From a3f46ec5724d40058f62a71e28ed2bd763c75764 Mon Sep 17 00:00:00 2001 From: GNiendorf Date: Thu, 16 Jan 2025 12:37:04 -0500 Subject: [PATCH 1/3] t3_dnn initial commit --- RecoTracker/LSTCore/interface/alpaka/Common.h | 21 +- .../LSTCore/src/alpaka/NeuralNetwork.h | 275 +++-- .../src/alpaka/T3NeuralNetworkWeights.h | 106 ++ RecoTracker/LSTCore/src/alpaka/Triplet.h | 12 + .../analysis/DNN/train_T3_DNN.ipynb | 964 ++++++++++++++++++ .../standalone/code/core/AccessHelper.cc | 18 +- .../standalone/code/core/AccessHelper.h | 2 + .../standalone/code/core/write_lst_ntuple.cc | 115 +++ .../standalone/code/core/write_lst_ntuple.h | 3 + 9 files changed, 1410 insertions(+), 106 deletions(-) create mode 100644 RecoTracker/LSTCore/src/alpaka/T3NeuralNetworkWeights.h create mode 100644 RecoTracker/LSTCore/standalone/analysis/DNN/train_T3_DNN.ipynb diff --git a/RecoTracker/LSTCore/interface/alpaka/Common.h b/RecoTracker/LSTCore/interface/alpaka/Common.h index fd59555fe8588..77ab860a46188 100644 --- a/RecoTracker/LSTCore/interface/alpaka/Common.h +++ b/RecoTracker/LSTCore/interface/alpaka/Common.h @@ -69,18 +69,27 @@ namespace ALPAKA_ACCELERATOR_NAMESPACE::lst { {0.4f, 0.4f, 0.4f, 0.4f, 0.4f, 0.4f, 0.4f, 0.4f, 0.18f, 0.18f, /*10*/ 0.18f, 0.18f, 0.18f, 0.18f, 0.18f}, {0.4f, 0.4f, 0.4f, 0.4f, 0.4f, 0.4f, 0.4f, 0.4f, 0.4f, 0.18f, /*10*/ 0.18f, 0.18f, 0.18f, 0.18f, 0.18f}}; + // Common constants used by both DNNs + ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kEta_norm = 2.5f; + ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kPhi_norm = kPi; + constexpr unsigned int kPtBins = 2; + constexpr unsigned int kEtaBins = 10; + + namespace t3dnn { + ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kZ_max = 224.149505f; + ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kR_max = 98.932365f; + // No pt binning for T3 + ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kWp[kEtaBins] = + {0.024, 0.0267, 0.052, 0.0658, 0.093, 0.0968, 0.1913, 0.2443, 0.4012, 0.5449}; + } // namespace t3dnn + namespace t5dnn { ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kZ_max = 267.2349854f; ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kR_max = 110.1099396f; - ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kEta_norm = 2.5f; - ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kPhi_norm = kPi; - // pt, eta binned - constexpr unsigned int kPtBins = 2; - constexpr unsigned int kEtaBins = 10; ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kWp[kPtBins][kEtaBins] = { {0.4493, 0.4939, 0.5715, 0.6488, 0.5709, 0.5938, 0.7164, 0.7565, 0.8103, 0.8593}, {0.4488, 0.4448, 0.5067, 0.5929, 0.4836, 0.4112, 0.4968, 0.4403, 0.5597, 0.5067}}; - } // namespace t5dnn + } // namespace t5dnn } // namespace ALPAKA_ACCELERATOR_NAMESPACE::lst #endif diff --git a/RecoTracker/LSTCore/src/alpaka/NeuralNetwork.h b/RecoTracker/LSTCore/src/alpaka/NeuralNetwork.h index cc1bffa3d928b..4c81c137ee089 100644 --- a/RecoTracker/LSTCore/src/alpaka/NeuralNetwork.h +++ b/RecoTracker/LSTCore/src/alpaka/NeuralNetwork.h @@ -7,8 +7,9 @@ #include "RecoTracker/LSTCore/interface/MiniDoubletsSoA.h" #include "NeuralNetworkWeights.h" +#include "T3NeuralNetworkWeights.h" -namespace ALPAKA_ACCELERATOR_NAMESPACE::lst::t5dnn { +namespace ALPAKA_ACCELERATOR_NAMESPACE::lst { template ALPAKA_FN_ACC ALPAKA_FN_INLINE void relu_activation(float (&input)[FEATURES]) { @@ -46,106 +47,182 @@ namespace ALPAKA_ACCELERATOR_NAMESPACE::lst::t5dnn { } else if (delta < -kPi) { delta += 2 * kPi; } - return delta; } - template - ALPAKA_FN_ACC ALPAKA_FN_INLINE bool runInference(TAcc const& acc, - MiniDoubletsConst mds, - const unsigned int mdIndex1, - const unsigned int mdIndex2, - const unsigned int mdIndex3, - const unsigned int mdIndex4, - const unsigned int mdIndex5, - const float innerRadius, - const float outerRadius, - const float bridgeRadius) { - // Constants - constexpr unsigned int kinputFeatures = 23; - constexpr unsigned int khiddenFeatures = 32; - - float eta1 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex1]); // inner T3 anchor hit 1 eta (t3_0_eta) - float eta2 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex2]); // inner T3 anchor hit 2 eta (t3_2_eta) - float eta3 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex3]); // inner T3 anchor hit 3 eta (t3_4_eta) - float eta4 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex4]); // outer T3 anchor hit 4 eta (t3_2_eta) - float eta5 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex5]); // outer T3 anchor hit 5 eta (t3_4_eta) - - float phi1 = mds.anchorPhi()[mdIndex1]; // inner T3 anchor hit 1 phi (t3_0_phi) - float phi2 = mds.anchorPhi()[mdIndex2]; // inner T3 anchor hit 2 phi (t3_2_phi) - float phi3 = mds.anchorPhi()[mdIndex3]; // inner T3 anchor hit 3 phi (t3_4_phi) - float phi4 = mds.anchorPhi()[mdIndex4]; // outer T3 anchor hit 4 phi (t3_2_phi) - float phi5 = mds.anchorPhi()[mdIndex5]; // outer T3 anchor hit 5 phi (t3_4_phi) - - float z1 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex1]); // inner T3 anchor hit 1 z (t3_0_z) - float z2 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex2]); // inner T3 anchor hit 2 z (t3_2_z) - float z3 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex3]); // inner T3 anchor hit 3 z (t3_4_z) - float z4 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex4]); // outer T3 anchor hit 4 z (t3_2_z) - float z5 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex5]); // outer T3 anchor hit 5 z (t3_4_z) - - float r1 = mds.anchorRt()[mdIndex1]; // inner T3 anchor hit 1 r (t3_0_r) - float r2 = mds.anchorRt()[mdIndex2]; // inner T3 anchor hit 2 r (t3_2_r) - float r3 = mds.anchorRt()[mdIndex3]; // inner T3 anchor hit 3 r (t3_4_r) - float r4 = mds.anchorRt()[mdIndex4]; // outer T3 anchor hit 4 r (t3_2_r) - float r5 = mds.anchorRt()[mdIndex5]; // outer T3 anchor hit 5 r (t3_4_r) - - // Build the input feature vector using pairwise differences after the first hit - float x[kinputFeatures] = { - eta1 / kEta_norm, // inner T3: First hit eta normalized - alpaka::math::abs(acc, phi1) / kPhi_norm, // inner T3: First hit phi normalized - z1 / kZ_max, // inner T3: First hit z normalized - r1 / kR_max, // inner T3: First hit r normalized - - eta2 - eta1, // inner T3: Difference in eta between hit 2 and 1 - delta_phi(phi2, phi1) / kPhi_norm, // inner T3: Difference in phi between hit 2 and 1 - (z2 - z1) / kZ_max, // inner T3: Difference in z between hit 2 and 1 normalized - (r2 - r1) / kR_max, // inner T3: Difference in r between hit 2 and 1 normalized - - eta3 - eta2, // inner T3: Difference in eta between hit 3 and 2 - delta_phi(phi3, phi2) / kPhi_norm, // inner T3: Difference in phi between hit 3 and 2 - (z3 - z2) / kZ_max, // inner T3: Difference in z between hit 3 and 2 normalized - (r3 - r2) / kR_max, // inner T3: Difference in r between hit 3 and 2 normalized - - eta4 - eta3, // outer T3: Difference in eta between hit 4 and 3 - delta_phi(phi4, phi3) / kPhi_norm, // inner T3: Difference in phi between hit 4 and 3 - (z4 - z3) / kZ_max, // outer T3: Difference in z between hit 4 and 3 normalized - (r4 - r3) / kR_max, // outer T3: Difference in r between hit 4 and 3 normalized - - eta5 - eta4, // outer T3: Difference in eta between hit 5 and 4 - delta_phi(phi5, phi4) / kPhi_norm, // inner T3: Difference in phi between hit 5 and 4 - (z5 - z4) / kZ_max, // outer T3: Difference in z between hit 5 and 4 normalized - (r5 - r4) / kR_max, // outer T3: Difference in r between hit 5 and 4 normalized - - alpaka::math::log10(acc, innerRadius), // T5 inner radius (t5_innerRadius) - alpaka::math::log10(acc, bridgeRadius), // T5 bridge radius (t5_bridgeRadius) - alpaka::math::log10(acc, outerRadius) // T5 outer radius (t5_outerRadius) - }; - - float x_1[khiddenFeatures]; // Layer 1 output - float x_2[khiddenFeatures]; // Layer 2 output - float x_3[1]; // Layer 3 linear output - - // Layer 1: Linear + Relu - linear_layer(x, x_1, wgtT_layer1, bias_layer1); - relu_activation(x_1); - - // Layer 2: Linear + Relu - linear_layer(x_1, x_2, wgtT_layer2, bias_layer2); - relu_activation(x_2); - - // Layer 3: Linear + Sigmoid - linear_layer(x_2, x_3, wgtT_output_layer, bias_output_layer); - float x_5 = sigmoid_activation(acc, x_3[0]); - - // Get the bin index based on abs(eta) of first hit and t5_pt - float t5_pt = innerRadius * lst::k2Rinv1GeVf * 2; - - uint8_t pt_index = (t5_pt > 5); - uint8_t bin_index = (eta1 > 2.5f) ? (kEtaBins - 1) : static_cast(eta1 / 0.25f); - - // Compare x_5 to the cut value for the relevant bin - return x_5 > kWp[pt_index][bin_index]; - } -} // namespace ALPAKA_ACCELERATOR_NAMESPACE::lst::t5dnn + namespace t3dnn { + template + ALPAKA_FN_ACC ALPAKA_FN_INLINE bool runInference(TAcc const& acc, + MiniDoubletsConst mds, + const unsigned int mdIndex1, + const unsigned int mdIndex2, + const unsigned int mdIndex3, + const float radius, + const float betaIn) { + // Constants for T3 DNN + constexpr unsigned int kinputFeatures = 14; + constexpr unsigned int khiddenFeatures = 32; + + // Extract hit information + float eta1 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex1]); // inner T3 anchor hit 1 eta + float eta2 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex2]); // inner T3 anchor hit 2 eta + float eta3 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex3]); // inner T3 anchor hit 3 eta + + float phi1 = mds.anchorPhi()[mdIndex1]; // inner T3 anchor hit 1 phi + float phi2 = mds.anchorPhi()[mdIndex2]; // inner T3 anchor hit 2 phi + float phi3 = mds.anchorPhi()[mdIndex3]; // inner T3 anchor hit 3 phi + + float z1 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex1]); // inner T3 anchor hit 1 z + float z2 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex2]); // inner T3 anchor hit 2 z + float z3 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex3]); // inner T3 anchor hit 3 z + + float r1 = mds.anchorRt()[mdIndex1]; // inner T3 anchor hit 1 r + float r2 = mds.anchorRt()[mdIndex2]; // inner T3 anchor hit 2 r + float r3 = mds.anchorRt()[mdIndex3]; // inner T3 anchor hit 3 r + + // Build input feature vector matching training order + float x[kinputFeatures] = { + eta1 / kEta_norm, // First hit eta normalized + alpaka::math::abs(acc, phi1) / kPhi_norm, // First hit phi normalized + z1 / kZ_max, // First hit z normalized + r1 / kR_max, // First hit r normalized + + eta2 - eta1, // Difference in eta between hit 2 and 1 + delta_phi(phi2, phi1) / kPhi_norm, // Difference in phi between hit 2 and 1 + (z2 - z1) / kZ_max, // Difference in z between hit 2 and 1 normalized + (r2 - r1) / kR_max, // Difference in r between hit 2 and 1 normalized + + eta3 - eta2, // Difference in eta between hit 3 and 2 + delta_phi(phi3, phi2) / kPhi_norm, // Difference in phi between hit 3 and 2 + (z3 - z2) / kZ_max, // Difference in z between hit 3 and 2 normalized + (r3 - r2) / kR_max, // Difference in r between hit 3 and 2 normalized + + alpaka::math::log10(acc, radius), // T3's circle radius + betaIn // Beta angle of inner segment + }; + + float x_1[khiddenFeatures]; // Layer 1 output + float x_2[khiddenFeatures]; // Layer 2 output + float x_3[1]; // Layer 3 linear output + + // Layer 1: Linear + Relu + linear_layer(x, x_1, t3dnn::wgtT_layer1, t3dnn::bias_layer1); + relu_activation(x_1); + + // Layer 2: Linear + Relu + linear_layer(x_1, x_2, t3dnn::wgtT_layer2, t3dnn::bias_layer2); + relu_activation(x_2); + + // Layer 3: Linear + Sigmoid + linear_layer(x_2, x_3, t3dnn::wgtT_output_layer, t3dnn::bias_output_layer); + float x_5 = sigmoid_activation(acc, x_3[0]); + + uint8_t bin_index = (eta1 > 2.5f) ? (kEtaBins - 1) : static_cast(eta1 / 0.25f); + + // Compare to cut value for relevant bin + return x_5 > kWp[bin_index]; + } + } // namespace t3dnn + + namespace t5dnn { + template + ALPAKA_FN_ACC ALPAKA_FN_INLINE bool runInference(TAcc const& acc, + MiniDoubletsConst mds, + const unsigned int mdIndex1, + const unsigned int mdIndex2, + const unsigned int mdIndex3, + const unsigned int mdIndex4, + const unsigned int mdIndex5, + const float innerRadius, + const float outerRadius, + const float bridgeRadius) { + // Constants + constexpr unsigned int kinputFeatures = 23; + constexpr unsigned int khiddenFeatures = 32; + + float eta1 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex1]); // inner T3 anchor hit 1 eta + float eta2 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex2]); // inner T3 anchor hit 2 eta + float eta3 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex3]); // inner T3 anchor hit 3 eta + float eta4 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex4]); // outer T3 anchor hit 4 eta + float eta5 = alpaka::math::abs(acc, mds.anchorEta()[mdIndex5]); // outer T3 anchor hit 5 eta + + float phi1 = mds.anchorPhi()[mdIndex1]; // inner T3 anchor hit 1 phi + float phi2 = mds.anchorPhi()[mdIndex2]; // inner T3 anchor hit 2 phi + float phi3 = mds.anchorPhi()[mdIndex3]; // inner T3 anchor hit 3 phi + float phi4 = mds.anchorPhi()[mdIndex4]; // outer T3 anchor hit 4 phi + float phi5 = mds.anchorPhi()[mdIndex5]; // outer T3 anchor hit 5 phi + + float z1 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex1]); // inner T3 anchor hit 1 z + float z2 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex2]); // inner T3 anchor hit 2 z + float z3 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex3]); // inner T3 anchor hit 3 z + float z4 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex4]); // outer T3 anchor hit 4 z + float z5 = alpaka::math::abs(acc, mds.anchorZ()[mdIndex5]); // outer T3 anchor hit 5 z + + float r1 = mds.anchorRt()[mdIndex1]; // inner T3 anchor hit 1 r + float r2 = mds.anchorRt()[mdIndex2]; // inner T3 anchor hit 2 r + float r3 = mds.anchorRt()[mdIndex3]; // inner T3 anchor hit 3 r + float r4 = mds.anchorRt()[mdIndex4]; // outer T3 anchor hit 4 r + float r5 = mds.anchorRt()[mdIndex5]; // outer T3 anchor hit 5 r + + // Build the input feature vector using pairwise differences after the first hit + float x[kinputFeatures] = { + eta1 / kEta_norm, // inner T3: First hit eta normalized + alpaka::math::abs(acc, phi1) / kPhi_norm, // inner T3: First hit phi normalized + z1 / kZ_max, // inner T3: First hit z normalized + r1 / kR_max, // inner T3: First hit r normalized + + eta2 - eta1, // inner T3: Difference in eta between hit 2 and 1 + delta_phi(phi2, phi1) / kPhi_norm, // inner T3: Difference in phi between hit 2 and 1 + (z2 - z1) / kZ_max, // inner T3: Difference in z between hit 2 and 1 normalized + (r2 - r1) / kR_max, // inner T3: Difference in r between hit 2 and 1 normalized + + eta3 - eta2, // inner T3: Difference in eta between hit 3 and 2 + delta_phi(phi3, phi2) / kPhi_norm, // inner T3: Difference in phi between hit 3 and 2 + (z3 - z2) / kZ_max, // inner T3: Difference in z between hit 3 and 2 normalized + (r3 - r2) / kR_max, // inner T3: Difference in r between hit 3 and 2 normalized + + eta4 - eta3, // outer T3: Difference in eta between hit 4 and 3 + delta_phi(phi4, phi3) / kPhi_norm, // inner T3: Difference in phi between hit 4 and 3 + (z4 - z3) / kZ_max, // outer T3: Difference in z between hit 4 and 3 normalized + (r4 - r3) / kR_max, // outer T3: Difference in r between hit 4 and 3 normalized + + eta5 - eta4, // outer T3: Difference in eta between hit 5 and 4 + delta_phi(phi5, phi4) / kPhi_norm, // inner T3: Difference in phi between hit 5 and 4 + (z5 - z4) / kZ_max, // outer T3: Difference in z between hit 5 and 4 normalized + (r5 - r4) / kR_max, // outer T3: Difference in r between hit 5 and 4 normalized + + alpaka::math::log10(acc, innerRadius), // T5 inner radius + alpaka::math::log10(acc, bridgeRadius), // T5 bridge radius + alpaka::math::log10(acc, outerRadius) // T5 outer radius + }; + + float x_1[khiddenFeatures]; // Layer 1 output + float x_2[khiddenFeatures]; // Layer 2 output + float x_3[1]; // Layer 3 linear output + + // Layer 1: Linear + Relu + linear_layer(x, x_1, t5dnn::wgtT_layer1, t5dnn::bias_layer1); + relu_activation(x_1); + + // Layer 2: Linear + Relu + linear_layer(x_1, x_2, t5dnn::wgtT_layer2, t5dnn::bias_layer2); + relu_activation(x_2); + + // Layer 3: Linear + Sigmoid + linear_layer(x_2, x_3, t5dnn::wgtT_output_layer, t5dnn::bias_output_layer); + float x_5 = sigmoid_activation(acc, x_3[0]); + + // Get the bin index based on abs(eta) of first hit and t5_pt + float t5_pt = innerRadius * lst::k2Rinv1GeVf * 2; + + uint8_t pt_index = (t5_pt > 5); + uint8_t bin_index = (eta1 > 2.5f) ? (kEtaBins - 1) : static_cast(eta1 / 0.25f); + + // Compare x_5 to the cut value for the relevant bin + return x_5 > kWp[pt_index][bin_index]; + } + } // namespace t5dnn + +} // namespace ALPAKA_ACCELERATOR_NAMESPACE::lst #endif diff --git a/RecoTracker/LSTCore/src/alpaka/T3NeuralNetworkWeights.h b/RecoTracker/LSTCore/src/alpaka/T3NeuralNetworkWeights.h new file mode 100644 index 0000000000000..59a9852fd99c7 --- /dev/null +++ b/RecoTracker/LSTCore/src/alpaka/T3NeuralNetworkWeights.h @@ -0,0 +1,106 @@ +#ifndef RecoTracker_LSTCore_src_alpaka_T3NeuralNetworkWeights_h +#define RecoTracker_LSTCore_src_alpaka_T3NeuralNetworkWeights_h + +#include + +namespace ALPAKA_ACCELERATOR_NAMESPACE::lst::t3dnn { + +ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_layer1[32] = { +-0.1918962f, -1.0575372f, -0.8276564f, -0.0243965f, -0.1577621f, 1.0067693f, 1.5348158f, 0.4439710f, 0.0041234f, -1.1558943f, -1.4180470f, 1.0221841f, -0.0592227f, -1.2107433f, -0.2100758f, 1.2193928f, -0.3124787f, -1.9197327f, -0.8064887f, -0.2178766f, -0.0111392f, -0.1638742f, 0.0029338f, -0.0157688f, 0.2662797f, 1.8194629f, 0.8465537f, -0.7592145f, -0.8783396f, 0.5602613f, -0.0764334f, -0.8502049f }; + +ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_layer1[14][32] = { +{ -0.2258795f, 0.4783621f, 0.4048500f, -0.2535419f, -0.2204617f, 0.2292106f, 0.5185162f, 1.0356801f, -0.1940614f, 0.7601448f, -0.2762348f, 0.4394473f, -0.2472922f, 0.6392686f, 0.0174135f, 1.1337029f, 0.5831500f, -0.5536784f, 0.1608927f, 0.3617360f, -0.0014275f, 0.1839313f, -0.1858799f, 0.0080405f, 0.1321980f, 0.4460649f, 0.4296648f, 0.4569599f, -0.2347775f, -0.3548780f, 0.1986685f, 0.2211701f }, +{ 0.1509047f, 0.0057684f, -0.0019561f, -0.1210250f, -0.2612511f, -0.0098326f, 0.0019430f, 0.0152595f, -0.2313585f, 0.0117159f, -0.0087940f, -0.0000195f, 0.2194081f, -0.0033250f, 0.1478879f, 0.0097880f, -0.0041943f, 0.0362815f, -0.0197458f, -0.0063192f, -0.0004957f, -0.0104775f, -0.2756116f, -0.0049759f, -0.0015340f, -0.0111502f, 0.0034782f, -0.0100520f, 0.0124440f, 0.0076567f, -0.0263710f, 0.0381163f }, +{ 0.1448244f, 0.0803795f, 1.6524559f, 0.1457684f, -0.1350329f, -3.6343203f, -1.6798589f, -0.0142065f, 0.0835910f, 0.0497492f, 0.6378320f, -0.4540107f, -0.2349942f, -0.0257038f, -0.2397820f, -4.3412380f, 0.1279643f, 0.9793525f, -0.5925660f, -0.8705363f, 0.0134403f, -0.1410540f, -0.0032170f, 0.0107955f, -1.5223076f, -0.1977515f, -0.9901226f, -0.5315630f, 0.5766137f, 0.8363163f, -0.2219186f, -1.3835622f }, +{ -0.1057612f, 0.8005342f, -1.4052893f, -0.1196175f, 0.1360446f, 1.2852736f, -4.4616752f, 0.3914140f, -0.2356829f, 1.0064709f, 1.5389245f, -3.2000272f, 0.0828165f, 0.8944567f, -0.1592458f, 0.5353487f, 0.8051971f, 0.2057788f, 1.7512299f, 0.9215795f, -0.0036782f, 2.1131773f, -0.3204709f, -0.0044941f, 0.0865040f, 0.4481153f, -1.2256490f, 1.7857696f, -0.6445597f, -0.7477201f, 0.2316555f, -0.7155212f }, +{ 0.0901890f, -1.1808448f, 1.4875709f, -0.0262624f, -0.1323451f, -0.8950851f, 0.7974008f, -2.7021067f, 0.0281707f, -1.3674084f, -4.0485940f, 5.8682752f, -0.1891649f, 7.2682476f, 0.1300428f, -2.2845411f, -8.7903214f, 0.6268425f, -1.9901284f, -2.5285044f, 0.0066100f, 7.4012928f, -0.0831600f, 0.0095476f, 6.1734300f, 0.8520991f, -8.3464308f, 2.4261341f, 3.6403832f, 5.9824433f, -0.0999645f, -0.2824311f }, +{ -0.0335586f, 16.4833145f, 1.4688981f, -0.1789742f, 0.2243387f, 0.1425887f, 0.1262731f, -4.6265879f, -0.0683203f, -12.3650198f, 1.3732860f, -0.2777069f, -0.0603657f, 0.8076705f, 0.0482012f, -0.7798629f, -1.6209395f, -0.0720773f, -0.3990867f, -1.0641636f, -0.0139909f, -0.5711431f, -0.0209516f, 