AVh`dZddlZddlmZddlmZddlmZddlm Z ddl m Z ddlmZdd l mZdd l mZdd l mZdd lmZdd lmZddlmZddlmZm Z m!Z!ddl"m#Z#ejHdgdZ%edddZ&d4de#e!e&fde#e!ejNfde(de(de)f dZ*edejVe*Z,de#e!e&fde#e!ejNfde(de(de)f dZ-ejHdgdZ.edddZ/d5de#e!e/fde#e!ejNfde)d e(fd!Z0ed"ejVe0Z1de#e!e/fde#e!ejNfde)d e(fd#Z2ejHd$d%d&gZ3ed'ddZ4d6de#e!e4fd(e#e!ejjfd)e#e!ejNfde#e!ejNfd*e)d+e)d,e)fd-Z6ed.ejVe6Z7de#e!e4fd(e#e!ejjfd)e#e!ejNfde#e!ejNfd*e)d+e)d,e)fd/Z8ejHd0d%d&gZ9d6de#e!ejtfd(e#e!ejjfd)e#e!ejNfde#e!ejNfd*e)d+e)d,e)fd1Z;ed2ejVe;Z= 1`. A scalar >= 0 (beam search beam width). top_paths: An `int` that is `>= 1`. A scalar >= 0, <= beam_width (controls output size). merge_repeated: An optional `bool`. Defaults to `True`. If true, merge repeated classes in output. name: A name for the operation (optional). Returns: A tuple of `Tensor` objects (decoded_indices, decoded_values, decoded_shape, log_probability). decoded_indices: A list of `top_paths` `Tensor` objects with type `int64`. decoded_values: A list of `top_paths` `Tensor` objects with type `int64`. decoded_shape: A list of `top_paths` `Tensor` objects with type `int64`. log_probability: A `Tensor`. Has the same type as `inputs`. rrrrN)rrrnamectxT)rrrrrrT)_contextr_thread_local_datais_eagerrTFE_Py_FastPathExecute_CTCBeamSearchDecoderOutput_make_core_NotOkStatusException_opsraise_from_not_ok_status_FallbackException&ctc_beam_search_decoder_eager_fallback_SymbolicException_executemake_int make_bool_op_def_library_apply_op_helpermust_record_gradient _get_attr_int_get_attr_bool_get_attr_typerrecord_gradient)rrrrrr_ctxtld_resulte__op_outputs_attrs _inputs_flats Q/home/dcms/DCMS/lib/python3.12/site-packages/tensorflow/python/ops/gen_ctc_ops.pyctc_beam_search_decoderrEsm<    0h..0$ #\\  11 $dFOj+y:Jg,11':g n  \:* ;7)N%%n6FG.'88v0?+5/=D J!QX QK' ""$C--l;[ ,.>  !12C  %'F::L  fg? Zi !GIJ$7 7' BQK71Q]34 4wq9}~7N N' BQK71Q]34 4wq9}~7N N' ' - -g 6' .E  & &- ##At,,  # #    3 /jI'd>>  # #   s05F33G:G!!G:9G:>HH'&H'zraw_ops.CTCBeamSearchDecoderctj|d}tj|d}|d}tj|d}tj|g|tj tj gtj \}\}tj|tj}||g}d|d|d|d|f} tjd||z|zdz|| ||} tjrtjd || | | d|g| |dz} | dd| dd|zgz| d|zdz} | dd | d d |zgz| d |zdz} tj| } | S) NrrTrr!sCTCBeamSearchDecoderr"rattrsr rrr#)r1r2r3args_to_matching_eager_dtypesfloat32float64r,convert_to_tensorint32rr6r:r(r)) rrrrrrr _attr_TrCrBr=s rDr/r/ns  \:* ;7)N%%n6FG.66xwX_XgXgFjlsl{l{|'9F**?GMMJ//*, *k9NC 2&   4i)6K&7')*7+3?v!$4 1'""$  fg? Zi !GIJ$7 7' BQK71Q]34 4wq9}~7N N' BQK71Q]34 4wq9}~7N N' ' - -g 6' .CTCGreedyDecoderTV_CTCGreedyDecoder_T blank_indexc tjxstj}|j}|jr5 t j |d|||d|d| }t j|}|S|d}tj |d}|d}tj"|d}t%j&d|||||\} } } } | dd}tj(rYd| j+dd| j-dd | j/d f} | j0} tj2d| | |t j|}|S#tj$r }tj||Yd}~nd}~wtj$rYnwxYw t||||||S#tj$rYOwxYw) a7Performs greedy decoding on the logits given in inputs. A note about the attribute merge_repeated: if enabled, when consecutive logits' maximum indices are the same, only the first of these is emitted. Labeling the blank '*', the sequence "A B B * B B" becomes "A B B" if merge_repeated = True and "A B B B B" if merge_repeated = False. Regardless of the value of merge_repeated, if the maximum index of a given time and batch corresponds to the blank, index `(num_classes - 1)`, no new element is emitted. Args: inputs: A `Tensor`. Must be one of the following types: `float32`, `float64`. 