+ 9iY ^RIt^RIt^RIt^RIHt^RIt^RIHtH t H t ^RI H t RRlt RRltR R ltRR R lltR RltRRltRRRlltRtRRRlltRRRlltRtRRltRt]R8Xd ]!4R#R#) N)Path) ModelProto TensorProto numpy_helper) OnnxModelc hV^8dQhR\R\R\R\R\P/#) input_ids batch_sizesequence_lengthdictionary_sizereturnrintnpndarray)formats"e/var/www/html/photoedit/myenv/lib/python3.14/site-packages/onnxruntime/transformers/bert_test_data.py __annotate__rs8(+>ATWZZc8VPPP\P\P \P 39gQh\PPW1V3\PR7pVPPP\P8Xd\P!V4pV#VPPP\P 8Xd\P!V4pV#)a@Create input tensor based on the graph input of input_ids Args: input_ids (TensorProto): graph input of the input_ids input tensor batch_size (int): batch size sequence_length (int): sequence length dictionary_size (int): vocabulary size of dictionary Returns: np.ndarray: the input tensor created )sizedtype) type tensor_type elem_typerFLOATINT32INT64rrandomrandintint32float32int64)r r r r datas&&&& rfake_input_ids_datar&s >> % % / /4   99  _3PXZX`X`  aD~~!!++{/@/@@zz$ K  # # - -1B1B Bxx~ Krc\V^8dQhR\R\R\R\P/#)r segment_idsr r rr)rs"rrr1s0{VY^`^h^hrc$VPPP\P\P \P 39gQh\P!W3\PR7pVPPP\P8Xd\P!V4pV#VPPP\P 8Xd\P!V4pV#)aCreate input tensor based on the graph input of segment_ids Args: segment_ids (TensorProto): graph input of the token_type_ids input tensor batch_size (int): batch size sequence_length (int): sequence length Returns: np.ndarray: the input tensor created r) rrrrrrrrzerosr"r#r$)r(r r r%s&&& rfake_segment_ids_datar,1s    ' ' 1 16   88Z1 BD##--1B1BBzz$ K    % % / /;3D3D Dxx~ Krc0V^8dQhR\R\/#)r max_sequence_lengthaverage_sequence_length)r)rs"rrrLsBB3BBrcV^8dW8:gQh^V,V8d&\P!^V,V, V4#\P!^^V,^, 4#))r r!)r.r/s&&rget_random_lengthr2Lsa "a ',C,ZZ Z ""%88~~a"99*QRUVQVtNQ,>'?$1u+'M\]!^$YZ]^Y^tNQ4F/G$STu+/UdeQ^a'!+,+:&1Q':>*'('B!^$/0/JQ^a'!+,+"",, 0A0AAzz$ K  $ $ . .+2C2C Cxx~ KrcfV^8dQhR\R\\\P3,/#)r directoryinputs)strdictrr)rs"rrrs'333T#rzz/-B3rc V\PPV4'g(\P!V4\ RV R24M\ RV R24\ VP44F|wpwr4\P!WC4p\\PPVRV R24R 4;_uu_4pVPVP44R R R 4K~ R # \ d\ RT R24Li;i +'giK;i) zOutput input tensors of test data to a directory Args: directory (str): path of a directory inputs (Dict[str, np.ndarray]): map from input name to value z#Successfully created the directory  zCreation of the directory z failedzWarning: directory z$ existed. Files will be overwritten.input_.pbwbN)ospathexistsmkdirprintOSError enumerateitemsr from_arrayopenjoinwriteSerializeToString)rBrCindexnamer%tensorfiles&& routput_test_datar\s 77>>) $ $ F HHY  7 {!D E #I;.RST(8|((4 "'',,yF5'*=> E E JJv//1 2F E 9 C .ykA B CF E EsC7 D7DD D( cV^8dQhR\R\R\R\R\R\R\R\R \R \R \R \/ #) r r r test_casesr verbose random_seedr r(r4r/r5r6rr7r)rs"rrrs77777 7  7  7777!7!77rc VfQh\PPV4\P!V4.p \V4Fp \ W`W4pVP V/pV'd\ WpV4WP &V'd\WWW4WP &V'd\V 4^8Xd \RV4V PV4K V #)aCreate given number of input data for testing Args: batch_size (int): batch size sequence_length (int): sequence length test_cases (int): number of test cases dictionary_size (int): vocabulary size of dictionary for input_ids verbose (bool): print more information or not random_seed (int): random seed input_ids (TensorProto): graph input of input IDs segment_ids (TensorProto): graph input of token type IDs input_mask (TensorProto): graph input of attention mask average_sequence_length (int): average sequence length excluding paddings random_sequence_length (bool): whether use uniform random number for sequence length mask_type (int): mask type 1 is mask index; 2 is 2D mask; 3 is key len, cumulated lengths of query and key Returns: List[Dict[str,numpy.ndarray]]: list of test cases, where each test case is a dictionary with input name as key and a tensor as value zExample inputs) rr seedr:r&rYr,r@lenrOappend)r r r^r r_r`r r(r4r/r5r6 all_inputs _test_caseinput_1rCs&&&&&&&&&&&& rfake_test_datarisD   IINN; KK JJ' %i_^..'