0.0946307f, 0.0096234f, 0.7337275f, 0.4466013f, 1.5696099f, -5.6807351f, -1.6136186f, -0.2170101f, -0.9905877f }, +{ -0.0965901f, 0.4846557f, -0.0652100f, -0.2353250f, -0.1215300f, -3.5360579f, -2.0285676f, -2.8374794f, -0.1559567f, 0.7083026f, -0.1030692f, -1.1805813f, 0.0706595f, -0.2951699f, -0.0989490f, -2.8418458f, 0.8427116f, -2.2046885f, -1.3581268f, 1.5271128f, -0.0045855f, 0.1438956f, -0.2343508f, -0.0425132f, 1.0763166f, 1.0795754f, 0.4891663f, 2.6029303f, -1.0363307f, 2.2133234f, -0.2510982f, 1.1560025f }, +{ 0.0074010f, -0.4152716f, 3.4255989f, -0.1664258f, 0.2028382f, 1.2840286f, 1.3538148f, -1.5683322f, 0.0429628f, -0.2171310f, 2.8748064f, -0.5141454f, 0.1099870f, -0.4614673f, -0.2515875f, -0.3862101f, 1.6164609f, -1.3735241f, -0.8896182f, 1.8310896f, 0.0188789f, -0.8612124f, -0.2143756f, 0.0385537f, -8.3583250f, 1.3280702f, -7.0806746f, -0.2001765f, 0.8213225f, -2.8566475f, 0.1479015f, 0.5155560f }, +{ 0.0511283f, -1.9676421f, 4.3773866f, -0.1648336f, 0.0505374f, -3.0326672f, 2.5867159f, 1.8804691f, -0.1412510f, -2.3595653f, 3.9877729f, -14.6059513f, -0.0444889f, 7.5017557f, -0.0684788f, 1.0775388f, -8.0584450f, -0.2013538f, -12.5508022f, 9.2100115f, -0.0011293f, -7.0289955f, 0.0521985f, 0.0155513f, -1.6320782f, -4.0694451f, 7.8784938f, 8.9188061f, 10.7590771f, -9.3559170f, -0.1319706f, -0.0336969f }, +{ -0.1288323f, 7.9389300f, -0.4022054f, 0.0851088f, -0.2593997f, -0.3426172f, 0.3062155f, 1.9866818f, -0.1748946f, -9.7667055f, -0.6524503f, 0.3385161f, 0.1884300f, 9.0051126f, -0.2027990f, 0.3781536f, -0.1085841f, -5.2012277f, -0.5169301f, 2.4512987f, -14.8858967f, 0.8265918f, 0.1330098f, 14.9286051f, 0.4946520f, 1.2087970f, -1.8545671f, 0.7865057f, -3.1178455f, 2.2735860f, 0.1769712f, 0.7800660f }, +{ -0.0041451f, -0.3216657f, 0.2317746f, -0.0770410f, -0.1181624f, -1.6830167f, 1.3180428f, 1.7278676f, -0.1011897f, -0.8716651f, 0.8045275f, 1.0895320f, 0.0222730f, -1.0280910f, 0.1943782f, -3.5416641f, -1.8209232f, -0.6835734f, 1.1706284f, -1.0224590f, -0.0421350f, 0.6345906f, -0.2030822f, -0.0540403f, 1.6704807f, 0.2792225f, 0.5046398f, 1.1982751f, 0.1087174f, -0.6976060f, 0.1717374f, 1.3827170f }, +{ 0.0014786f, -1.5822506f, -1.4512368f, 0.1379846f, -0.0946356f, 1.0877693f, -1.4706703f, 0.5806997f, 0.1940563f, -1.1956979f, 0.9592837f, -0.8354632f, 0.2176267f, -0.7358339f, -0.0474411f, 0.2848974f, -0.6754610f, -0.9115687f, 2.1276686f, -1.0713644f, -0.0001056f, 1.8921490f, -0.0347501f, 0.0047789f, 1.3341833f, 1.8055003f, 1.1767987f, -3.9586561f, -0.6598469f, -0.3037908f, -0.1673333f, 0.2427263f }, +{ -0.2013803f, 0.1559301f, 0.2713688f, -0.0187786f, 0.0075337f, -0.0453327f, 0.0510727f, -0.2533994f, -0.0093928f, 0.0939973f, 0.0683563f, 0.0257523f, -0.1638049f, 0.2167263f, 0.0139614f, -0.0689526f, -0.2007709f, 0.8788205f, 0.1043992f, 0.0529033f, 0.0041733f, -0.2248188f, 0.0029659f, 0.0044919f, 0.1728916f, -1.2823037f, 0.0284686f, -0.0879781f, 0.6332331f, 0.0599467f, -0.2467749f, 0.7796255f }, +{ 0.1541705f, -2.8967228f, 0.2088300f, 0.0289306f, -0.1897649f, -0.1835614f, 0.1872510f, -0.5846522f, -0.0145777f, 2.2226386f, 0.0885817f, 0.0293056f, -0.0056043f, 2.9454181f, -0.0623621f, 0.0230481f, -0.8140234f, -4.3990140f, -0.2562745f, 0.6827632f, -4.7245188f, 0.1150251f, 0.0615204f, 4.7553473f, 0.1170709f, 0.0822542f, -0.6365855f, 0.3014538f, -2.6740234f, 0.1919117f, -0.0003937f, -0.0227543f }, +}; + +ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_layer2[32] = { +0.5612384f, -0.0400684f, -0.1890610f, -0.8038151f, -0.0184527f, -0.8351601f, -1.7656847f, -0.4417709f, -0.3462953f, 1.0924704f, 0.1019267f, 0.0497286f, 0.3448936f, -0.0442495f, -0.1294845f, -0.1740453f, -0.6256254f, 1.0588725f, 0.1306455f, 0.0451779f, 1.2896419f, -0.0145429f, 0.2775581f, -0.6205941f, 0.0369313f, -0.0632537f, -0.1257888f, -0.2130138f, 0.0593540f, 0.8294140f, -0.5174863f, -0.0208223f }; + +ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_layer2[32][32] = { +{ 0.0802436f, -0.1068745f, -0.1320603f, 0.0222937f, 0.0260173f, -0.0999878f, 0.0045460f, 0.0257662f, -0.1026595f, 0.2246336f, -0.0252555f, 0.0313543f, -0.2062458f, 0.1684278f, -0.0972477f, -0.2176001f, -0.1525148f, -0.0692302f, 0.1213232f, -0.1288088f, -0.1729289f, -0.0895176f, -0.0199082f, -0.0172326f, 0.0126052f, 0.0598741f, 0.0853796f, -0.1326686f, 0.0753646f, -0.0108036f, -0.1034372f, -0.0026847f }, +{ -0.5740064f, -0.1143318f, 0.1388070f, 1.3980483f, 0.0779263f, -1.2151324f, 2.2568495f, -0.5645000f, 2.2761426f, -0.2111119f, -0.3852313f, -1.2643801f, -0.8515005f, -0.1352748f, -0.1276971f, -0.9761337f, -1.1221037f, -0.1373484f, -6.1303482f, -0.6726473f, -3.5227783f, 0.3276748f, -0.8825266f, -0.0658490f, -0.1995521f, 0.4330961f, -0.0200665f, 0.1250316f, -0.1700243f, 1.1084150f, 1.6254629f, 0.3429218f }, +{ 1.1285430f, -0.2066509f, -0.0835447f, 1.6354681f, -0.1452742f, -0.4063630f, 0.9329628f, -0.3926201f, 0.1437214f, 0.2107810f, 1.6312121f, 1.3499900f, 0.3157536f, -0.0890818f, 0.1018713f, 0.1947485f, 0.2054133f, -0.2302379f, -0.1797780f, -0.5907550f, 0.9738173f, 0.6502790f, 0.2307118f, -0.2878186f, 0.7077893f, -0.2159140f, -0.0506834f, -0.1203295f, 0.0040051f, 0.9362236f, -0.8645781f, -5.5910249f }, +{ 0.1194026f, -0.0770950f, 0.0977710f, -0.0434290f, -0.0754589f, -0.0162118f, 0.1764810f, -0.0389150f, 0.1742229f, -0.0958543f, -0.1752356f, -0.0456456f, 0.1090995f, -0.0889283f, 0.0165932f, 0.0426875f, 0.0146571f, 0.0522842f, -0.0530897f, 0.0810161f, 0.0913054f, 0.0682782f, 0.0375430f, 0.0441132f, 0.0616512f, 0.1282314f, -0.0735286f, 0.0566166f, 0.1289222f, 0.0390186f, 0.0632017f, 0.0279914f }, +{ -0.1397050f, 0.1469029f, -0.1035157f, -0.1537557f, 0.0723362f, 0.1360317f, 0.1632015f, -0.0519030f, 0.1664423f, -0.1674401f, 0.1221813f, 0.1218546f, 0.0581016f, -0.0302506f, 0.1745771f, 0.0057055f, -0.0728391f, -0.1049056f, -0.1592961f, 0.0329875f, 0.0789358f, -0.0462632f, 0.0748179f, 0.0180050f, 0.0803397f, -0.1458464f, -0.1164891f, -0.0082622f, 0.0343072f, 0.0366396f, -0.1715891f, -0.1089048f }, +{ 0.5968121f, -0.1619606f, -0.1518290f, 0.7109355f, -0.2739125f, -1.2199029f, -0.0848204f, -0.9633313f, 1.5520710f, 0.9341283f, 0.2848818f, 1.7079298f, -0.0926562f, 0.0886664f, 0.0828461f, 2.1451118f, 1.2788274f, -0.1305077f, -0.1538649f, 0.6987577f, -1.6689410f, -0.0606114f, 0.8084964f, -0.4838576f, 0.0563342f, 0.7853296f, -0.0552194f, -0.1284064f, -0.1123770f, 1.0149344f, 0.1754804f, 0.2576492f }, +{ -2.3144670f, -0.0794675f, -0.1847414f, -0.1791797f, -0.0605982f, -3.8482890f, 0.9826989f, -1.9203123f, -0.8249935f, -0.6718075f, -1.6555610f, 2.0798690f, 1.5665153f, -0.2050105f, -0.1993937f, 1.1070000f, 0.0836252f, 1.6307852f, 0.2448520f, -2.1885965f, -0.2890749f, 2.9986718f, 2.8526762f, -0.7994567f, -0.3042918f, 0.0960785f, 0.0377693f, -0.0463109f, -0.0870433f, -0.8293707f, -1.3597184f, 1.3583397f }, +{ -0.3422491f, -0.0783274f, -0.1543446f, 0.0939283f, -0.3309467f, 0.6933140f, 0.6562154f, -0.6217518f, 0.7983661f, -0.1371530f, 0.5118276f, 0.7320337f, 0.5217202f, -0.0545936f, -0.1052059f, -0.1444394f, 0.3567903f, 1.2509772f, -0.6311634f, -0.7454629f, -1.0077031f, 0.8453025f, -0.2257671f, -0.2347097f, -0.3497045f, -0.1478627f, -0.0333489f, -0.2663794f, -0.0317280f, -0.3712497f, 0.8136677f, -2.5842018f }, +{ 0.0997253f, 0.0620212f, 0.0460688f, 0.1300038f, 0.1323504f, -0.1669361f, 0.0732264f, -0.0860083f, -0.0762404f, -0.1628099f, -0.0881021f, 0.1323460f, -0.0460503f, 0.1401906f, 0.0602387f, 0.1474468f, 0.0935555f, -0.1531849f, 0.1754870f, 0.1422898f, 0.1418013f, -0.0302381f, -0.1058683f, -0.0258421f, -0.0110884f, 0.1016085f, 0.0643939f, -0.0729882f, -0.0572575f, -0.0603358f, 0.0267279f, 0.0611828f }, +{ -0.6703482f, -0.0389322f, -0.1896956f, 1.0978996f, -0.1892595f, -2.0757272f, 1.7673731f, -0.7337908f, 2.1260054f, -0.6923556f, -0.6350248f, -1.8122731f, -1.2794832f, 0.0727381f, -0.1702943f, -1.0052446f, -1.1308233f, -0.1550428f, -5.3850775f, -0.5972114f, -0.3461530f, -0.4958812f, -1.2339170f, 0.0774786f, -0.8347382f, 0.2885106f, -0.1653138f, -0.0159324f, -0.1990542f, 1.6405339f, 1.7268503f, 0.8528640f }, +{ 0.0038449f, -0.2388104f, -0.1149939f, -1.2830256f, -0.0052918f, 0.3492746f, 0.3040406f, 0.6817812f, 0.8697712f, 0.4297135f, -1.2266355f, -0.6838435f, 0.4784633f, -0.2080981f, 0.1026195f, 0.4500216f, -0.4817615f, -1.5641849f, 1.6826199f, -0.6830738f, 0.3725868f, 0.4554783f, -0.0576175f, 1.0322572f, 1.9159367f, -0.3336473f, 0.1027016f, -0.2515334f, -0.1622610f, -1.8702512f, -0.7941946f, -2.9838107f }, +{ -0.2906766f, -0.0472568f, -0.2609593f, -1.5798510f, -0.1301123f, -1.1353090f, -2.5198724f, 1.5027611f, 1.2625716f, 3.2891662f, -2.3402910f, -0.0245398f, -9.8846655f, -0.2448200f, -0.0981539f, -0.2132508f, 0.7027491f, -3.4207478f, 1.3422097f, 4.3238688f, -1.1800685f, -2.7913725f, 1.8557802f, 5.5698090f, 0.2008359f, 0.2571939f, -0.0491005f, -0.1192926f, -0.0141392f, 0.1872108f, -0.8192848f, 0.6364858f }, +{ 0.0671812f, -0.0234023f, 0.1400131f, 0.0778011f, 0.1308578f, -0.1675161f, 0.0237332f, 0.0215410f, 0.1514422f, 0.0736446f, 0.0181612f, -0.0219220f, -0.0099684f, 0.0102909f, 0.1243076f, 0.0897413f, -0.0682666f, 0.0389046f, 0.1245468f, 0.0098897f, -0.0425716f, 0.1595688f, 0.0397469f, -0.0664724f, -0.1641733f, 0.0605745f, 0.1712325f, 0.1596854f, 0.0224220f, -0.1328268f, -0.1743169f, 0.0608113f }, +{ -1.1302176f, -0.0267920f, -0.2404846f, 0.6629997f, 0.1574058f, 4.7085104f, 0.5321816f, 0.4496275f, -0.2375129f, 1.2458097f, 1.8413728f, -1.8281459f, 2.3066695f, -0.2194766f, -0.0977243f, -0.4933192f, 0.5315795f, -0.8266373f, -1.6038543f, -0.6524141f, -0.2899630f, -4.3681431f, 0.2379541f, -0.3776518f, 1.5330435f, 0.3396196f, -0.0837295f, -0.0557159f, 0.1054505f, 1.1391810f, 0.6029354f, 0.9084895f }, +{ -0.0898137f, 0.0973598f, -0.1583033f, -0.1445000f, 0.0044857f, -0.1162652f, -0.1396598f, 0.1172944f, 0.0714815f, -0.1043701f, 0.0497041f, 0.0381792f, 0.1395592f, -0.0310079f, -0.1161861f, -0.0776435f, 0.0771079f, -0.0311096f, -0.1498033f, 0.0838249f, 0.1343088f, 0.1152903f, 0.0749620f, -0.1248982f, -0.0346379f, -0.0780947f, -0.0730843f, 