lmoneta · lmoneta · Aug 25, 2023 · Aug 25, 2023 · Aug 25, 2023
@@ -0,0 +1,12 @@
+tensor_batchnormalization1movingvariance0 64
+0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394 0.999500394
+tensor_batchnormalization1beta0 64
+0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
+tensor_batchnormalization1gamma0 64
+1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
+tensor_dense17kernel0 448
+0.0108048618 0.215931386 0.0123945177 0.285069078 -0.104697779 -0.209598348 0.00423592329 -0.2823928 0.116487086 -0.12513034 0.0747551024 -0.00770813227 0.239529043 0.159453094 -0.28320843 -0.00874188542 -0.157378957 0.0602476597 -0.205106258 0.228819519 0.156571299 -0.234788939 -0.0700670928 -0.263461113 0.143216401 0.0773885548 -0.00458329916 -0.170526713 -0.145017579 0.122938514 -0.290278703 0.0688058138 0.0361943543 -0.0493358076 -0.00392708182 -0.284400582 0.166610718 0.122950584 -0.0912979394 -0.0632623136 -0.19157052 -0.0420778245 -0.0250469148 0.277549595 -0.246620253 0.153357059 -0.193938583 0.0913421512 0.17182675 -0.161206797 -0.146403342 0.168200165 0.0514654517 0.285589248 0.177236408 0.239046842 0.271209925 -0.0295130908 0.243002325 -0.177540243 -0.168115333 -0.228779823 0.0283973515 0.179595679 -0.198399067 0.215017587 0.0389697552 -0.0777019709 -0.0739576519 -0.0860316157 -0.0285128355 -0.142276645 -0.137453258 -0.157670751 0.11001423 -0.251886994 -0.290008962 0.238182813 -0.156130373 -0.248245955 -0.124903694 -0.129840136 0.0469535887 -0.156804115 0.0401303768 0.221745759 -0.0454875827 0.179246753 -0.0117903352 0.0204212666 -0.268830091 0.218235701 -0.264822543 0.151397437 -0.220389888 0.217418164 0.253188759 0.0389984548 0.202692091 0.0823853612 -0.223403066 0.273629099 0.230908662 -0.107497707 -0.110331237 -0.120720729 -0.235759467 0.0507353544 -0.122409225 0.112901539 -0.189501375 -0.118903041 -0.269068241 0.0767331123 -0.209109575 -0.0241218507 0.249974757 0.228964657 0.251895219 -0.0699959695 0.245350212 -0.271214575 -0.226293862 0.245217532 0.0605310798 -0.105240524 -0.181201681 0.276239663 -0.154074326 0.0106938779 0.230460852 0.115790337 0.161239386 0.154280871 0.215945214 0.287186056 -0.226439416 -0.244549721 0.25764963 0.132598132 -0.167660266 -0.124990329 0.0322424471 -0.144228578 -0.0685434043 -0.263268828 0.287399143 -0.167323902 -0.175259233 -0.220007718 0.247749418 0.249331802 0.0275261402 0.285681635 0.224241287 0.279460758 0.0549612343 0.136446476 -0.1061299 -0.211200207 0.190035731 -0.148243546 -0.000311821699 0.231444687 -0.0228230059 -0.199860305 -0.269175649 0.0455857217 -0.259420276 -0.131201625 -0.117612928 -0.189057663 0.185323805 0.0275135934 0.0949172378 -0.0632302165 0.187787294 0.021389991 -0.122828677 -0.231739581 -0.0210861564 0.130167186 0.230645269 -0.204621851 -0.0380035043 0.0184370279 0.0357265174 0.274054497 0.154319197 -0.280433387 0.276268095 0.192360759 0.181914926 -0.10357748 -0.0122078657 0.0994132161 -0.189529791 0.14017576 0.194115937 -0.0822871476 0.190085024 0.28514728 0.196474433 -0.126136974 0.0876276493 0.22874549 -0.0943528563 -0.21709615 -0.166011333 -0.204499945 -0.228483602 -0.029565841 -0.221954465 0.169807643 0.116328508 0.10293144 0.238334745 0.0139791071 -0.279547811 -0.057903856 0.116980433 0.211231321 -0.0950945914 0.0762212574 -0.122185215 -0.2635867 -0.189521894 -0.133030325 0.276025444 -0.274003267 0.0397931337 -0.0262677968 -0.0421609879 -0.205390483 0.265192062 0.169212639 -0.0588397384 -0.113253638 -0.286851257 -0.173200786 0.0343206823 -0.128531307 0.120894819 0.16205284 -0.26693207 -0.103117675 -0.0463741124 0.234895021 0.0869942904 -0.195388302 -0.243550718 0.111745328 0.259278566 0.232909232 0.269730121 0.128117055 -0.285545975 -0.015607506 0.0385112166 0.223638445 0.236599356 0.191709965 -0.19348073 0.067645669 -0.14132981 0.241881877 -0.254734188 -0.228491291 -0.239838064 -0.0670351833 0.154426962 -0.0801543146 0.132534206 -0.157252058 0.139873505 0.103443086 -0.184833378 0.231345206 0.272313625 0.144612879 -0.0427127481 -0.116279498 0.266867608 0.148977041 -0.11838904 -0.243089601 -0.220981985 0.273647815 0.265305191 -0.108035192 -0.00552237034 0.0897967219 0.0963402092 0.257930547 0.00589051843 0.24955228 -0.287052602 -0.216473341 -0.118245646 -0.229325905 0.00889033079 -0.141335711 0.16459924 -0.279934347 0.264600664 0.189448148 0.20955357 0.0990448296 -0.027635932 0.0795616508 0.243308693 0.0665873885 -0.193698078 0.178786546 0.0275886059 0.19980213 0.0518908501 -0.227167904 -0.0288101733 -0.217772186 0.0663120151 0.0809931457 0.23769322 -0.284707844 -0.277737409 0.202413201 -0.0784830749 0.200194508 -0.191186607 -0.163768932 -0.0959100127 -0.24673669 -0.11973758 -0.18587245 0.237517565 0.266499251 -0.0500470549 0.0682872534 -0.114201352 -0.203679815 0.0936342776 -0.0233978629 -0.0219255388 0.190061122 -0.156872734 0.167493641 -0.229817986 -0.132643729 -0.282223672 0.0858635902 -0.12809591 0.14055869 -0.203111768 -0.153313249 -0.0905048549 0.032363534 0.0640204847 -0.11933282 -0.0381491482 0.288048357 0.0544538796 0.0737758577 -0.186934054 -0.0764555931 -0.118738636 0.0113504529 -0.23289682 -0.24103862 0.183179826 -0.0324722826 0.0281182528 -0.106172249 0.137558669 -0.101038292 -0.0941964239 -0.109811425 0.205613256 -0.0358462334 -0.202844858 0.259382457 0.00479933619 0.279637247 -0.250478655 0.201448083 0.0918923318 0.0203225613 0.133025885 -0.160158649 -0.139158651 0.103551477 -0.0765164495 0.235947281 0.0446474254 0.0191914439 -0.115890756 -0.239997402 0.0594975352 -0.206928521 0.283032924 -0.15147914 0.251403838 -0.049104169 -0.129442304 -0.115165785 0.0916467011 -0.0739351213 -0.00163686275 0.0691559017 -0.0499850214 0.13808167 -0.151729494 0.08843413 0.257064074 -0.224859029 -0.0825117826 -0.0263075829 -0.257945806 -0.198194474 0.200157225 -0.120868146 -0.17015183 -0.268784493 0.235713631 0.208676606 -0.0274443626 -0.278948426 -0.220054567 0.093952328 -0.00847533345 0.169271052 -0.109093592 0.111329764 -0.232666582 -0.130706757 0.0706871152 -0.215322942 -0.287200451 -0.0953007936 0.101649582 0.0647162497 -0.266912311 -0.0801778883 -0.205314517 -0.285931677 0.214427441 0.217154175 0.148746103 0.0555344224
+tensor_dense17bias0 64
+0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
+tensor_batchnormalization1movingmean0 64
+0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
@@ -0,0 +1,160 @@
+//Code generated automatically by TMVA for Inference of Model file [BatchNormalizationtest.h5] at [Thu Aug 24 08:55:44 202] 
+
+#ifndef ROOT_TMVA_SOFIE_BATCHNORMALIZATIONTEST
+#define ROOT_TMVA_SOFIE_BATCHNORMALIZATIONTEST
+
+#include<algorithm>
+#include<vector>
+#include "TMVA/SOFIE_common.hxx"
+#include <fstream>
+
+namespace TMVA_SOFIE_BatchNormalizationtest{
+namespace BLAS{
+	extern "C" void saxpy_(const int * n, const float * alpha, const float * x,
+	                         const int * incx, float * y, const int * incy);
+	extern "C" void scopy_(const int *n, const float* x, const int *incx, float* y, const int* incy);
+	extern "C" void sgemm_(const char * transa, const char * transb, const int * m, const int * n, const int * k,
+	                       const float * alpha, const float * A, const int * lda, const float * B, const int * ldb,