* '<[Ve'fF## $ &:Rh'F?? # s:!+ "F +&!( rcV^8dQhR\R\R\R\R\R\R\R\R \R \R \R \/ #) r r r r^rcr_r r(r4r/r5r6r ra)rs"rrrs11111  1  1  111!1!111rc h\VVVV VVVVVVV V 4 p \V 4V8wd \R4V #)auCreate given number of input data for testing Args: batch_size (int): batch size sequence_length (int): sequence length test_cases (int): number of test cases seed (int): random seed verbose (bool): print more information or not input_ids (TensorProto): graph input of input IDs segment_ids (TensorProto): graph input of token type IDs input_mask (TensorProto): graph input of attention mask average_sequence_length (int): average sequence length excluding paddings random_sequence_length (bool): whether use uniform random number for sequence length mask_type (int): mask type 1 is mask index; 2 is 2D mask; 3 is key len, cumulated lengths of query and key Returns: List[Dict[str,numpy.ndarray]]: list of test cases, where each test case is a dictionary with input name as key and a tensor as value z$Failed to create test data for test.)rirdrO) r r r^rcr_r r(r4r/r5r6r rfs &&&&&&&&&&&& rgenerate_test_datarlsPB   J :*$ 45 rcV\VP48dR#VPV,pVPV4pVfIVPW4pVe4VPR8Xd#VPVP^,4pV#)NCast)rdinputfind_graph_input get_parentop_type) onnx_model embed_node input_indexro graph_input parent_nodes&&& rget_graph_input_from_embed_noderx$sc***++   [ )E--e4K ++JD  "{':':f'D$55k6G6G6JKK rcV^8dQhR\R\R,R\R,R\R,R\\PR,\PR,\PR,3,/#)r rsinput_ids_nameNsegment_ids_nameinput_mask_namer)rrDtuplerr)rs"rrr1s{YXYXYX$JYXDjYX4Z YX  2:: bjj4/d1B BC YXrcVP4pVeVPV4pVf\RV 24hRpV'd$VPV4pVf\RV 24hRpV'd$VPV4pVf\RV 24h^V'd^M^,V'd^M^,p\V4V8wd\RV R\V4 24hWVV3#\V4^8wd\R\V4 24hVP R4p \V 4^8XdsV ^,p \ W ^4p\ W ^4p\ W ^4pVf/VF(p V P P4p RV 9gK&T pK* Vf \R4hWVV3#RpRpRpVF<p V P P4p RV 9dT pK(R V 9gR V 9dT pK:T pK> V'dV'd V'dWVV3#\R 4h) aFind graph inputs for BERT model. First, we will deduce inputs from EmbedLayerNormalization node. If not found, we will guess the meaning of graph inputs based on naming. Args: onnx_model (OnnxModel): onnx model object input_ids_name (str, optional): Name of graph input for input IDs. Defaults to None. segment_ids_name (str, optional): Name of graph input for segment IDs. Defaults to None. input_mask_name (str, optional): Name of graph input for attention mask. Defaults to None. Raises: ValueError: Graph does not have input named of input_ids_name or segment_ids_name or input_mask_name ValueError: Expected graph input number does not match with specified input_ids_name, segment_ids_name and input_mask_name Returns: Tuple[Optional[np.ndarray], Optional[np.ndarray], Optional[np.ndarray]]: input tensors of input_ids, segment_ids and input_mask Nz Graph does not have input named zExpect the graph to have z inputs. Got z'Expect the graph to have 3 inputs. Got EmbedLayerNormalizationmaskz#Failed to find attention mask inputtokensegmentz?Fail to assign 3 inputs. You might try rename the graph inputs.)'get_graph_inputs_excluding_initializersrp ValueErrorrdget_nodes_by_op_typerxrYlower) rsrzr{r| graph_inputsr r(r4expected_inputs embed_nodesrtroinput_name_lowers &&&& rfind_bert_inputsr1s24EEGL!//?  ??OPQ Q $556FGK" #CDTCU!VWW #44_EJ! #COCT!UVVKqQ7 1PQR |  /88IWZ[gWhVijk kz11 <AB3|CTBUVWW112KLK ;1 ^ 3JAN 5jaP 4ZQO  %#(::#3#3#5 --!