3-D, shape: `(max_time x batch_size x num_classes)`, the logits. sequence_length: A `Tensor` of type `int32`. A vector containing sequence lengths, size `(batch_size)`. merge_repeated: An optional `bool`. Defaults to `False`. If True, merge repeated classes in output. blank_index: An optional `int`. Defaults to `-1`. name: A name for the operation (optional). Returns: A tuple of `Tensor` objects (decoded_indices, decoded_values, decoded_shape, log_probability). decoded_indices: A `Tensor` of type `int64`. decoded_values: A `Tensor` of type `int64`. decoded_shape: A `Tensor` of type `int64`. log_probability: A `Tensor`. Has the same type as `inputs`. rQrrSN)rrSrr F)rrrrSrr!)r$rr%r&rr'_CTCGreedyDecoderOutputr)r*r+r,r-r.!ctc_greedy_decoder_eager_fallbackr0r1r3r2r4r5r6r8r7r9rr:)rrrrSrr;r<r=r>r?r@rArBrCs rDctc_greedy_decoderrXs>    0h..0$ #\\  11  $.-Fg(--g6g nN%%n6FG.K!!+}=+'886?+9(3$@!QX QK' ""$ 2 23C DS..}=s  %'F::L L&'; # ) )' 2' .=  & &- ##At,,  # #    . /.!$88  # #   s03EF$E??FFF--GGzraw_ops.CTCGreedyDecoderc|d}tj|d}|d}tj|d}tj|g|tj tj gtj \}\}tj|tj}||g}d|d|d|f}tjdd||||} tjrtjd ||| tj| } | S) NFrrUrSr!sCTCGreedyDecoderrGrQ)r1r3r2rIrJrKrLr,rMrNrr6r:rVr)) rrrrSrr rOrCrBr=s rDrWrWsN%%n6FG.K!!+}=+66xwX_XgXgFjlsl{l{|'9F**?GMMJ//*, nm[#  &   0!L#)s ?' ""$ L&'; # ) )' 2' .rPCTCLosslossgradient TV_CTCLoss_Tlabels_indices labels_valuespreprocess_collapse_repeatedctc_merge_repeated!ignore_longer_outputs_than_inputsctjxstj}|j} | jr9 t j |d|||||d|d|d| } t j| } | S|d}tj |d}|d}tj |d}|d}tj |d}t#j$d|||||||| \} } } }|dd} tj&rjd| j)dd| j)dd| j)dd | j+d f}| j,}tj.d||| t j| } | S#tj$r } tj| |Yd} ~ nd} ~ wtj$rYnwxYw t||||||||| S#tj$rYwxYw) aCalculates the CTC Loss (log probability) for each batch entry. Also calculates the gradient. This class performs the softmax operation for you, so inputs should be e.g. linear projections of outputs by an LSTM. Args: inputs: A `Tensor`. Must be one of the following types: `float32`, `float64`. 3-D, shape: `(max_time x batch_size x num_classes)`, the logits. labels_indices: A `Tensor` of type `int64`. The indices of a `SparseTensor`. `labels_indices(i, :) == [b, t]` means `labels_values(i)` stores the id for `(batch b, time t)`. labels_values: A `Tensor` of type `int32`. The values (labels) associated with the given batch and time. sequence_length: A `Tensor` of type `int32`. A vector containing sequence lengths (batch). preprocess_collapse_repeated: An optional `bool`. Defaults to `False`. Scalar, if true then repeated labels are collapsed prior to the CTC calculation. ctc_merge_repeated: An optional `bool`. Defaults to `True`. Scalar. If set to false, *during* CTC calculation repeated non-blank labels will not be merged and are interpreted as individual labels. This is a simplified version of CTC. ignore_longer_outputs_than_inputs: An optional `bool`. Defaults to `False`. Scalar. If set to true, during CTC calculation, items that have longer output sequences than input sequences are skipped: they don't contribute to the loss term and have zero-gradient. name: A name for the operation (optional). Returns: A tuple of `Tensor` objects (loss, gradient). loss: A `Tensor`. Has the same type as `inputs`. gradient: A `Tensor`. Has the same type as `inputs`. r[rarbrcNrarbrcrr FTrr_r`rrarbrcrr!)r$rr%r&rr'_CTCLossOutputr)r*r+r,r-r.ctc_loss_eager_fallbackr0r1r3r4r5r6r8r9rr:rr_r`rrarbrcrr;r<r=r>r?r@rArBrCs rDctc_lossrjs2H    0h..0$ #\\  11 iv~}7$&:?) +g $$W-g n")#( !)