0.1654648f, -0.1482577f, -0.0118278f, 0.1078758f, -0.1372479f }, +{ -0.7470447f, -0.0269854f, 0.0793572f, -0.7553226f, -0.0888210f, 1.8323143f, 0.2497575f, -10.9856701f, -0.7587091f, 1.1695130f, -0.0198075f, -0.8016124f, -0.2920889f, -0.1624553f, -0.0925370f, -1.3623806f, -1.3067284f, 0.2490501f, 0.0292220f, -0.2309522f, 0.0131977f, -0.7858045f, 0.1076498f, 0.6439033f, -0.1686209f, -0.8376603f, 0.0470587f, 0.0656591f, 0.0300643f, 1.4373403f, 0.2995886f, -1.3597747f }, +{ -0.8746193f, -0.0201447f, -0.1603909f, -0.9373384f, 0.0254792f, 0.7993639f, 0.1087815f, -0.4859151f, -1.9559643f, -2.9621441f, -1.7735658f, -2.2227323f, -0.3725011f, 0.0284650f, -0.2266676f, -0.7529051f, -0.4477961f, 0.2071678f, 2.2770953f, 1.1170574f, -0.7023426f, 0.6896675f, -1.2416189f, -1.0012786f, -1.8752691f, -0.0359559f, -0.0125555f, -0.0457818f, -0.0775177f, 0.6092747f, 0.6639680f, 1.0951738f }, +{ -1.7250717f, -0.0309622f, 0.1054536f, 2.0844676f, -0.2957073f, -0.0859962f, -1.5239947f, -0.2195731f, -1.5450290f, 0.4916542f, -1.3940116f, -0.5043938f, -0.7850562f, -0.1393226f, 0.0421535f, -0.2805571f, 0.5149463f, -0.4160199f, 1.3064783f, 0.3980424f, 0.1537250f, 0.2291165f, -1.5061877f, -0.0060016f, 0.1236818f, -0.1598873f, 0.0182953f, -0.0649871f, -0.1681922f, -0.8940877f, -0.6040494f, -0.5640311f }, +{ 0.1721584f, -0.0760296f, 0.0164177f, 1.6552234f, -0.2948481f, -0.8678465f, 0.9449416f, -2.5188146f, -0.3830298f, -0.9600880f, 1.3968240f, 0.4318309f, 1.2421557f, -0.2275233f, -0.2093997f, -0.4352153f, -0.2142241f, -1.6719555f, -1.5515612f, -1.5296252f, 0.6067497f, 2.3952096f, -1.4322679f, -1.7078539f, -0.0283693f, -0.0217880f, -0.0268028f, -0.0223932f, -0.0441584f, 0.2260311f, -0.3117388f, 0.0257804f }, +{ -0.3736052f, -0.1436419f, -0.1791924f, 0.6735403f, -0.1233307f, 0.9371016f, -0.5032559f, 0.0065924f, 0.2331814f, 0.5494700f, 1.4163370f, 0.7603047f, 1.8672758f, -0.1058595f, -0.0214494f, -0.1112118f, 0.4055682f, -0.2838995f, -1.8654461f, -0.6464235f, 0.3173830f, 0.1275925f, -0.0674378f, -0.1359007f, 0.8348094f, 0.2912215f, -0.1404384f, -0.2908648f, -0.1660577f, 0.4171410f, -0.5499387f, -0.9420084f }, +{ -10.1053333f, -0.1126838f, 0.0914457f, 2.1485925f, -0.2125181f, -0.6991695f, -1.5648814f, 3.5901635f, 0.4265155f, 2.2461650f, 1.5659872f, -4.5079947f, -1.8503289f, 0.0036438f, -0.0552862f, 6.9531665f, 7.5919185f, -5.2455883f, 1.9158086f, -0.4571260f, -8.5806217f, 3.1702795f, 4.6932554f, 0.9312032f, -0.2189541f, -35.8472443f, -0.0437394f, 0.0193345f, -0.0759754f, -0.8393922f, -8.7812824f, 1.2664827f }, +{ -0.4285544f, -0.0112712f, -0.1986143f, -0.6809276f, -0.2643681f, 0.8245937f, -0.1275320f, 1.8360085f, -1.0321223f, 1.0097107f, -0.4381949f, -0.0334159f, -0.1864567f, 0.0119804f, -0.1997918f, -0.3271705f, -0.0386158f, 1.0598557f, 0.9783336f, 1.9556490f, -0.2163550f, -2.8529193f, 0.6785525f, 1.3684491f, 1.2711524f, -0.1338721f, -0.0004762f, -0.0270106f, -0.1851630f, 0.1504625f, 0.1159483f, 1.3600112f }, +{ 0.1121258f, 0.0640089f, -0.1267886f, 0.0409490f, -0.1376833f, -0.0064835f, 0.1204074f, -0.0622866f, 0.1238085f, 0.0468351f, -0.0761505f, 0.0465574f, 0.0356306f, -0.0866522f, 0.1416768f, 0.1586308f, 0.1173862f, -0.0805042f, 0.1809647f, -0.1865381f, -0.1134762f, -0.1078163f, 0.1637889f, 0.3176169f, 0.1323416f, 0.0248524f, -0.0535714f, 0.1136851f, 0.0925486f, 0.0475878f, -0.0636055f, -0.0584198f }, +{ -14.6186295f, -0.1993816f, 0.1559632f, 2.3735945f, -0.0626924f, -8.2991180f, 0.2650338f, 3.5373747f, 1.6012716f, 2.8416808f, 1.1672604f, -4.3095527f, -1.1607343f, 0.0817207f, 0.1491408f, 7.4242001f, 7.2067833f, -4.3056359f, 0.8474119f, -0.9166479f, 0.6955321f, 2.0489492f, 4.0285969f, 0.0758955f, -1.0957255f, -38.1951637f, -0.0826333f, -0.1540980f, -0.2915384f, 1.0447723f, -9.4216700f, 0.7907819f }, +{ 1.0854393f, -0.2057757f, -0.2333630f, -2.8073475f, -0.0092103f, -0.0027702f, 0.0400727f, 1.1957222f, 2.1408370f, 2.4174137f, -4.4619761f, -2.1267519f, -2.3692110f, -0.0215195f, 0.0940003f, 1.5145874f, 0.4090779f, 1.0525370f, 0.7181456f, 2.2475765f, -4.0287180f, 0.2974668f, 1.9458562f, 0.4356304f, -1.8651968f, 0.1091178f, -0.0625651f, 0.0101182f, -0.1866876f, 2.0775406f, -0.0924411f, -0.9942529f }, +{ 0.6088188f, -0.1557965f, -0.1098123f, -2.0238423f, -0.0628158f, 1.8817046f, -2.8116503f, -1.9974719f, -2.8561919f, -1.5064508f, -0.5036830f, 0.0460033f, -1.2808908f, -0.2553190f, -0.2242925f, 0.3286708f, 0.6866547f, -0.8440441f, -3.4105279f, 1.5247352f, 0.5397177f, -0.5575525f, -0.4773433f, -0.8334278f, -1.8625259f, -0.5315101f, -0.1528733f, -0.1654517f, -0.1824549f, -0.9775680f, -2.9137912f, -0.4873172f }, +{ -3.4014626f, -0.0843555f, 0.0550500f, 1.7411346f, -0.1923270f, 0.9657442f, -2.5730155f, -0.4918169f, 0.7225664f, -1.8889602f, 3.7809584f, 1.7930225f, -1.6396630f, -0.2156895f, -0.1968789f, -0.4836586f, -1.0472771f, 3.1005175f, -3.8917956f, -4.2640653f, 0.9576553f, -0.5442499f, -0.7966165f, -7.2177167f, 3.5426965f, 0.3307767f, -0.1710942f, -0.1729530f, -0.0668134f, -0.4916894f, 1.9444537f, 0.5960958f }, +{ -0.9639474f, -0.0439117f, 0.0067572f, -0.8821743f, 0.1859997f, -0.7608641f, 1.5826563f, -1.4137437f, -0.3524639f, 0.4026311f, -0.0760814f, 0.7603636f, 2.5665376f, 0.0416676f, -0.1360063f, -0.6073055f, -1.3086349f, 1.9021115f, 1.3700111f, -0.9890064f, -0.1628284f, -1.3746781f, 0.2909067f, 0.3049279f, 0.8926148f, -0.4370743f, -0.0742112f, -0.0174381f, 0.0485841f, -0.0802396f, 3.5397923f, -0.6531604f }, +{ 0.0491318f, -0.0591713f, 0.0498613f, 1.1349666f, -0.0889358f, -1.5782114f, 0.4471801f, 1.2610462f, -0.4626705f, 0.9874090f, 1.3419105f, 0.9187180f, 0.3291495f, -0.0520998f, -0.0861611f, 0.0016949f, -0.3544937f, -0.4579405f, -0.3237021f, -0.2511895f, -3.9122045f, -0.4933105f, 1.1840253f, 0.0769721f, 1.6942110f, -0.0282394f, -0.0797413f, -0.0965752f, -0.1654065f, 0.6972140f, 0.0176173f, 0.2189070f }, +{ -1.1536870f, -0.1555014f, -0.3485503f, -1.5355002f, -0.2638741f, 1.3639680f, 1.6288967f, 0.6296268f, -0.5865426f, -0.5947766f, -1.4163440f, -1.5454646f, -0.7648816f, -0.1768368f, -0.1375362f, 0.3310686f, -0.2621009f, 0.6404898f, 1.0589499f, 0.2490305f, 0.3255049f, -0.1451550f, -0.3024975f, 0.8173084f, -2.1893589f, -0.0099213f, -0.1596652f, 0.1068924f, 0.0260035f, 1.4544648f, -0.7779374f, -0.7673930f }, +{ 0.0435760f, -0.0246287f, -0.1238125f, -0.0320143f, 0.1576336f, 0.0702628f, -0.1744860f, -0.1600130f, -0.0870369f, 0.0819791f, 0.0445500f, 0.0019862f, 0.1529912f, 0.0886548f, 0.0478378f, -0.0487068f, 0.0401030f, 0.0869702f, 0.1359502f, 0.1713400f, -0.1529109f, 0.1224933f, 0.0072967f, -0.0784628f, 0.1494704f, -0.1378658f, -0.0747428f, -0.0850826f, -0.0940957f, 0.0511818f, 0.1520686f, -0.0680635f }, +{ 1.3802143f, 0.0130436f, -0.0118297f, 0.9667702f, -0.4034582f, 0.3238442f, 0.4695910f, -0.3288718f, 1.1822759f, 1.5390285f, 0.2593471f, -0.0801804f, 0.4593041f, 0.0215906f, -0.2340092f, 0.4232760f, -0.0674659f, 0.7093915f, -0.0001982f, 0.2270681f, -0.8337699f, -0.2296698f, 0.9161285f, 0.3906515f, -0.1553026f, 0.2775815f, -0.0620411f, -0.1116518f, -0.0008762f, -0.3016884f, 1.4699873f, 0.6260686f }, +}; + +ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_output_layer[1] = { +0.7275639f }; + +ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_output_layer[32][1] = { +{ 1.4243358f }, +{ 0.0335807f }, +{ 0.0551641f }, +{ -0.9836086f }, +{ -0.0249541f }, +{ -1.5375688f }, +{ -0.7714168f }, +{ -0.9649364f }, +{ -1.1769278f }, +{ 1.3249911f }, +{ -1.6541473f }, +{ 1.4079021f }, +{ -0.8831168f }, +{ 0.0122874f }, +{ 0.0511134f }, +{ -2.6734750f }, +{ 2.8394303f }, +{ 0.9675560f }, +{ -1.4186903f }, +{ -2.0796514f }, +{ -1.7693948f }, +{ -0.8502544f }, +{ -1.5927037f }, +{ -1.1028550f }, +{ 0.8137528f }, +{ 6.3073616f }, +{ 0.1059108f }, +{ -0.0468376f }, +{ 0.1322162f }, +{ 0.7481517f }, +{ -1.2260461f }, +{ -0.9095332f }, +}; + +} + +#endif \ No newline at end of file diff --git a/RecoTracker/LSTCore/src/alpaka/Triplet.h b/RecoTracker/LSTCore/src/alpaka/Triplet.h index ae2faecb080a6..7726fe23f8277 100644 --- a/RecoTracker/LSTCore/src/alpaka/Triplet.h +++ b/RecoTracker/LSTCore/src/alpaka/Triplet.h @@ -13,6 +13,8 @@ #include "MiniDoublet.h" #include "Hit.h" +#include "NeuralNetwork.h" + namespace ALPAKA_ACCELERATOR_NAMESPACE::lst { ALPAKA_FN_ACC ALPAKA_FN_INLINE void addTripletToMemory(ModulesConst modules, @@ -730,6 +732,16 @@ namespace ALPAKA_ACCELERATOR_NAMESPACE::lst { ptCut)) return false; + bool inference = lst::t3dnn::runInference(acc, + mds, + firstMDIndex, + secondMDIndex, + thirdMDIndex, + circleRadius, + betaIn); + if (!inference) // T3-building cut + return false; + return true; } diff --git a/RecoTracker/LSTCore/standalone/analysis/DNN/train_T3_DNN.ipynb b/RecoTracker/LSTCore/standalone/analysis/DNN/train_T3_DNN.ipynb new file mode 100644 index 0000000000000..b3bf66c2140ea --- /dev/null +++ b/RecoTracker/LSTCore/standalone/analysis/DNN/train_T3_DNN.ipynb @@ -0,0 +1,964 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": 1, + "metadata": {}, + "outputs": [], + "source": [ + "import os\n", + "import uproot\n", + "import numpy as np\n", + "\n", + "def load_root_file(file_path, branches=None, print_branches=False):\n", + " all_branches = {}\n", + " with uproot.open(file_path) as file:\n", + " tree = file[\"tree\"]\n", + " # Load all ROOT branches into array if not specified\n", + " if branches is None:\n", + " branches = tree.keys()\n", + " # Option to print the branch names\n", + " if print_branches:\n", + " print(\"Branches:\", tree.keys())\n", + " # Each branch is added to the dictionary\n", + " for branch in branches:\n", + " try:\n", + " all_branches[branch] = (tree[branch].array(library=\"np\"))\n", + " except uproot.KeyInFileError as e:\n", + " print(f\"KeyInFileError: {e}\")\n", + " # Number of events in file\n", + " all_branches['event'] = tree.num_entries\n", + " return all_branches\n", + "\n", + "branches_list = [\n", + " # Core T3 properties from TripletsSoA\n", + " 't3_betaIn',\n", + " 't3_centerX',\n", + " 't3_centerY',\n", + " 't3_radius',\n", + " 't3_partOfPT5',\n", + " 't3_partOfT5',\n", + " 't3_partOfPT3',\n", + " 