6bi"jSrSSKrSSKJs Jr SSKJr S Sjr S Sjr "SS5r Sr S r S rg) zTraining-related utilities.N) generic_utilsc Sn[U[5(dSnU/n[SU55(asU(dNUVVs/sHoAVs/sH oTXUS-PM snPM nnnUVs/sHn[R"USS9PM nnO4UVs/sHoDUSUSS-PM nnO[ R "X5nU(aUSnU$s snfs snnfs snfs snf)aSlices batches out of provided arrays (workaround for eager tensors). Unfortunately eager tensors don't have the same slicing behavior as Numpy arrays (they follow the same slicing behavior as symbolic TF tensors), hence we cannot use `generic_utils.slice_arrays` directly and we have to implement this workaround based on `concat`. This has a performance cost. Args: arrays: Single array or list of arrays. indices: List of indices in the array that should be included in the output batch. contiguous: Boolean flag indicating whether the indices are contiguous. Returns: Slice of data (either single array or list of arrays). FTc3N# UHn[R"U5v M g7fN)tf is_tensor).0xs \/home/james-whalen/.local/lib/python3.13/site-packages/tf_keras/src/engine/training_utils.py slice_arrays..-s +Fq2<<??Fs#%r)axis) isinstancelistanyrconcatr slice_arrays)arraysindices contiguousconverted_to_listr ientriesslicess r rrs$ fd # #  +F +++@FG1g6g!!e g6GG4;.Os? #1aTM>c3(# UHoSLv M g7frrr s r r r"Rs: -!I~r#temporalrr)rrtuplerrnestassert_same_structure list_to_tupleflatten enumeratenpndarrayshapeappendones)outputssample_weightssample_weight_modescheck_all_flatany_sample_weightpartial_sample_weightnew_sample_weightsrswas_numpyoutput output_shape is_temporalsw_shapes r handle_partial_sample_weightsr>;s  ntUm 4 4*$6 1 Ln6L*$6 3? #1? < !2! c: -: 7  &== 2GGG %% . ) "''//.9 :  %% ' "M"''//'2J$K   * GG ) )#RWW__5H%I >* :!'*bjj9HZF+36<<&9IL$4/9'*j8  a,q/2"1o'   % %%-!27783D   % %b )++. () c*\rSrSrSrSrSrSrSrg)RespectCompiledTrainableStatea]Set and restore trainable state if it has changed since compile. The keras API guarantees that the value of each Layer's `trainable` property at `Model.compile` time will be used when training that model. In order to respect this requirement, it may be necessary to set the trainable value of layers to their compile time values before beginning a training endpoint and restore the values before returning from said endpoint. This scope checks if any layer's trainable state has changed since Model compile, and performs this set and un-set bookkeeping. However, the trainable state of a layer changes quite infrequently, if ever, for many kinds of workflows. Moreover, updating every layer in a model is an expensive operation. As a result, we will only explicitly set and unset the trainable state of a model if a trainable value has changed since compile. c:XlSUlSUlSUlgNF)_model_current_trainable_state_compiled_trainable_state_should_set_trainable)selfmodels r __init__&RespectCompiledTrainableState.__init__s (,%)-&%*"r?cURR5UlURRUlURR 5H1upXR;dMX RU:wdM*SUl O UR (a&URR UR5 gg)NT)rE_get_trainable_staterFrGitemsrH_set_trainable_state)rIlayer trainables r __enter__'RespectCompiledTrainableState.__enter__s(, (H(H(J%)-)N)N&!% > > D D F E666!>!>u!EE-1* !G  % % KK , ,T-K-K L &r?cpUR(a%URRUR5 grD)rHrErPrF)rItype_arg value_arg traceback_args r __exit__&RespectCompiledTrainableState.__exit__s(  % % KK , ,T-J-J Kr?)rGrFrErHN) __name__ __module__ __qualname____firstlineno____doc__rKrSrY__static_attributes__rr?r rArAs + M$r?rAcSnU"U5(a:UR(d [S5eURSnU"U5(aM:[USS5(aURUR4$g)a2Retrieves input shape and input dtype of layer if applicable. Args: layer: Layer (or model) instance. Returns: Tuple (input_shape, input_dtype). Both could be None if the layer does not have a defined input shape. Raises: ValueError: in case an empty Sequential or Functional model is passed. c[US5=(a UR=(d URRS:H$)N_is_graph_network Sequential)hasattrrc __class__r[)rQs r _is_graph_model2get_input_shape_and_dtype.._is_graph_models4 E. / KE4K4K6 __ % % 5 6r?z)An empty Model cannot be used as a Layer.r_batch_input_shapeN)NN)layers ValueErrorgetattrridtype)rQrgs r get_input_shape_and_dtypernsk6 % ||HI I Q % u*D11''44 r?c[U5upUb6[RRR US5R $g)zyGets the static batch size of a Layer. Args: layer: a `Layer` instance. Returns: The static batch size of a Layer. Nr)rnrcompatv1 Dimensionvalue)rQbatch_input_shape_s r get_static_batch_sizervs@5U;$yy||%%&7&:;AAA r?cF[U[5(a [U5$U$)zADatasets will stack the list of tensor, so switch them to tuples.)rrr&) maybe_lists r r)r)s *d##Z  r?)T)F)r_numpyr,tensorflow.compat.v2rpv2rtf_keras.src.utilsrrr>rArnrvr)rr?r r}sE"!!,!JBGIX..h@ r?