!3!34PRp!q))*<>RS&.(-%&.&8&89Z\'A#'88&!.#20L&85V!QX QK' ""$,  !?@"C$6$67K$L1  !DEs  % 'F ::L <2   )' .W  & &- ##At,,  # #    $ .-'C/,M   # #   s07FG F11G  G G""G98G9zraw_ops.CTCLossc |d}tj|d}|d}tj|d}|d}tj|d}tj|g|tjtj gtj\} \}t j|tj}t j|tj}t j|tj}||||g} d|d|d|d| f} tjdd| | || } tjrtjd | | | tj| } | S) NFraTrbrcr!sCTCLossr#rGr[)r1r3rIrJrKrLr,rMint64rNrr6r:rgr)) rr_r`rrarbrcrr rOrCrBr=s rDrhrhRsc!)#( !)!3!34PRp!q))*<>RS&.(-%&.&8&89Z\'A#66xwX_XgXgFjlsl{l{|'9F)).'--H.(( F-**?GMMJ/.-I, *,H*%'H#  &   Z`. `labels_indices(i, :) == [b, t]` means `labels_values(i)` stores the id for `(batch b, time t)`. labels_values: A `Tensor` of type `int32`. The values (labels) associated with the given batch and time. sequence_length: A `Tensor` of type `int32`. A vector containing sequence lengths (batch). preprocess_collapse_repeated: An optional `bool`. Defaults to `False`. Scalar, if true then repeated labels are collapsed prior to the CTC calculation. ctc_merge_repeated: An optional `bool`. Defaults to `True`. Scalar. If set to false, *during* CTC calculation repeated non-blank labels will not be merged and are interpreted as individual labels. This is a simplified version of CTC. ignore_longer_outputs_than_inputs: An optional `bool`. Defaults to `False`. Scalar. If set to true, during CTC calculation, items that have longer output sequences than input sequences are skipped: they don't contribute to the loss term and have zero-gradient. name: A name for the operation (optional). Returns: A tuple of `Tensor` objects (loss, gradient). loss: A `Tensor` of type `float32`. gradient: A `Tensor` of type `float32`. rmrarbrcNreFTrf)r$rr%r&rr'_CTCLossV2Outputr)r*r+r,r-r.ctc_loss_v2_eager_fallbackr0r1r3r4r5r6r8rr:ris rD ctc_loss_v2rqrs#J    0h..0$ #\\  11 k47$&:?) +g !&&w/g n")#( !)!3!34PRp!q))*<>RS&.(-%&.&8&89Z\'A#'88F>#0%42N(:7X !QX QK' ""$,  !?@"C$6$67K$L1  !DE GF ::L \674  " "7 +' .U  & &- ##At,,  # #    ' .-'C/,M   # #   s07E22F9F  F98F9=GG('G(zraw_ops.CTCLossV2c |d}tj|d}|d}tj|d}|d}tj|d}tj|tj }tj|tj }tj|tj}tj|tj}||||g} d|d|d|f} tjdd| | ||} tjrtjd | | | tj| } | S) NFraTrbrcs CTCLossV2r#rGrm) r1r3r,rMrJrKrlrNrr6r:ror)) rr_r`rrarbrcrr rCrBr=s rDrprps?!)#( !)!3!34PRp!q))*<>RS&.(-%&.&8&89Z\'A#  ! !&'// :&)).'--H.(( F-**?GMMJ/.-I, *,H*%'H J&   \1\#)s ?' ""$ \674  " "7 +' .rP)TN)FrUN)FTFN)>__doc__ collectionstensorflow.pythonrtensorflow.python.eagerrr$rr*rr1tensorflow.python.frameworkrrJtensorflow.security.fuzzing.pyr_atypesr _op_def_registryr r,r r4"tensorflow.python.util.deprecationr tensorflow.python.utilr _dispatch tensorflow.python.util.tf_exportrtypingrrrtyping_extensionsr namedtupler(rInt32intboolrE to_raw_oprr/rVrRrXrQrWrgr^Int64rjr[rhroFloat32rqrmrprPrDrs 67179FK3IC86%%'4k44MO $$?ARTefJIc3L.L$MJ`ijmovo|o|j|`}JLOJ\_JptJXAy!?@PgAhi9SB[=[3\oxy|FLLzLpM[^kn@D.1+00MO  79JL]^Fy.C)CDFW`adfmfsfsasWtFFJF^aFP9978HZ9[\i=R8R.Sfopsu|vCvCqCgDVZil((''  Z ~'8:KL [YsL01[9SRYR_R_M_C`[qz{~AHANAN|NrO[bkloqxq~q~l~b[^b[}A[jn[z ')% &~t~~h'? @Ic<.?$@R[\_ahanan\nRoAJKNPWP]P]K]A^qz{~@G@M@M{MqNnrHLqu6*;)) Z [ #w"67[SVX_XeXeSeIf[xABEGNGTGTBTxU[hqruw~wDwDrDhE[dh[CG[pt[z +I) *>4>>++F G ygoo1E'FXabegngtgtbtXuGPQTV]VcVcQcGdw@ADFMFSFSASwTtxNRw{rP