't3_layer_binary',\n", + " 't3_pMatched',\n", + " 't3_matched_simIdx',\n", + "]\n", + "\n", + "# Hit-dependent branches\n", + "suffixes = ['r', 'z', 'eta', 'phi', 'layer']\n", + "branches_list += [f't3_hit_{i}_{suffix}' for i in [0, 1, 2, 3, 4, 5] for suffix in suffixes]\n", + "\n", + "file_path = \"t3_dnn_train_175.root\"\n", + "branches = load_root_file(file_path, branches_list)" + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Z max: 224.14950561523438, R max: 98.93236541748047, Eta max: 2.5\n" + ] + } + ], + "source": [ + "z_max = np.max([np.max(event) for event in branches[f't3_hit_3_z']])\n", + "r_max = np.max([np.max(event) for event in branches[f't3_hit_3_r']])\n", + "eta_max = 2.5\n", + "phi_max = np.pi\n", + "\n", + "print(f'Z max: {z_max}, R max: {r_max}, Eta max: {eta_max}')\n", + "\n", + "def delta_phi(phi1, phi2):\n", + " delta = phi1 - phi2\n", + " # Adjust delta to be within the range [-pi, pi]\n", + " if delta > np.pi:\n", + " delta -= 2 * np.pi\n", + " elif delta < -np.pi:\n", + " delta += 2 * np.pi\n", + " return delta" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "metadata": {}, + "outputs": [], + "source": [ + "features_list = []\n", + "eta_list = []\n", + "\n", + "for event in range(branches['event']):\n", + " num_elements = len(branches['t3_betaIn'][event])\n", + " for i in range(num_elements):\n", + " features_iter = []\n", + " eta_iter = []\n", + " \n", + " # Get properties for the 3 selected hits (0, 2, 4)\n", + " eta1 = np.abs(branches['t3_hit_0_eta'][event][i])\n", + " eta3 = np.abs(branches['t3_hit_2_eta'][event][i])\n", + " eta5 = np.abs(branches['t3_hit_4_eta'][event][i])\n", + "\n", + " phi1 = branches['t3_hit_0_phi'][event][i]\n", + " phi3 = branches['t3_hit_2_phi'][event][i]\n", + " phi5 = branches['t3_hit_4_phi'][event][i]\n", + "\n", + " z1 = np.abs(branches['t3_hit_0_z'][event][i])\n", + " z3 = np.abs(branches['t3_hit_2_z'][event][i])\n", + " z5 = np.abs(branches['t3_hit_4_z'][event][i])\n", + "\n", + " r1 = branches['t3_hit_0_r'][event][i]\n", + " r3 = branches['t3_hit_2_r'][event][i]\n", + " r5 = branches['t3_hit_4_r'][event][i]\n", + "\n", + " # T3-specific properties\n", + " radius = branches['t3_radius'][event][i]\n", + " centerX = branches['t3_centerX'][event][i]\n", + " centerY = branches['t3_centerY'][event][i]\n", + " betaIn = branches['t3_betaIn'][event][i]\n", + "\n", + " features_iter = [\n", + " eta1 / eta_max, # First hit eta, normalized\n", + " np.abs(phi1) / phi_max, # First hit phi, normalized\n", + " z1 / z_max, # First hit z, normalized\n", + " r1 / r_max, # First hit r, normalized\n", + "\n", + " eta3 - eta1, # Difference in eta between hit 3 and 1\n", + " delta_phi(phi3, phi1) / phi_max, # Difference in phi between hit 3 and 1\n", + " (z3 - z1) / z_max, # Difference in z between hit 3 and 1, normalized\n", + " (r3 - r1) / r_max, # Difference in r between hit 3 and 1, normalized\n", + "\n", + " eta5 - eta3, # Difference in eta between hit 5 and 3\n", + " delta_phi(phi5, phi3) / phi_max, # Difference in phi between hit 5 and 3\n", + " (z5 - z3) / z_max, # Difference in z between hit 5 and 3, normalized\n", + " (r5 - r3) / r_max, # Difference in r between hit 5 and 3, normalized\n", + "\n", + " np.log10(radius), # T3's circle radius\n", + " betaIn # Beta angle of inner segment\n", + " ]\n", + "\n", + " eta_iter.extend([eta1]) # Use first hit eta for cut thresholds\n", + " features_list.append(features_iter)\n", + " eta_list.append(eta_iter)\n", + "\n", + "features = np.array(features_list).T\n", + "eta_list = np.array(eta_list).T" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "metadata": {}, + "outputs": [], + "source": [ + "import torch\n", + "\n", + "# Stack features along a new axis to form a single array suitable for NN input\n", + "input_features_numpy = np.stack(features, axis=-1)\n", + "\n", + "# Identify rows with NaN or Inf values\n", + "mask = ~np.isnan(input_features_numpy) & ~np.isinf(input_features_numpy)\n", + "\n", + "# Apply mask across all columns: retain a row only if all its entries are neither NaN nor Inf\n", + "filtered_input_features_numpy = input_features_numpy[np.all(mask, axis=1)]\n", + "t5_isFake_filtered = (np.concatenate(branches['t3_pMatched']) < 0.9)[np.all(mask, axis=1)]\n", + "\n", + "# Convert to PyTorch tensor when ready to use with NN\n", + "input_features_tensor = torch.tensor(filtered_input_features_numpy, dtype=torch.float32)" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Using device: cuda\n", + "Initial dataset size: 15131951\n", + "Dataset size after initial 100.0% downsampling: 15131951\n", + "Class distribution after initial downsampling - Class 0: 13896802, Class 1: 1235149\n", + "Final class distribution after balancing - Class 0: 1235149, Class 1: 1235149\n", + "Epoch [1/150], Loss: 0.2612, Test Acc: 89.49%\n", + "Epoch [2/150], Loss: 0.2496, Test Acc: 91.23%\n", + "Epoch [3/150], Loss: 0.2351, Test Acc: 92.18%\n", + "Epoch [4/150], Loss: 0.1740, Test Acc: 92.70%\n", + "Epoch [5/150], Loss: 0.2017, Test Acc: 93.42%\n", + "Epoch [6/150], Loss: 0.1938, Test Acc: 93.49%\n", + "Epoch [7/150], Loss: 0.1705, Test Acc: 93.64%\n", + "Epoch [8/150], Loss: 0.1661, Test Acc: 93.77%\n", + "Epoch [9/150], Loss: 0.1995, Test Acc: 93.77%\n", + "Epoch [10/150], Loss: 0.1610, Test Acc: 93.71%\n", + "Epoch [11/150], Loss: 0.1647, Test Acc: 93.94%\n", + "Epoch [12/150], Loss: 0.1356, Test Acc: 93.88%\n", + "Epoch [13/150], Loss: 0.1737, Test Acc: 93.91%\n", + "Epoch [14/150], Loss: 0.1596, Test Acc: 93.91%\n", + "Epoch [15/150], Loss: 0.1907, Test Acc: 93.97%\n", + "Epoch [16/150], Loss: 0.2036, Test Acc: 94.02%\n", + "Epoch [17/150], Loss: 0.1486, Test Acc: 93.96%\n", + "Epoch [18/150], Loss: 0.1830, Test Acc: 94.00%\n", + "Epoch [19/150], Loss: 0.1912, Test Acc: 93.86%\n", + "Epoch [20/150], Loss: 0.1674, Test Acc: 94.03%\n", + "Epoch [21/150], Loss: 0.1751, Test Acc: 93.84%\n", + "Epoch [22/150], Loss: 0.1354, Test Acc: 93.87%\n", + "Epoch [23/150], Loss: 0.1845, Test Acc: 93.99%\n", + "Epoch [24/150], Loss: 0.1562, Test Acc: 94.14%\n", + "Epoch [25/150], Loss: 0.1947, Test Acc: 93.74%\n", + "Epoch [26/150], Loss: 0.1526, Test Acc: 94.03%\n", + "Epoch [27/150], Loss: 0.1891, Test Acc: 94.14%\n", + "Epoch [28/150], Loss: 0.1824, Test Acc: 94.09%\n", + "Epoch [29/150], Loss: 0.1931, Test Acc: 94.13%\n", + "Epoch [30/150], Loss: 0.1595, Test Acc: 93.95%\n", + "Epoch [31/150], Loss: 0.1625, Test Acc: 94.10%\n", + "Epoch [32/150], Loss: 0.1604, Test Acc: 94.07%\n", + "Epoch [33/150], Loss: 0.1826, Test Acc: 94.03%\n", + "Epoch [34/150], Loss: 0.1632, Test Acc: 93.78%\n", + "Epoch [35/150], Loss: 0.1870, Test Acc: 93.95%\n", + "Epoch [36/150], Loss: 0.1606, Test Acc: 94.01%\n", + "Epoch [37/150], Loss: 0.1368, Test Acc: 94.04%\n", + "Epoch [38/150], Loss: 0.1544, Test Acc: 94.15%\n", + "Epoch [39/150], Loss: 0.1447, Test Acc: 94.14%\n", + "Epoch [40/150], Loss: 0.1664, Test Acc: 94.16%\n", + "Epoch [41/150], Loss: 0.1780, Test Acc: 93.99%\n", + "Epoch [42/150], Loss: 0.1573, Test Acc: 94.16%\n", + "Epoch [43/150], Loss: 0.1633, Test Acc: 94.08%\n", + "Epoch [44/150], Loss: 0.1728, Test Acc: 94.17%\n", + "Epoch [45/150], Loss: 0.1685, Test Acc: 94.20%\n", + "Epoch [46/150], Loss: 0.2017, Test Acc: 93.97%\n", + "Epoch [47/150], Loss: 0.1557, Test Acc: 94.03%\n", + "Epoch [48/150], Loss: 0.1593, Test Acc: 94.23%\n", + "Epoch [49/150], Loss: 0.1554, Test Acc: 94.16%\n", + "Epoch [50/150], Loss: 0.1530, Test Acc: 94.16%\n", + "Epoch [51/150], Loss: 0.1641, Test Acc: 94.25%\n", + "Epoch [52/150], Loss: 0.1831, Test Acc: 94.24%\n", + "Epoch [53/150], Loss: 0.1391, Test Acc: 94.17%\n", + "Epoch [54/150], Loss: 0.1704, Test Acc: 94.19%\n", + "Epoch [55/150], Loss: 0.1355, Test Acc: 94.18%\n", + "Epoch [56/150], Loss: 0.1590, Test Acc: 94.21%\n", + "Epoch [57/150], Loss: 0.1926, Test Acc: 94.27%\n", + "Epoch [58/150], Loss: 0.1704, Test Acc: 94.28%\n", + "Epoch [59/150], Loss: 0.1541, Test Acc: 94.08%\n", + "Epoch [60/150], Loss: 0.1755, Test Acc: 94.31%\n", + "Epoch [61/150], Loss: 0.1585, Test Acc: 94.16%\n", + "Epoch [62/150], Loss: 0.1579, Test Acc: 94.21%\n", + "Epoch [63/150], Loss: 0.1508, Test Acc: 94.12%\n", + "Epoch [64/150], Loss: 0.1600, Test Acc: 94.31%\n", + "Epoch [65/150], Loss: 0.1678, Test Acc: 94.32%\n", + "Epoch [66/150], Loss: 0.1572, Test Acc: 94.30%\n", + "Epoch [67/150], Loss: 0.1731, Test Acc: 94.24%\n", + "Epoch [68/150], Loss: 0.1651, Test Acc: 93.88%\n", + "Epoch [69/150], Loss: 0.1626, Test Acc: 94.10%\n", + "Epoch [70/150], Loss: 0.1486, Test Acc: 94.13%\n", + "Epoch [71/150], Loss: 0.1659, Test Acc: 93.81%\n", + "Epoch [72/150], Loss: 0.1617, Test Acc: 93.97%\n", + "Epoch [73/150], Loss: 0.1658, Test Acc: 94.17%\n", + "Epoch [74/150], Loss: 0.1795, Test Acc: 93.94%\n", + "Epoch [75/150], Loss: 0.1849, Test Acc: 94.26%\n", + "Epoch [76/150], Loss: 0.1364, Test Acc: 94.30%\n", + "Epoch [77/150], Loss: 0.1865, Test Acc: 94.08%\n", + "Epoch [78/150], Loss: 0.1890, Test Acc: 94.24%\n", + "Epoch [79/150], Loss: 0.1618, Test Acc: 94.24%\n", + "Epoch [80/150], Loss: 0.1703, Test Acc: 94.31%\n", + "Epoch [81/150], Loss: 0.1239, Test Acc: 94.28%\n", + "Epoch [82/150], Loss: 0.1638, Test Acc: 94.38%\n", + "Epoch [83/150], Loss: 0.1834, Test Acc: 94.37%\n", + "Epoch [84/150], Loss: 0.1581, Test Acc: 94.19%\n", + "Epoch [85/150], Loss: 0.1982, Test Acc: 94.29%\n", + "Epoch [86/150], Loss: 0.1707, Test Acc: 94.31%\n", + "Epoch [87/150], Loss: 0.1748, Test Acc: 94.28%\n", + "Epoch [88/150], Loss: 0.1615, Test Acc: 94.27%\n", + "Epoch [89/150], Loss: 0.1569, Test Acc: 94.28%\n", + "Epoch [90/150], Loss: 0.1377, Test Acc: 94.29%\n", + "Epoch [91/150], Loss: 0.1430, Test Acc: 94.33%\n", + "Epoch [92/150], Loss: 0.1575, Test Acc: 93.93%\n", + "Epoch [93/150], Loss: 0.1722, Test Acc: 94.36%\n", + "Epoch [94/150], Loss: 0.1522, Test Acc: 94.36%\n", + "Epoch [95/150], Loss: 0.1474, Test Acc: 94.39%\n", + "Epoch [96/150], Loss: 0.1479, Test Acc: 94.25%\n", + "Epoch [97/150], Loss: 0.1311, Test Acc: 