+	                       const float * beta, float * C, const int * ldc);
+	extern "C" void sgemv_(const char * trans, const int * m, const int * n, const float * alpha, const float * A,
+	                       const int * lda, const float * X, const int * incx, const float * beta, const float * Y, const int * incy);
+}//BLAS
+struct Session {
+std::vector<float> fTensor_batchnormalization1movingvariance0 = std::vector<float>(64);
+float * tensor_batchnormalization1movingvariance0 = fTensor_batchnormalization1movingvariance0.data();
+std::vector<float> fTensor_batchnormalization1beta0 = std::vector<float>(64);
+float * tensor_batchnormalization1beta0 = fTensor_batchnormalization1beta0.data();
+std::vector<float> fTensor_batchnormalization1gamma0 = std::vector<float>(64);
+float * tensor_batchnormalization1gamma0 = fTensor_batchnormalization1gamma0.data();
+std::vector<float> fTensor_dense17kernel0 = std::vector<float>(448);
+float * tensor_dense17kernel0 = fTensor_dense17kernel0.data();
+std::vector<float> fTensor_dense17bias0 = std::vector<float>(64);
+float * tensor_dense17bias0 = fTensor_dense17bias0.data();
+std::vector<float> fTensor_batchnormalization1movingmean0 = std::vector<float>(64);
+float * tensor_batchnormalization1movingmean0 = fTensor_batchnormalization1movingmean0.data();
+std::vector<float> fTensor_batchnormalization1batchnormadd10 = std::vector<float>(64);
+float * tensor_batchnormalization1batchnormadd10 = fTensor_batchnormalization1batchnormadd10.data();
+std::vector<float> fTensor_dense17BiasAdd0 = std::vector<float>(64);
+float * tensor_dense17BiasAdd0 = fTensor_dense17BiasAdd0.data();
+std::vector<float> fTensor_dense17bias0bcast = std::vector<float>(64);
+float * tensor_dense17bias0bcast = fTensor_dense17bias0bcast.data();
+
+
+Session(std::string filename ="") {
+   if (filename.empty()) filename = "BatchNormalizationtest.dat";
+   std::ifstream f;
+   f.open(filename);
+   if (!f.is_open()){
+      throw std::runtime_error("tmva-sofie failed to open file for input weights");
+   }
+   std::string tensor_name;
+   int length;
+   f >> tensor_name >> length;
+   if (tensor_name != "tensor_batchnormalization1movingvariance0" ) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor name; expected name is tensor_batchnormalization1movingvariance0 , read " + tensor_name;
+      throw std::runtime_error(err_msg);
+    }
+   if (length != 64) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor size; expected size is 64 , read " + std::to_string(length) ;
+      throw std::runtime_error(err_msg);
+    }
+    for (int i =0; i < length; ++i) 
+       f >> tensor_batchnormalization1movingvariance0[i];
+   f >> tensor_name >> length;
+   if (tensor_name != "tensor_batchnormalization1beta0" ) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor name; expected name is tensor_batchnormalization1beta0 , read " + tensor_name;
+      throw std::runtime_error(err_msg);
+    }
+   if (length != 64) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor size; expected size is 64 , read " + std::to_string(length) ;
+      throw std::runtime_error(err_msg);
+    }
+    for (int i =0; i < length; ++i) 
+       f >> tensor_batchnormalization1beta0[i];
+   f >> tensor_name >> length;
+   if (tensor_name != "tensor_batchnormalization1gamma0" ) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor name; expected name is tensor_batchnormalization1gamma0 , read " + tensor_name;
+      throw std::runtime_error(err_msg);
+    }
+   if (length != 64) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor size; expected size is 64 , read " + std::to_string(length) ;
