1VhTdZddlZddlmZddlmZmZmZdZ dZ dZ dZ d Z d Zd Zd Zd ZdZdZdZdZdZdZiej0e ej2e ej4eej6eej8eej:eej<eej>eej@eejBeejDeejFeejHeejJeejLeejNjPeejRjPeejTjPeejVjPeejXjPeiZ-ej\jPe iZ/dZ0dZ1dZ2dZ3e4ej2ej4ej6ej:ejDejBej8ejNejRejTejVejXej>ej@gZ5GddZ6dZ7dZ8dZ9dZ:dZ;dZiej~e>gejHec||jj|Dcgc]}|jc}ycc}wr)bool_valrrrs r SlowAppendBoolArrayToTensorProtorO[s*  L!Aq!&&(!AB!Ar)c|jD]\}}||k(s |cStjD]\}}||k(s |cSyr)items BACKUP_DICT) dtype_dictr keyvals rGetFromNumpyDTypeDictrV|sY$$&S %<J %%'S %<J rc|jtjk(s|jtjk(rtSt t |Sr)typer bytes_str_rLrV_NP_TO_APPEND_FNr s rGetNumpyAppendFnr\s7  zzRYY%**"711 !15 99rcrt|ttfsgS|sdgSt|gt |dzS)z9Returns the inferred dense dimensions of a list of lists.r) isinstancelisttuplelen_GetDenseDimensions) list_of_listss rrbrbs= mdE] 3 s M"#&9-:J&KKKrc#xKt|ttfr|D]}t|D]}|y|ywr)r^r_r`_FlattenToStrings)nested_stringsinnerflattened_strings rreresF.4-0# 'E$5e$< ' && ' 's8:ceZdZdZdZy)_Messagec||_yr_message)selfmessages r__init__z_Message.__init__s  rc|jSrrl)rns r__repr__z_Message.__repr__s }}rN)__name__ __module__ __qualname__rprrrrrjrjs  rrjc|D]}||cSyrrv)lrs r _FirstNotNonerys  =H rc | tdS|S)NNone)rj)rBs r_NotNoner|syrc vt|ttfs|St|trtd|Dsyt|trCtd|Ds1t |Dcgc]}t|ttfrdn|c}St |Dcgc] }t |c}Scc}wcc}w)Nc3HK|]}t|ttfywrr^r_r`.0rs r z_FilterTuple..;A:a$/; "c3HK|]}t|ttfywrrrs rrz_FilterTuple..rr)r^r_r`anyry _FilterTuplerBrs rrrs a$ '!U;;;!T;;; FGHAe}51<H  15a,q/5 66I5s '!B1B6ct|ttfr"t|Dcgc] }t |c}St|t j tjfrdSt|Scc}wr) r^r_r`ry _FilterIntrintegral_typesr Dimensionr|rs rrrsd!dE]#Q7jm788 a&//1G1GH I a[8sA0ct|ttfr"t|Dcgc] }t |c}St|t j rdSt|Scc}wr)r^r_r`ry _FilterFloatr real_typesr|rs rrrsM!dE]#q9!l1o9::a!2!234D!D:A ct|ttfr"t|Dcgc] }t |c}St|t j rdSt|Scc}wr)r^r_r`ry_FilterComplexr complex_typesr|rs rrrsN!dE]#;AnQ/;<<a!5!564GHQKG4>+B+B D  c  t|tjr|S|rtj|}|tj tj tjtjtjfv}t|tjtjfr#|r|j|j}n|}ntt!|ddstt!|ddt"rtj$||}|}nx| t'd|r|j(r |j}nd}|@tj*|tj,dk(rtj.||}nwt1||tj2||}t5|j6t9|k7r3|s1t'd|dt5|j6d t9|d |j:tj<k(r"| |jtj>}nV|j:tj,k(r9|7|jtj@}tjB||r|}tj|j:}|tEd |j:z|r|}|CtG|d r|jH|jHk7rtEd |d|j:d|||j6}d} |jJ} n|D cgc] } tM| }} tj*|tj,} | |jJk(} |r=|j6tO|k(s%tEdtO|d|j6d |jJ| kDrt'd| |jJfztj|jPtSjT|jW} | rK|tXvrC| dkDr>|jJ|jZzdk\r t'd|j]| _/| S|tj`k(ret|tjsKtc|} | Dcgc]}tejf|}}| jjjm|| S|jo} tq| j:}|tEd|jrz|| | | Scc} wcc}w#tD$rtEdti|d|dwxYw)aCreate a TensorProto. Args: values: Values to put in the TensorProto. dtype: Optional tensor_pb2 DataType value. shape: List of integers representing the dimensions of tensor. verify_shape: Boolean that enables verification of a shape of values. Returns: A `TensorProto`. Depending on the type, it may contain data in the "tensor_content" attribute, which is not directly useful to Python programs. To access the values you should convert the proto back to a numpy ndarray with `tensor_util.MakeNdarray(proto)`. If `values` is a `TensorProto`, it is immediately returned; `dtype` and `shape` are ignored. Raises: TypeError: if unsupported types are provided. ValueError: if arguments have inappropriate values or if verify_shape is True and shape of values is not equals to a shape from the argument. make_tensor_proto accepts "values" of a python scalar, a python list, a numpy ndarray, or a numpy scalar. If "values" is a python scalar or a python list, make_tensor_proto first convert it to numpy ndarray. If dtype is None, the conversion tries its best to infer the right numpy data type. Otherwise, the resulting numpy array has a convertible data type with the given dtype. In either case above, the numpy ndarray (either the caller provided or the auto converted) must have the convertible type with