&J&  BC Cz11IKJ ::++- % %J ' '98H+HKI[Zz11 V WWrcV^8dQhR\R\R,R\R,R\R,R\\PR,\PR,\PR,3,/#)r onnx_filerzNr{r|r)rDr}rr)rs"rrrst[[[$J[Dj[4Z [  2:: bjj4/d1B BC [rc\4p\VR4;_uu_4pVPVP44RRR4\ V4p\ WaW#4# +'giL';i)aFind graph inputs for BERT model. First, we will deduce inputs from EmbedLayerNormalization node. If not found, we will guess the meaning of graph inputs based on naming. Args: onnx_file (str): onnx model path input_ids_name (str, optional): Name of graph input for input IDs. Defaults to None. segment_ids_name (str, optional): Name of graph input for segment IDs. Defaults to None. input_mask_name (str, optional): Name of graph input for attention mask. Defaults to None. Returns: Tuple[Optional[np.ndarray], Optional[np.ndarray], Optional[np.ndarray]]: input tensors of input_ids, segment_ids and input_mask rbN)rrTParseFromStringreadrr)rrzr{r|modelr[rss&&&& rget_bert_inputsrsV( LE i  $ diik* 5!J J8H ZZ  s A A. cv\P!4pVPRR\RR7VPRR\RRR7VPR R\^R R7VPR R\^R R7VPR R\RRR7VPRR\RRR7VPRR\RRR7VPRR\^RR7VPRR\^RR7VPRRRRR7VP RR7VPRRRRR7VP RR7VPRR R)\R!R"7VPR#R$RRR%R7VP RR&7VPR'R\^R(R7VP 4pV#)*z--modelTzbert onnx model path.)requiredrhelpz --output_dirFNz4output test data path. Default is current directory.)rrdefaultrz --batch_sizezbatch size of inputz--sequence_lengthz maximum sequence length of inputz--input_ids_namezinput name for input idsz--segment_ids_namezinput name for segment idsz--input_mask_namezinput name for attention maskz --samplesz$number of test cases to be generatedz--seedz random seedz --verbose store_truezprint verbose information)ractionr)r_z--only_input_tensorsz-only save input tensors and no output tensors)only_input_tensorsz-az--average_sequence_lengthz)average sequence length excluding padding)rrrz-rz--random_sequence_lengthz3use uniform random instead of fixed sequence length)r5z --mask_typez^mask type: (1: mask index, 2: raw 2D mask, 3: key lengths, cumulated lengths of query and key))argparseArgumentParser add_argumentrDr set_defaults parse_args)parserargss rparse_argumentsrs'  $ $ &F  DsAXY   C  S!Rgh   /    '    )    ,    3  5sAMZ  (   &  <   51  #  8   " B  u5   m     D KrcV^8dQhR\R\R\R\R\R\R\R\R ,R \R ,R \R ,R \R \R\R\/#)r r output_dirr r r^rcr_rzNr{r|rr/r5r6)rDrr7)rs"rrr sI>I> I>I>I> I>  I>  I>I>$JI>DjI>4ZI>I>!I>!I>I>rc \WW4wrp\VVVVVVVVV V V 4 p\V4FBwpp\PP VR\ V4,4p\VV4KD V 'dR#^RIpRVP49dRR.MR.pVPVVR7pVP4Uu.uFpVPNK pp\V4Fwpp\PP VR\ V4,4pVPVV4p\V4Fwpp\P!\ P"!VV,4V4p\%\PP VRV R24R4;_uu_4pVP'VP)44RRR4K K R#uupi +'giK;i) a Create test data for a model, and save test data to a directory. Args: model (str): path of ONNX bert model output_dir (str): output directory batch_size (int): batch size sequence_length (int): sequence length test_cases (int): number of test cases seed (int): random seed verbose (bool): whether print more information input_ids_name (str): graph input name of input_ids segment_ids_name (str): graph input name of segment_ids input_mask_name (str): graph input name of input_mask only_input_tensors (bool): only save input tensors, average_sequence_length (int): average sequence length excluding paddings random_sequence_length (bool): whether use uniform random number for sequence length mask_type(int): mask type test_data_set_NCUDAExecutionProviderCPUExecutionProvider) providersoutput_rIrJ)rrlrQrKrLrUrDr\ onnxruntimeget_available_providersInferenceSession get_outputsrYrunrrSrasarrayrTrVrW)rrr r r^rcr_rzr{r|rr/r5r6r r(r4rfr<rCrBrrsessionoutput output_namesresult output_name tensor_resultr[s&&&&&&&&&&&&&& rcreate_and_save_test_datar sD*9P`)r&IJ#  Jz* 6GGLL-=A-FG F++ #k&I&I&K K !"89$ % **5I*FG.5.A.A.CD.CFFKK.CLDz* 6GGLL-=A-FG \62' 5NA{(33BJJvay4I;WMbggll9s#.>?FF$ =::<=GF6+EGFFsG( G  G0 c\4pVP^8:dVPVnVPpVfZ\ VP 4p\ PPVPRVP RVP 24pVe \ V4pVPRRR7M \R4\VP VVPVPVPVPVP VP"VP$VP&VP(VPVP*VP,4\RV4R#)rNbatch__seq_T)parentsexist_okz7Directory existed. test data files will be overwritten.z Test data is saved to directory:)rr/r rrrrKrLrUparentr rNrOrsamplesrcr_rzr{r|rr5r6)rrprLs rmainrYs  D ##q('+';';$J  WW\\!((fT__4EU4K_K_J`,ab J 4$ / GH           $$ ## " ,j9r__main__)r )i')NNN)rrKr pathlibrnumpyronnxrrrrsrr&r,r2r@r\rirlrxrrrrr__name__rrrs 66 <6BFR3.7t1h YXx[8aHI>X$:N zFr