94.31%\n", + "Epoch [98/150], Loss: 0.1569, Test Acc: 94.31%\n", + "Epoch [99/150], Loss: 0.1634, Test Acc: 94.35%\n", + "Epoch [100/150], Loss: 0.1541, Test Acc: 94.38%\n", + "Epoch [101/150], Loss: 0.1606, Test Acc: 94.30%\n", + "Epoch [102/150], Loss: 0.1845, Test Acc: 94.26%\n", + "Epoch [103/150], Loss: 0.1361, Test Acc: 94.26%\n", + "Epoch [104/150], Loss: 0.1923, Test Acc: 94.40%\n", + "Epoch [105/150], Loss: 0.1552, Test Acc: 94.27%\n", + "Epoch [106/150], Loss: 0.1352, Test Acc: 94.36%\n", + "Epoch [107/150], Loss: 0.1963, Test Acc: 94.21%\n", + "Epoch [108/150], Loss: 0.1819, Test Acc: 94.41%\n", + "Epoch [109/150], Loss: 0.1593, Test Acc: 94.30%\n", + "Epoch [110/150], Loss: 0.1741, Test Acc: 94.16%\n", + "Epoch [111/150], Loss: 0.1899, Test Acc: 94.35%\n", + "Epoch [112/150], Loss: 0.1653, Test Acc: 94.32%\n", + "Epoch [113/150], Loss: 0.1597, Test Acc: 94.34%\n", + "Epoch [114/150], Loss: 0.1502, Test Acc: 94.40%\n", + "Epoch [115/150], Loss: 0.1510, Test Acc: 94.36%\n", + "Epoch [116/150], Loss: 0.1306, Test Acc: 94.41%\n", + "Epoch [117/150], Loss: 0.1805, Test Acc: 94.38%\n", + "Epoch [118/150], Loss: 0.1518, Test Acc: 94.15%\n", + "Epoch [119/150], Loss: 0.2202, Test Acc: 94.32%\n", + "Epoch [120/150], Loss: 0.1514, Test Acc: 94.27%\n", + "Epoch [121/150], Loss: 0.1433, Test Acc: 94.35%\n", + "Epoch [122/150], Loss: 0.1583, Test Acc: 94.41%\n", + "Epoch [123/150], Loss: 0.1809, Test Acc: 94.36%\n", + "Epoch [124/150], Loss: 0.1712, Test Acc: 94.40%\n", + "Epoch [125/150], Loss: 0.1765, Test Acc: 94.41%\n", + "Epoch [126/150], Loss: 0.1582, Test Acc: 94.38%\n", + "Epoch [127/150], Loss: 0.1797, Test Acc: 94.26%\n", + "Epoch [128/150], Loss: 0.1559, Test Acc: 94.25%\n", + "Epoch [129/150], Loss: 0.1222, Test Acc: 94.31%\n", + "Epoch [130/150], Loss: 0.1351, Test Acc: 94.42%\n", + "Epoch [131/150], Loss: 0.1282, Test Acc: 94.38%\n", + "Epoch [132/150], Loss: 0.1631, Test Acc: 94.44%\n", + "Epoch [133/150], Loss: 0.1531, Test Acc: 94.18%\n", + "Epoch [134/150], Loss: 0.1375, Test Acc: 94.45%\n", + "Epoch [135/150], Loss: 0.1421, Test Acc: 94.43%\n", + "Epoch [136/150], Loss: 0.1712, Test Acc: 93.80%\n", + "Epoch [137/150], Loss: 0.1655, Test Acc: 94.41%\n", + "Epoch [138/150], Loss: 0.1699, Test Acc: 94.34%\n", + "Epoch [139/150], Loss: 0.1899, Test Acc: 94.36%\n", + "Epoch [140/150], Loss: 0.1488, Test Acc: 94.40%\n", + "Epoch [141/150], Loss: 0.1621, Test Acc: 94.36%\n", + "Epoch [142/150], Loss: 0.1432, Test Acc: 94.33%\n", + "Epoch [143/150], Loss: 0.1767, Test Acc: 94.27%\n", + "Epoch [144/150], Loss: 0.1421, Test Acc: 94.46%\n", + "Epoch [145/150], Loss: 0.1765, Test Acc: 94.42%\n", + "Epoch [146/150], Loss: 0.1434, Test Acc: 94.36%\n", + "Epoch [147/150], Loss: 0.1741, Test Acc: 94.39%\n", + "Epoch [148/150], Loss: 0.1763, Test Acc: 94.30%\n", + "Epoch [149/150], Loss: 0.2035, Test Acc: 94.36%\n", + "Epoch [150/150], Loss: 0.1627, Test Acc: 94.44%\n" + ] + } + ], + "source": [ + "from torch import nn\n", + "from torch.optim import Adam\n", + "from torch.utils.data import DataLoader, TensorDataset, random_split\n", + "import torch\n", + "\n", + "# Set device\n", + "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n", + "print(f\"Using device: {device}\")\n", + "\n", + "# Create labels tensor\n", + "labels_tensor = 1 - torch.tensor(t5_isFake_filtered, dtype=torch.float32)\n", + "\n", + "# Set initial downsample fraction\n", + "initial_downsample_fraction = 1.0 # Adjust this value as needed\n", + "\n", + "class MyNeuralNetwork(nn.Module):\n", + " def __init__(self):\n", + " super(MyNeuralNetwork, self).__init__()\n", + " self.layer1 = nn.Linear(input_features_numpy.shape[1], 32)\n", + " self.layer2 = nn.Linear(32, 32)\n", + " self.output_layer = nn.Linear(32, 1)\n", + "\n", + " def forward(self, x):\n", + " x = self.layer1(x)\n", + " x = nn.ReLU()(x)\n", + " x = self.layer2(x)\n", + " x = nn.ReLU()(x)\n", + " x = self.output_layer(x)\n", + " x = torch.sigmoid(x)\n", + " return x\n", + "\n", + "class WeightedBCELoss(nn.Module):\n", + " def __init__(self):\n", + " super(WeightedBCELoss, self).__init__()\n", + " \n", + " def forward(self, outputs, targets):\n", + " eps = 1e-7\n", + " losses = -((targets * torch.log(outputs + eps) + \n", + " (1 - targets) * torch.log(1 - outputs + eps)))\n", + " return losses.mean()\n", + "\n", + "\n", + "# Print initial dataset size\n", + "print(f\"Initial dataset size: {len(labels_tensor)}\")\n", + "\n", + "# Remove rows with NaN and update weights accordingly\n", + "nan_mask = torch.isnan(input_features_tensor).any(dim=1) | torch.isnan(labels_tensor)\n", + "filtered_inputs = input_features_tensor[~nan_mask]\n", + "filtered_labels = labels_tensor[~nan_mask]\n", + "\n", + "# Initial downsampling of entire dataset\n", + "if initial_downsample_fraction < 1.0:\n", + " total_samples = len(filtered_labels)\n", + " samples_to_keep = int(total_samples * initial_downsample_fraction)\n", + " indices = torch.randperm(total_samples)[:samples_to_keep]\n", + " filtered_inputs = filtered_inputs[indices]\n", + " filtered_labels = filtered_labels[indices]\n", + "\n", + "print(f\"Dataset size after initial {initial_downsample_fraction*100}% downsampling: {len(filtered_labels)}\")\n", + "\n", + "# Count samples in each class after initial downsampling\n", + "class_counts = torch.bincount(filtered_labels.int())\n", + "print(f\"Class distribution after initial downsampling - Class 0: {class_counts[0]}, Class 1: {class_counts[1]}\")\n", + "\n", + "# Balance classes while maintaining weights\n", + "minority_class = 0 if class_counts[0] < class_counts[1] else 1\n", + "minority_indices = (filtered_labels == minority_class).nonzero(as_tuple=True)[0]\n", + "majority_indices = (filtered_labels == (1 - minority_class)).nonzero(as_tuple=True)[0]\n", + "downsampled_majority_indices = majority_indices[torch.randperm(len(majority_indices))[:len(minority_indices)]]\n", + "balanced_indices = torch.cat((minority_indices, downsampled_majority_indices))\n", + "\n", + "# Create balanced dataset with weights\n", + "balanced_inputs = filtered_inputs[balanced_indices]\n", + "balanced_labels = filtered_labels[balanced_indices]\n", + "\n", + "# Verify balanced distribution\n", + "balanced_counts = torch.bincount(balanced_labels.int())\n", + "print(f\"Final class distribution after balancing - Class 0: {balanced_counts[0]}, Class 1: {balanced_counts[1]}\")\n", + "\n", + "# Create dataset with weights\n", + "dataset = TensorDataset(balanced_inputs, balanced_labels)\n", + "\n", + "# Split into train and test sets\n", + "train_size = int(0.8 * len(dataset))\n", + "test_size = len(dataset) - train_size\n", + "train_dataset, test_dataset = random_split(dataset, [train_size, test_size])\n", + "\n", + "# Create data loaders\n", + "train_loader = DataLoader(train_dataset, batch_size=1024, shuffle=True, num_workers=10, pin_memory=True)\n", + "test_loader = DataLoader(test_dataset, batch_size=1024, shuffle=False, num_workers=10, pin_memory=True)\n", + "\n", + "# Initialize model and optimizer\n", + "model = MyNeuralNetwork().to(device)\n", + "loss_function = WeightedBCELoss()\n", + "optimizer = Adam(model.parameters(), lr=0.0025)\n", + "\n", + "def evaluate_model(loader):\n", + " model.eval()\n", + " total = 0\n", + " correct = 0\n", + " with torch.no_grad():\n", + " for inputs, targets in loader:\n", + " inputs, targets = inputs.to(device), targets.to(device)\n", + " outputs = model(inputs)\n", + " predicted = outputs.squeeze() > 0.5\n", + " total += targets.size(0)\n", + " correct += (predicted == targets.bool()).sum().item()\n", + " model.train()\n", + " return 100 * correct / total\n", + "\n", + "# Training loop\n", + "num_epochs = 150\n", + "loss_log = []\n", + "\n", + "for epoch in range(num_epochs):\n", + " for inputs, targets in train_loader:\n", + " inputs, targets = inputs.to(device), targets.to(device)\n", + " \n", + " # Forward pass\n", + " outputs = model(inputs)\n", + " loss = loss_function(outputs.squeeze(), targets)\n", + " \n", + " loss_log.append(loss.item())\n", + "\n", + " # Backward and optimize\n", + " optimizer.zero_grad()\n", + " loss.backward()\n", + " optimizer.step()\n", + "\n", + " test_accuracy = evaluate_model(test_loader)\n", + " print(f'Epoch [{epoch+1}/{num_epochs}], Loss: {loss.item():.4f}, Test Acc: {test_accuracy:.2f}%')" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": {}, + "outputs": [], + "source": [ + "torch.save(model.state_dict(), \"model.pth\")" + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Baseline accuracy: 0.9399382472038269\n", + "Feature importances:\n", + "Feature 13 importance: 0.0324\n", + "Feature 9 importance: 0.0289\n", + "Feature 8 importance: 0.0175\n", + "Feature 4 importance: 0.0169\n", + "Feature 0 importance: 0.0092\n", + "Feature 12 importance: 0.0072\n", + "Feature 3 importance: 0.0037\n", + "Feature 6 importance: 0.0036\n", + "Feature 2 importance: 0.0032\n", + "Feature 10 importance: 0.0012\n", + "Feature 5 importance: 0.0011\n", + "Feature 11 importance: 0.0002\n", + "Feature 1 importance: -0.0000\n", + "Feature 7 importance: -0.0005\n" + ] + } + ], + "source": [ + "from sklearn.metrics import accuracy_score\n", + "\n", + "# Convert tensors to numpy for simplicity in permutation\n", + "input_features_np = input_features_tensor.numpy()\n", + "labels_np = labels_tensor.numpy()\n", + "\n", + "def model_accuracy(features, labels, model):\n", + " model.eval() # Set the model to evaluation mode\n", + " inputs = features.to(device)\n", + " labels = labels.to(device)\n", + " with torch.no_grad():\n", + " outputs = model(inputs)\n", + " predicted = (outputs.squeeze() > 0.5).float() # Update threshold as necessary\n", + " accuracy = (predicted == labels).float().mean().item()\n", + " return accuracy\n", + "\n", + "# Use the original input_features_tensor and labels_tensor directly\n", + "baseline_accuracy = model_accuracy(input_features_tensor, labels_tensor, model)\n", + "print(f\"Baseline accuracy: {baseline_accuracy}\")\n", + "\n", + "# Initialize