+      throw std::runtime_error(err_msg);
+    }
+    for (int i =0; i < length; ++i) 
+       f >> tensor_batchnormalization1gamma0[i];
+   f >> tensor_name >> length;
+   if (tensor_name != "tensor_dense17kernel0" ) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor name; expected name is tensor_dense17kernel0 , read " + tensor_name;
+      throw std::runtime_error(err_msg);
+    }
+   if (length != 448) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor size; expected size is 448 , read " + std::to_string(length) ;
+      throw std::runtime_error(err_msg);
+    }
+    for (int i =0; i < length; ++i) 
+       f >> tensor_dense17kernel0[i];
+   f >> tensor_name >> length;
+   if (tensor_name != "tensor_dense17bias0" ) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor name; expected name is tensor_dense17bias0 , read " + tensor_name;
+      throw std::runtime_error(err_msg);
+    }
+   if (length != 64) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor size; expected size is 64 , read " + std::to_string(length) ;
+      throw std::runtime_error(err_msg);
+    }
+    for (int i =0; i < length; ++i) 
+       f >> tensor_dense17bias0[i];
+   f >> tensor_name >> length;
+   if (tensor_name != "tensor_batchnormalization1movingmean0" ) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor name; expected name is tensor_batchnormalization1movingmean0 , read " + tensor_name;
+      throw std::runtime_error(err_msg);
+    }
+   if (length != 64) {
+      std::string err_msg = "TMVA-SOFIE failed to read the correct tensor size; expected size is 64 , read " + std::to_string(length) ;
+      throw std::runtime_error(err_msg);
+    }
+    for (int i =0; i < length; ++i) 
+       f >> tensor_batchnormalization1movingmean0[i];
+   f.close();
+   {
+      float * data = TMVA::Experimental::SOFIE::UTILITY::UnidirectionalBroadcast<float>(tensor_dense17bias0,{ 64 }, { 1 , 64 });
+      std::copy(data, data + 64, tensor_dense17bias0bcast);
+      delete [] data;
+   }
+}
+
+std::vector<float> infer(float* tensor_dense17input){
+
+//--------- Gemm
+   char op_0_transA = 'n';
+   char op_0_transB = 'n';
+   int op_0_m = 1;
+   int op_0_n = 64;
+   int op_0_k = 7;
+   float op_0_alpha = 1;
+   float op_0_beta = 1;
+   int op_0_lda = 7;
+   int op_0_ldb = 64;
+   std::copy(tensor_dense17bias0bcast, tensor_dense17bias0bcast + 64, tensor_dense17BiasAdd0);
+   BLAS::sgemm_(&op_0_transB, &op_0_transA, &op_0_n, &op_0_m, &op_0_k, &op_0_alpha, tensor_dense17kernel0, &op_0_ldb, tensor_dense17input, &op_0_lda, &op_0_beta, tensor_dense17BiasAdd0, &op_0_n);
+   constexpr int op_1_N =64;
+   constexpr int op_1_incx = 1;
+   constexpr int op_1_incy = 1;
+   BLAS::scopy_(&op_1_N, tensor_dense17BiasAdd0, &op_1_incx,tensor_batchnormalization1batchnormadd10, &op_1_incy);
+
+   float op_1_alpha = -1;
+   BLAS::saxpy_(&op_1_N, &op_1_alpha, tensor_batchnormalization1movingmean0, &op_1_incx,tensor_batchnormalization1batchnormadd10, &op_1_incy);
+
+    for (size_t i = 0; i < 64; i++) {
+      tensor_batchnormalization1batchnormadd10[i] *= tensor_batchnormalization1gamma0[i] * tensor_batchnormalization1movingvariance0[i]; 
+   }
+   op_1_alpha = 1;
+   BLAS::saxpy_(&op_1_N, &op_1_alpha, tensor_batchnormalization1beta0, &op_1_incx, tensor_batchnormalization1batchnormadd10, &op_1_incy);
+
+   std::vector<float> ret (tensor_batchnormalization1batchnormadd10, tensor_batchnormalization1batchnormadd10 + 64);
+   return ret;
+}
+};
+} //TMVA_SOFIE_BatchNormalizationtest
+
+#endif  // ROOT_TMVA_SOFIE_BATCHNORMALIZATIONTEST