr & ~_V\\NL & Y(=FKuu66&7    !3):#:F 277;;v& &CAJ"''++a&j1 11C 4Krcd}tj|j}|dvrd}t|d}|rt |||\}}nt |||\}}|rt j|t jjt jj}t j|t jjt jj}t||}t||}||fS)NF)float16bfloat16Tr) rrrr_compute_moments_sync_compute_momentsr r`r minmax)r axesrr synchronized need_cast ori_dtypemeanvariances rmomentsr,'sI))!''2I++ I .q$Ah)!T8<hbjjnnbjjnnE##HbjjnnbjjnnMD)$), >rctjj}|s t|||Stj|d}tj ||d}tj tj ||d}tj ||d}|jtjjj|}|jtjjj|}|jtjjj|} tjj|| } tjj|| } tj| tj | z d} |s,tj| |} tj| |} | | fS)Ncount)nameTrqr)r distributeget_replica_contextr# ones_likerwr9 all_reduceReduceOpSUMr& divide_no_nanrPr) r r&rr replica_ctx local_count local_sumlocal_squared_sumy_sum y_squared_sum count_sumr*y_squared_meanr+s rr"r">sY--335K 422,,qw/K adT:I biilM-- $FK  " "2==#9#9#=#=y IE** ""$5M&&r}}'='='A'A;OI 77  2DWW**=)DNzz.299T?:C@H zz$%::h- >rcFtjj|||S)Nr )r r r,)r r&rrs rr#r#^s 55==D8= 44rcd|dk7rdgt|jz}|jd||<tj||}tj||}|tj||}|tj||}tjj ||||||S)Nrrr)r r*r+offsetscalevariance_epsilon)rvrWr rjr batch_normalization)r r*r+rVrArBrrWs rrDrDbs rzc!''l"jjmd zz$&::h.  ZZ.F  JJue,E 55 $ $  % rcPt|}t|}tj|d}tj|j d}|dk(rdn|}tj||}tj j|||||d}tj||S)Nint32rrfloat64F)r r label_length logit_length blank_indexlogits_time_major)rr rr result_typerr ctc_loss)rr target_length output_length mask_index result_dtype compute_dtypelosss rrMrMysv &F v &F WWV7 +F &&v||Y?L!-!:I M WWV] +F 55>>""  D 774 &&rct|}tj|}|d|d} }tj|d}t j |j d} tj|| }t|d}|dk(r'tjj||||\} } nx|d k(rd|;|d d|f} |d ||dzf}|d |dzdf}tj| ||gd }tjj|||| \} } ntd|dg}| D]_}tj|j|j|| f}|j!tj"j%|d atj&|d }tj|d}|d k(r,|*|dkr||d z}tj(||k\|dz|}|| fS)Nrr)rrr2rrFrgreedy)rsequence_lengthmerge_repeatedrJ beam_search.rri)rrV beam_width top_pathszInvalid strategy z2. Supported values are 'greedy' and 'beam_search'.)sp_inputrf)rr rWrrrLrrr ctc_greedy_decoderrctc_beam_search_decoderrXrrrappendrto_densestackr )rsequence_lengthsstrategyrYrZrWrPr num_samples num_stepsrdecodedscores inputs_before inputs_mask inputs_after decoded_densests r ctc_decoderls'v &F((6"K(^[^K \\&) ,F    i 8E WWVU #F()9I8EE44,)" 5 & ] "  !"3  #34M j:>&A!ABK!#zA~'7"78LYY k:FEE99,! : &z** *  MP __RZZ[)4L MRYY//2/NOPHH]3MGGM73M= Z%; >#k"o5J Z '):M  &  rcBddlm}|j|jk7r&td|jd|jdt ||j }t jt j||z }d||zd||zz }|S) Nr)log10z Input shapes z and z" must match for PSNR calculation. r ) "keras.src.backend.tensorflow.numpyrnrWrXrrr reduce_meanr9)rZr[max_valrnmsepsnrs rrurus8 xx288BHH:U288*5+ +   rxx8G ..27+ ,C g eCj 0D Krcrtj|}|dk(rdnd}tj|dz|S)Nr g@g̓$Ggffffffr)rrr r!)rvals r_get_large_negativerxs5  % %e ,Ei''ZC ;;sTz //rc||s|Stj|d}|tj||}|rytj|}|d|d}}tjj tj ||fddd}|ddddddf}tj||}tj||t|j}|S)Nrrr2rrr) r r2 logical_andrWlinalg band_partonesr rxr)rmask is_causal combined_mask logits_shapeTS padded_logitss r _apply_masksrs |I LLv6M }d; xx' A Q1yy""277Aq66#:BBD$1$%}d; HHv26<<@M rctj|jd}tjd||d}tj ||}tj |tj ||j}|4tj|tj ||j}t|||} tj| jd} tj tjjtj | | d|j} tjd| |dS)NrzBTNH,BSNH->BNTSoptimal)optimizerrizBNTS,BSNH->BTNH) rrLrr einsumrmultiplyaddrr rk) querykeyvaluebiasr~rrB logits_dtyperr probs_dtypeprobss r_dot_product_attention_xlars&&u{{I>L YY(%y IF WWV\ *F [[ != >F fll ;< y9M%%m&9&99EK GG  bggm[9 CSYY E 99&uy IIrc |d}|r tdt|}t|}t|}t|jdk7r3td|jd|jd|jdt j|d} |,d t j t j | d z n|}t|||||||S) NFz7Flash attention is not supported in tensorflow backend.rzG`dot_product_attention` only supports 4D inputs. Received: query.shape=z , key.shape=z, value.shape=rrrr)rXrrvrWr r8rr) rrrrr~rBrflash_attentionattn_logits_soft_capHs rdot_product_attentionrs E   e $E C C e $E 5;;1 %%*[[Mcii[I ;;-q *   bA6;mS2772771i01 1E % sE4y% r)r)r)g?)r)T)r)NvalidN)rrNr)rrNNr)rrF)Fr)F)FF)NNgMbP?)r)rUdrTr)NNNFNN)@r&r tensorflowr keras.srcr&keras.src.backend.common.backend_utilsr!keras.src.backend.tensorflow.corerrr rrr"r$r'r)r+r0r4r6r<r>rArDrFrHrKrNrSr\rardr/rkrmrrrrrrrrrrrrrr rrr,r"r#rDrMrlrurxrrrrrrrs 3?"5L 12  H@5 +$$    >   8 F   %V   ;D  <D  %P!H?D !FP>f<~-`.@5 ?C8 '@>!B 0 *J.  "r