an array to store feature importances\n", + "feature_importances = np.zeros(input_features_tensor.shape[1])\n", + "\n", + "# Permute each feature and calculate the drop in accuracy\n", + "for i in range(input_features_tensor.shape[1]):\n", + " permuted_features = input_features_tensor.clone()\n", + " permuted_features[:, i] = permuted_features[torch.randperm(permuted_features.size(0)), i] # Permute feature\n", + "\n", + " permuted_accuracy = model_accuracy(permuted_features, labels_tensor, model)\n", + " feature_importances[i] = baseline_accuracy - permuted_accuracy\n", + "\n", + "# Ranking features by importance\n", + "important_features_indices = np.argsort(feature_importances)[::-1] # Indices of features in descending importance\n", + "important_features_scores = np.sort(feature_importances)[::-1] # Importance scores in descending order\n", + "\n", + "print(\"Feature importances:\")\n", + "for idx, score in zip(important_features_indices, important_features_scores):\n", + " print(f\"Feature {idx} importance: {score:.4f}\")" + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "metadata": {}, + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/tmp/ipykernel_1791539/52354147.py:7: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n", + " inputs = torch.tensor(features, dtype=torch.float32).to(device)\n" + ] + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "from sklearn.metrics import roc_curve, auc\n", + "import matplotlib.pyplot as plt\n", + "\n", + "def model_outputs(features, model):\n", + " model.eval() # Set the model to evaluation mode\n", + " with torch.no_grad():\n", + " inputs = torch.tensor(features, dtype=torch.float32).to(device)\n", + " outputs = model(inputs).squeeze().cpu().numpy()\n", + " return outputs\n", + "\n", + "# Calculate model outputs\n", + "probabilities = model_outputs(filtered_inputs, model)\n", + "\n", + "# Calculate ROC curve and AUC\n", + "fpr, tpr, thresholds = roc_curve(filtered_labels, probabilities)\n", + "roc_auc = auc(fpr, tpr)\n", + "\n", + "# Plot ROC curve\n", + "plt.figure()\n", + "lw = 2 # Line width\n", + "plt.plot(fpr, tpr, color='darkorange', lw=lw, label='ROC curve (area = %0.3f)' % roc_auc)\n", + "plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')\n", + "plt.xlim([0.0, 1.0])\n", + "plt.ylim([0.0, 1.05])\n", + "plt.xlabel('False Positive Rate')\n", + "plt.ylabel('True Positive Rate')\n", + "plt.title('Receiver Operating Characteristic')\n", + "plt.legend(loc=\"lower right\")\n", + "plt.show()" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_layer1[32] = {\n", + "-0.1918962f, -1.0575372f, -0.8276564f, -0.0243965f, -0.1577621f, 1.0067693f, 1.5348158f, 0.4439710f, 0.0041234f, -1.1558943f, -1.4180470f, 1.0221841f, -0.0592227f, -1.2107433f, -0.2100758f, 1.2193928f, -0.3124787f, -1.9197327f, -0.8064887f, -0.2178766f, -0.0111392f, -0.1638742f, 0.0029338f, -0.0157688f, 0.2662797f, 1.8194629f, 0.8465537f, -0.7592145f, -0.8783396f, 0.5602613f, -0.0764334f, -0.8502049f };\n", + "\n", + "ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_layer1[14][32] = {\n", + "{ -0.2258795f, 0.4783621f, 0.4048500f, -0.2535419f, -0.2204617f, 0.2292106f, 0.5185162f, 1.0356801f, -0.1940614f, 0.7601448f, -0.2762348f, 0.4394473f, -0.2472922f, 0.6392686f, 0.0174135f, 1.1337029f, 0.5831500f, -0.5536784f, 0.1608927f, 0.3617360f, -0.0014275f, 0.1839313f, -0.1858799f, 0.0080405f, 0.1321980f, 0.4460649f, 0.4296648f, 0.4569599f, -0.2347775f, -0.3548780f, 0.1986685f, 0.2211701f },\n", + "{ 0.1509047f, 0.0057684f, -0.0019561f, -0.1210250f, -0.2612511f, -0.0098326f, 0.0019430f, 0.0152595f, -0.2313585f, 0.0117159f, -0.0087940f, -0.0000195f, 0.2194081f, -0.0033250f, 0.1478879f, 0.0097880f, -0.0041943f, 0.0362815f, -0.0197458f, -0.0063192f, -0.0004957f, -0.0104775f, -0.2756116f, -0.0049759f, -0.0015340f, -0.0111502f, 0.0034782f, -0.0100520f, 0.0124440f, 0.0076567f, -0.0263710f, 0.0381163f },\n", + "{ 0.1448244f, 0.0803795f, 1.6524559f, 0.1457684f, -0.1350329f, -3.6343203f, -1.6798589f, -0.0142065f, 0.0835910f, 0.0497492f, 0.6378320f, -0.4540107f, -0.2349942f, -0.0257038f, -0.2397820f, -4.3412380f, 0.1279643f, 0.9793525f, -0.5925660f, -0.8705363f, 0.0134403f, -0.1410540f, -0.0032170f, 0.0107955f, -1.5223076f, -0.1977515f, -0.9901226f, -0.5315630f, 0.5766137f, 0.8363163f, -0.2219186f, -1.3835622f },\n", + "{ -0.1057612f, 0.8005342f, -1.4052893f, -0.1196175f, 0.1360446f, 1.2852736f, -4.4616752f, 0.3914140f, -0.2356829f, 1.0064709f, 1.5389245f, -3.2000272f, 0.0828165f, 0.8944567f, -0.1592458f, 0.5353487f, 0.8051971f, 0.2057788f, 1.7512299f, 0.9215795f, -0.0036782f, 2.1131773f, -0.3204709f, -0.0044941f, 0.0865040f, 0.4481153f, -1.2256490f, 1.7857696f, -0.6445597f, -0.7477201f, 0.2316555f, -0.7155212f },\n", + "{ 0.0901890f, -1.1808448f, 1.4875709f, -0.0262624f, -0.1323451f, -0.8950851f, 0.7974008f, -2.7021067f, 0.0281707f, -1.3674084f, -4.0485940f, 5.8682752f, -0.1891649f, 7.2682476f, 0.1300428f, -2.2845411f, -8.7903214f, 0.6268425f, -1.9901284f, -2.5285044f, 0.0066100f, 7.4012928f, -0.0831600f, 0.0095476f, 6.1734300f, 0.8520991f, -8.3464308f, 2.4261341f, 3.6403832f, 5.9824433f, -0.0999645f, -0.2824311f },\n", + "{ -0.0335586f, 16.4833145f, 1.4688981f, -0.1789742f, 0.2243387f, 0.1425887f, 0.1262731f, -4.6265879f, -0.0683203f, -12.3650198f, 1.3732860f, -0.2777069f, -0.0603657f, 0.8076705f, 0.0482012f, -0.7798629f, -1.6209395f, -0.0720773f, -0.3990867f, -1.0641636f, -0.0139909f, -0.5711431f, -0.0209516f, 0.0946307f, 0.0096234f, 0.7337275f, 0.4466013f, 1.5696099f, -5.6807351f, -1.6136186f, -0.2170101f, -0.9905877f },\n", + "{ -0.0965901f, 0.4846557f, -0.0652100f, -0.2353250f, -0.1215300f, -3.5360579f, -2.0285676f, -2.8374794f, -0.1559567f, 0.7083026f, -0.1030692f, -1.1805813f, 0.0706595f, -0.2951699f, -0.0989490f, -2.8418458f, 0.8427116f, -2.2046885f, -1.3581268f, 1.5271128f, -0.0045855f, 0.1438956f, -0.2343508f, -0.0425132f, 1.0763166f, 1.0795754f, 0.4891663f, 2.6029303f, -1.0363307f, 2.2133234f, -0.2510982f, 1.1560025f },\n", + "{ 0.0074010f, -0.4152716f, 3.4255989f, -0.1664258f, 0.2028382f, 1.2840286f, 1.3538148f, -1.5683322f, 0.0429628f, -0.2171310f, 2.8748064f, -0.5141454f, 0.1099870f, -0.4614673f, -0.2515875f, -0.3862101f, 1.6164609f, -1.3735241f, -0.8896182f, 1.8310896f, 0.0188789f, -0.8612124f, -0.2143756f, 0.0385537f, -8.3583250f, 1.3280702f, -7.0806746f, -0.2001765f, 0.8213225f, -2.8566475f, 0.1479015f, 0.5155560f },\n", + "{ 0.0511283f, -1.9676421f, 4.3773866f, -0.1648336f, 0.0505374f, -3.0326672f, 2.5867159f, 1.8804691f, -0.1412510f, -2.3595653f, 3.9877729f, -14.6059513f, -0.0444889f, 7.5017557f, -0.0684788f, 1.0775388f, -8.0584450f, -0.2013538f, -12.5508022f, 9.2100115f, -0.0011293f, -7.0289955f, 0.0521985f, 0.0155513f, -1.6320782f, -4.0694451f, 7.8784938f, 8.9188061f, 10.7590771f, -9.3559170f, -0.1319706f, -0.0336969f },\n", + "{ -0.1288323f, 7.9389300f, -0.4022054f, 0.0851088f, -0.2593997f, -0.3426172f, 0.3062155f, 1.9866818f, -0.1748946f, -9.7667055f, -0.6524503f, 0.3385161f, 0.1884300f, 9.0051126f, -0.2027990f, 0.3781536f, -0.1085841f, -5.2012277f, -0.5169301f, 2.4512987f, -14.8858967f, 0.8265918f, 0.1330098f, 14.9286051f, 0.4946520f, 1.2087970f, -1.8545671f, 0.7865057f, -3.1178455f, 2.2735860f, 0.1769712f, 0.7800660f },\n", + "{ -0.0041451f, -0.3216657f, 0.2317746f, -0.0770410f, -0.1181624f, -1.6830167f, 1.3180428f, 1.7278676f, -0.1011897f, -0.8716651f, 0.8045275f, 1.0895320f, 0.0222730f, -1.0280910f, 0.1943782f, -3.5416641f, -1.8209232f, -0.6835734f, 1.1706284f, -1.0224590f, -0.0421350f, 0.6345906f, -0.2030822f, -0.0540403f, 1.6704807f, 0.2792225f, 0.5046398f, 1.1982751f, 0.1087174f, -0.6976060f, 0.1717374f, 1.3827170f },\n", + "{ 0.0014786f, -1.5822506f, -1.4512368f, 0.1379846f, -0.0946356f, 1.0877693f, -1.4706703f, 0.5806997f, 0.1940563f, -1.1956979f, 0.9592837f, -0.8354632f, 0.2176267f, -0.7358339f, -0.0474411f, 0.2848974f, -0.6754610f, -0.9115687f, 2.1276686f, -1.0713644f, -0.0001056f, 1.8921490f, -0.0347501f, 0.0047789f, 1.3341833f, 1.8055003f, 1.1767987f, -3.9586561f, -0.6598469f, -0.3037908f, -0.1673333f, 0.2427263f },\n", + "{ -0.2013803f, 0.1559301f, 0.2713688f, -0.0187786f, 0.0075337f, -0.0453327f, 0.0510727f, -0.2533994f, -0.0093928f, 0.0939973f, 0.0683563f, 0.0257523f, -0.1638049f, 0.2167263f, 0.0139614f, -0.0689526f, -0.2007709f, 0.8788205f, 0.1043992f, 0.0529033f, 0.0041733f, -0.2248188f, 0.0029659f, 0.0044919f, 0.1728916f, -1.2823037f, 0.0284686f, -0.0879781f, 0.6332331f, 0.0599467f, -0.2467749f, 0.7796255f },\n", + "{ 0.1541705f, -2.8967228f, 0.2088300f, 0.0289306f, -0.1897649f, -0.1835614f, 0.1872510f, -0.5846522f, -0.0145777f, 2.2226386f, 0.0885817f, 0.0293056f, -0.0056043f, 2.9454181f, -0.0623621f, 0.0230481f, -0.8140234f, -4.3990140f, -0.2562745f, 0.6827632f, -4.7245188f, 0.1150251f, 0.0615204f, 4.7553473f, 0.1170709f, 0.0822542f, -0.6365855f, 0.3014538f, -2.6740234f, 0.1919117f, -0.0003937f, -0.0227543f },\n", + "};\n", + "\n", + "ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_layer2[32] = {\n", + "0.5612384f, -0.0400684f, -0.1890610f, -0.8038151f, -0.0184527f, -0.8351601f, -1.7656847f, -0.4417709f, -0.3462953f, 1.0924704f, 0.1019267f, 0.0497286f, 0.3448936f, -0.0442495f, -0.1294845f, -0.1740453f, -0.6256254f, 1.0588725f, 0.1306455f, 0.0451779f, 1.2896419f, -0.0145429f, 0.2775581f, -0.6205941f, 0.0369313f, -0.0632537f, -0.1257888f, -0.2130138f, 0.0593540f, 0.8294140f, -0.5174863f, -0.0208223f };\n", + "\n", + "ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_layer2[32][32] = {\n", + "{ 0.0802436f, -0.1068745f, -0.1320603f, 0.0222937f, 0.0260173f, -0.0999878f, 0.0045460f, 0.0257662f, -0.1026595f, 0.2246336f, -0.0252555f, 0.0313543f, -0.2062458f, 0.1684278f, -0.0972477f, -0.2176001f, -0.1525148f, -0.0692302f, 0.1213232f, -0.1288088f, -0.1729289f, -0.0895176f, -0.0199082f, -0.0172326f, 0.0126052f, 0.0598741f, 0.0853796f, -0.1326686f, 0.0753646f, -0.0108036f, -0.1034372f, -0.0026847f },\n", + "{ -0.5740064f, -0.1143318f, 0.1388070f, 1.3980483f, 0.0779263f, -1.2151324f, 2.2568495f, -0.5645000f, 