dtype. make_tensor_proto then converts the numpy array to a tensor proto. If "shape" is None, the resulting tensor proto represents the numpy array precisely. Otherwise, "shape" specifies the tensor's shape and the numpy array can not have more elements than what "shape" specifies. __array__N__array_interface__r zNone values not supported.rz!Argument must be a dense tensor: z - got shape z , but wanted .zUnrecognized data type: %s base_dtypezInconvertible types: z vs. z . Value is TzExpected Tensor's shape: rz>Too many elements provided. Needed at most %d, but received %d)r rlz>Cannot create a tensor proto whose content is larger than 2GB.z!Failed to convert object of type z to Tensor. Contents: z0. Consider casting elements to a supported type.z-Element type not supported in TensorProto: %s):r^r TensorProtoras_dtypeqint8quint8qint16quint16qint32r ndarraygenericastyper!callablegetattrdictr ValueErroris_numpy_compatibleprodint64emptyrarrayr_shaperbr float64float32int32 array_equalrhasattrrsizeintr`as_datatype_enumras_shapeas_proto_TENSOR_CONTENT_TYPESitemsizetobytestensor_contentstringrerrKrXrJrravelr\r)rr r verify_shape is_quantizednparraynp_dtdowncasted_array numpy_dtype is_same_size shape_sizedimrrr str_values append_fns rmake_tensor_protor$sV&*001  &    L&2::rzz23 mmE$8$89GG '&+t4 5-t4d: **V51 >9: : U..((EE  bhh!?1!DhhuE2G vu -hhvU3GW]]#':6'BB$ tGMM24G4OQ MMRZZ 'U]nnRZZ0Gmmrxx'U]&~~bhh7 ~~.8*//'--0K4w}}DEE   E< (   {55 5gmmV -   }  \\ %*+cS++WWU"((3 !W\\1 ==E%L0U|W]]4 <<* $Pw||,-  ))**!**51::<L  '<<a <<'** *w 7P '.oo&7 #fmm#Jvrzz,J(0  6BC&//!,CJC &&z2==?L !3!34I ;k>N>N N  lL) G,^D &*&\6;  s$7T8/U3T=U=U%U'c 4|jjDcgc]}|j}}tj|tj }t j|j}|j}|jr>tj|j|jj|S|t jk(s|t jk(rt!|j"dk(rhtj$|j"dtj&}|j|_tj(||j|Stj*|j"tj&}|j|_|j|S|t j,k(rt!|j.dk(rGtj(tj$|j.d||j|Stj*|j.|j|S|t j0k(rt!|j2dk(rGtj(tj$|j2d||j|Stj*|j2|j|S|t j4t j6t j&t j8t j:t j<t j>t j@t jBt jDf vrt!|jFdk(rGtj(tj$|jFd||j|Stj*|jF|j|S|t j k(rt!|jHdk(rGtj(tj$|jHd||j|Stj*|jH|j|S|t jJk(rt!|jLdk(rGtj(tj$|jLd||j|Stj$tO|jL|j|S|t jPk(rtS|jT}t!|jTdk(r^tj(tj$tW|jTd|jTd||j|Stj$tY||Dcgc]}tW|d|dc}|j|S|t jZk(rtS|j\}t!|j\dk(r^tj(tj$tW|j\d|j\d||j|Stj$tY||Dcgc]}tW|d|dc}|j|S|t j^k(rt!|j`dk(rGtj(tj$|j`d||j|Stj*|j`|j|Stcd|jzcc}wcc}wcc}w)aCreate a numpy ndarray from a tensor. Create a numpy ndarray with the same shape and data as the tensor. Args: tensor: A TensorProto. Returns: A numpy array with the tensor contents. Raises: TypeError: if tensor has unsupported type. r rrzUnsupported tensor type: %s)2rrrr rrrrr r!r frombuffercopyreshaper r rarrrrepeatfromiterrr&rr+ruint8int16int8rrrrrr0r3rrJr_ complex64iterr>complexzip complex128rGrrNr) tensordr num_elements tensor_dtyper tmpitrs r make_ndarrayrs$0044 5QVV 5E 57751L??6<<0L  ' 'E  MM&//u = TV WU^  '<6??+J v 1 $((6??1-RYY?C$33CI99S,/77> >++fooRYY?C$33CI;;u% %  ' v A %99))!,E:Lgen ;;v//u=EEeL L  ' v  !Q &99**1-U;\gen ;;v00>FFuM M             v~~ ! #99*%8,gen ;;v~~U;CCEJ J  % v A %99))!,E:Lgen ;;v//u=EEeL L  & v  !Q &99**1-U;\gen 88D!2!235AII%P P )) ) &%% & v"" #q (99F//2F4G4G4JK  gen  88.1"bk:1qt$:%gen  ** * &%% & v"" #q (99F//2F4G4G4JK  gen  88.1"bk:1qt$:%gen   $ v 1 $99+59<gen ;;ve<DDUK K5 DEEQ 6`;;s` 4` `)NNF)E__doc__numpyr tensorboard.compat.protor"tensorboard.compat.tensorflow_stubrrrrrr"r$r'r,r1r4r6r9r<rCrHrLrOr rrrrrruint32uint64rrrrobject_bool_rr!rrrrr[r rRrVr\rbre frozensetrrjryr|rrrrrrrrrrrrrvrrrs,/KKB  DEBDEEE  OCJJ3JJ3JJ3HH-  HH/  HH- II.II1II1GG ,HH-LL7MM9JJ2HH.  LL!A!" MM  "B#$ MM  "B NN!!#C MM  "B)2 OO""$H :L"           ( 7E H 8  KK+ ' ~& NN\N   NN\N   NN\N  LL:, LL:, LL:, KK* MMJ - MMJ - LL:|, NNZ. MMJ -  MMJ