2.2761426f, -0.2111119f, -0.3852313f, -1.2643801f, -0.8515005f, -0.1352748f, -0.1276971f, -0.9761337f, -1.1221037f, -0.1373484f, -6.1303482f, -0.6726473f, -3.5227783f, 0.3276748f, -0.8825266f, -0.0658490f, -0.1995521f, 0.4330961f, -0.0200665f, 0.1250316f, -0.1700243f, 1.1084150f, 1.6254629f, 0.3429218f },\n", + "{ 1.1285430f, -0.2066509f, -0.0835447f, 1.6354681f, -0.1452742f, -0.4063630f, 0.9329628f, -0.3926201f, 0.1437214f, 0.2107810f, 1.6312121f, 1.3499900f, 0.3157536f, -0.0890818f, 0.1018713f, 0.1947485f, 0.2054133f, -0.2302379f, -0.1797780f, -0.5907550f, 0.9738173f, 0.6502790f, 0.2307118f, -0.2878186f, 0.7077893f, -0.2159140f, -0.0506834f, -0.1203295f, 0.0040051f, 0.9362236f, -0.8645781f, -5.5910249f },\n", + "{ 0.1194026f, -0.0770950f, 0.0977710f, -0.0434290f, -0.0754589f, -0.0162118f, 0.1764810f, -0.0389150f, 0.1742229f, -0.0958543f, -0.1752356f, -0.0456456f, 0.1090995f, -0.0889283f, 0.0165932f, 0.0426875f, 0.0146571f, 0.0522842f, -0.0530897f, 0.0810161f, 0.0913054f, 0.0682782f, 0.0375430f, 0.0441132f, 0.0616512f, 0.1282314f, -0.0735286f, 0.0566166f, 0.1289222f, 0.0390186f, 0.0632017f, 0.0279914f },\n", + "{ -0.1397050f, 0.1469029f, -0.1035157f, -0.1537557f, 0.0723362f, 0.1360317f, 0.1632015f, -0.0519030f, 0.1664423f, -0.1674401f, 0.1221813f, 0.1218546f, 0.0581016f, -0.0302506f, 0.1745771f, 0.0057055f, -0.0728391f, -0.1049056f, -0.1592961f, 0.0329875f, 0.0789358f, -0.0462632f, 0.0748179f, 0.0180050f, 0.0803397f, -0.1458464f, -0.1164891f, -0.0082622f, 0.0343072f, 0.0366396f, -0.1715891f, -0.1089048f },\n", + "{ 0.5968121f, -0.1619606f, -0.1518290f, 0.7109355f, -0.2739125f, -1.2199029f, -0.0848204f, -0.9633313f, 1.5520710f, 0.9341283f, 0.2848818f, 1.7079298f, -0.0926562f, 0.0886664f, 0.0828461f, 2.1451118f, 1.2788274f, -0.1305077f, -0.1538649f, 0.6987577f, -1.6689410f, -0.0606114f, 0.8084964f, -0.4838576f, 0.0563342f, 0.7853296f, -0.0552194f, -0.1284064f, -0.1123770f, 1.0149344f, 0.1754804f, 0.2576492f },\n", + "{ -2.3144670f, -0.0794675f, -0.1847414f, -0.1791797f, -0.0605982f, -3.8482890f, 0.9826989f, -1.9203123f, -0.8249935f, -0.6718075f, -1.6555610f, 2.0798690f, 1.5665153f, -0.2050105f, -0.1993937f, 1.1070000f, 0.0836252f, 1.6307852f, 0.2448520f, -2.1885965f, -0.2890749f, 2.9986718f, 2.8526762f, -0.7994567f, -0.3042918f, 0.0960785f, 0.0377693f, -0.0463109f, -0.0870433f, -0.8293707f, -1.3597184f, 1.3583397f },\n", + "{ -0.3422491f, -0.0783274f, -0.1543446f, 0.0939283f, -0.3309467f, 0.6933140f, 0.6562154f, -0.6217518f, 0.7983661f, -0.1371530f, 0.5118276f, 0.7320337f, 0.5217202f, -0.0545936f, -0.1052059f, -0.1444394f, 0.3567903f, 1.2509772f, -0.6311634f, -0.7454629f, -1.0077031f, 0.8453025f, -0.2257671f, -0.2347097f, -0.3497045f, -0.1478627f, -0.0333489f, -0.2663794f, -0.0317280f, -0.3712497f, 0.8136677f, -2.5842018f },\n", + "{ 0.0997253f, 0.0620212f, 0.0460688f, 0.1300038f, 0.1323504f, -0.1669361f, 0.0732264f, -0.0860083f, -0.0762404f, -0.1628099f, -0.0881021f, 0.1323460f, -0.0460503f, 0.1401906f, 0.0602387f, 0.1474468f, 0.0935555f, -0.1531849f, 0.1754870f, 0.1422898f, 0.1418013f, -0.0302381f, -0.1058683f, -0.0258421f, -0.0110884f, 0.1016085f, 0.0643939f, -0.0729882f, -0.0572575f, -0.0603358f, 0.0267279f, 0.0611828f },\n", + "{ -0.6703482f, -0.0389322f, -0.1896956f, 1.0978996f, -0.1892595f, -2.0757272f, 1.7673731f, -0.7337908f, 2.1260054f, -0.6923556f, -0.6350248f, -1.8122731f, -1.2794832f, 0.0727381f, -0.1702943f, -1.0052446f, -1.1308233f, -0.1550428f, -5.3850775f, -0.5972114f, -0.3461530f, -0.4958812f, -1.2339170f, 0.0774786f, -0.8347382f, 0.2885106f, -0.1653138f, -0.0159324f, -0.1990542f, 1.6405339f, 1.7268503f, 0.8528640f },\n", + "{ 0.0038449f, -0.2388104f, -0.1149939f, -1.2830256f, -0.0052918f, 0.3492746f, 0.3040406f, 0.6817812f, 0.8697712f, 0.4297135f, -1.2266355f, -0.6838435f, 0.4784633f, -0.2080981f, 0.1026195f, 0.4500216f, -0.4817615f, -1.5641849f, 1.6826199f, -0.6830738f, 0.3725868f, 0.4554783f, -0.0576175f, 1.0322572f, 1.9159367f, -0.3336473f, 0.1027016f, -0.2515334f, -0.1622610f, -1.8702512f, -0.7941946f, -2.9838107f },\n", + "{ -0.2906766f, -0.0472568f, -0.2609593f, -1.5798510f, -0.1301123f, -1.1353090f, -2.5198724f, 1.5027611f, 1.2625716f, 3.2891662f, -2.3402910f, -0.0245398f, -9.8846655f, -0.2448200f, -0.0981539f, -0.2132508f, 0.7027491f, -3.4207478f, 1.3422097f, 4.3238688f, -1.1800685f, -2.7913725f, 1.8557802f, 5.5698090f, 0.2008359f, 0.2571939f, -0.0491005f, -0.1192926f, -0.0141392f, 0.1872108f, -0.8192848f, 0.6364858f },\n", + "{ 0.0671812f, -0.0234023f, 0.1400131f, 0.0778011f, 0.1308578f, -0.1675161f, 0.0237332f, 0.0215410f, 0.1514422f, 0.0736446f, 0.0181612f, -0.0219220f, -0.0099684f, 0.0102909f, 0.1243076f, 0.0897413f, -0.0682666f, 0.0389046f, 0.1245468f, 0.0098897f, -0.0425716f, 0.1595688f, 0.0397469f, -0.0664724f, -0.1641733f, 0.0605745f, 0.1712325f, 0.1596854f, 0.0224220f, -0.1328268f, -0.1743169f, 0.0608113f },\n", + "{ -1.1302176f, -0.0267920f, -0.2404846f, 0.6629997f, 0.1574058f, 4.7085104f, 0.5321816f, 0.4496275f, -0.2375129f, 1.2458097f, 1.8413728f, -1.8281459f, 2.3066695f, -0.2194766f, -0.0977243f, -0.4933192f, 0.5315795f, -0.8266373f, -1.6038543f, -0.6524141f, -0.2899630f, -4.3681431f, 0.2379541f, -0.3776518f, 1.5330435f, 0.3396196f, -0.0837295f, -0.0557159f, 0.1054505f, 1.1391810f, 0.6029354f, 0.9084895f },\n", + "{ -0.0898137f, 0.0973598f, -0.1583033f, -0.1445000f, 0.0044857f, -0.1162652f, -0.1396598f, 0.1172944f, 0.0714815f, -0.1043701f, 0.0497041f, 0.0381792f, 0.1395592f, -0.0310079f, -0.1161861f, -0.0776435f, 0.0771079f, -0.0311096f, -0.1498033f, 0.0838249f, 0.1343088f, 0.1152903f, 0.0749620f, -0.1248982f, -0.0346379f, -0.0780947f, -0.0730843f, 0.1654648f, -0.1482577f, -0.0118278f, 0.1078758f, -0.1372479f },\n", + "{ -0.7470447f, -0.0269854f, 0.0793572f, -0.7553226f, -0.0888210f, 1.8323143f, 0.2497575f, -10.9856701f, -0.7587091f, 1.1695130f, -0.0198075f, -0.8016124f, -0.2920889f, -0.1624553f, -0.0925370f, -1.3623806f, -1.3067284f, 0.2490501f, 0.0292220f, -0.2309522f, 0.0131977f, -0.7858045f, 0.1076498f, 0.6439033f, -0.1686209f, -0.8376603f, 0.0470587f, 0.0656591f, 0.0300643f, 1.4373403f, 0.2995886f, -1.3597747f },\n", + "{ -0.8746193f, -0.0201447f, -0.1603909f, -0.9373384f, 0.0254792f, 0.7993639f, 0.1087815f, -0.4859151f, -1.9559643f, -2.9621441f, -1.7735658f, -2.2227323f, -0.3725011f, 0.0284650f, -0.2266676f, -0.7529051f, -0.4477961f, 0.2071678f, 2.2770953f, 1.1170574f, -0.7023426f, 0.6896675f, -1.2416189f, -1.0012786f, -1.8752691f, -0.0359559f, -0.0125555f, -0.0457818f, -0.0775177f, 0.6092747f, 0.6639680f, 1.0951738f },\n", + "{ -1.7250717f, -0.0309622f, 0.1054536f, 2.0844676f, -0.2957073f, -0.0859962f, -1.5239947f, -0.2195731f, -1.5450290f, 0.4916542f, -1.3940116f, -0.5043938f, -0.7850562f, -0.1393226f, 0.0421535f, -0.2805571f, 0.5149463f, -0.4160199f, 1.3064783f, 0.3980424f, 0.1537250f, 0.2291165f, -1.5061877f, -0.0060016f, 0.1236818f, -0.1598873f, 0.0182953f, -0.0649871f, -0.1681922f, -0.8940877f, -0.6040494f, -0.5640311f },\n", + "{ 0.1721584f, -0.0760296f, 0.0164177f, 1.6552234f, -0.2948481f, -0.8678465f, 0.9449416f, -2.5188146f, -0.3830298f, -0.9600880f, 1.3968240f, 0.4318309f, 1.2421557f, -0.2275233f, -0.2093997f, -0.4352153f, -0.2142241f, -1.6719555f, -1.5515612f, -1.5296252f, 0.6067497f, 2.3952096f, -1.4322679f, -1.7078539f, -0.0283693f, -0.0217880f, -0.0268028f, -0.0223932f, -0.0441584f, 0.2260311f, -0.3117388f, 0.0257804f },\n", + "{ -0.3736052f, -0.1436419f, -0.1791924f, 0.6735403f, -0.1233307f, 0.9371016f, -0.5032559f, 0.0065924f, 0.2331814f, 0.5494700f, 1.4163370f, 0.7603047f, 1.8672758f, -0.1058595f, -0.0214494f, -0.1112118f, 0.4055682f, -0.2838995f, -1.8654461f, -0.6464235f, 0.3173830f, 0.1275925f, -0.0674378f, -0.1359007f, 0.8348094f, 0.2912215f, -0.1404384f, -0.2908648f, -0.1660577f, 0.4171410f, -0.5499387f, -0.9420084f },\n", + "{ -10.1053333f, -0.1126838f, 0.0914457f, 2.1485925f, -0.2125181f, -0.6991695f, -1.5648814f, 3.5901635f, 0.4265155f, 2.2461650f, 1.5659872f, -4.5079947f, -1.8503289f, 0.0036438f, -0.0552862f, 6.9531665f, 7.5919185f, -5.2455883f, 1.9158086f, -0.4571260f, -8.5806217f, 3.1702795f, 4.6932554f, 0.9312032f, -0.2189541f, -35.8472443f, -0.0437394f, 0.0193345f, -0.0759754f, -0.8393922f, -8.7812824f, 1.2664827f },\n", + "{ -0.4285544f, -0.0112712f, -0.1986143f, -0.6809276f, -0.2643681f, 0.8245937f, -0.1275320f, 1.8360085f, -1.0321223f, 1.0097107f, -0.4381949f, -0.0334159f, -0.1864567f, 0.0119804f, -0.1997918f, -0.3271705f, -0.0386158f, 1.0598557f, 0.9783336f, 1.9556490f, -0.2163550f, -2.8529193f, 0.6785525f, 1.3684491f, 1.2711524f, -0.1338721f, -0.0004762f, -0.0270106f, -0.1851630f, 0.1504625f, 0.1159483f, 1.3600112f },\n", + "{ 0.1121258f, 0.0640089f, -0.1267886f, 0.0409490f, -0.1376833f, -0.0064835f, 0.1204074f, -0.0622866f, 0.1238085f, 0.0468351f, -0.0761505f, 0.0465574f, 0.0356306f, -0.0866522f, 0.1416768f, 0.1586308f, 0.1173862f, -0.0805042f, 0.1809647f, -0.1865381f, -0.1134762f, -0.1078163f, 0.1637889f, 0.3176169f, 0.1323416f, 0.0248524f, -0.0535714f, 0.1136851f, 0.0925486f, 0.0475878f, -0.0636055f, -0.0584198f },\n", + "{ -14.6186295f, -0.1993816f, 0.1559632f, 2.3735945f, -0.0626924f, -8.2991180f, 0.2650338f, 3.5373747f, 1.6012716f, 2.8416808f, 1.1672604f, -4.3095527f, -1.1607343f, 0.0817207f, 0.1491408f, 7.4242001f, 7.2067833f, -4.3056359f, 0.8474119f, -0.9166479f, 0.6955321f, 2.0489492f, 4.0285969f, 0.0758955f, -1.0957255f, -38.1951637f, -0.0826333f, -0.1540980f, -0.2915384f, 1.0447723f, -9.4216700f, 0.7907819f },\n", + "{ 1.0854393f, -0.2057757f, -0.2333630f, -2.8073475f, -0.0092103f, -0.0027702f, 0.0400727f, 1.1957222f, 2.1408370f, 2.4174137f, -4.4619761f, -2.1267519f, -2.3692110f, -0.0215195f, 0.0940003f, 1.5145874f, 0.4090779f, 1.0525370f, 0.7181456f, 2.2475765f, -4.0287180f, 0.2974668f, 1.9458562f, 0.4356304f, -1.8651968f, 0.1091178f, -0.0625651f, 0.0101182f, -0.1866876f, 2.0775406f, -0.0924411f, -0.9942529f },\n", + "{ 0.6088188f, -0.1557965f, -0.1098123f, -2.0238423f, -0.0628158f, 1.8817046f, -2.8116503f, -1.9974719f, -2.8561919f, -1.5064508f, -0.5036830f, 0.0460033f, -1.2808908f, -0.2553190f, -0.2242925f, 0.3286708f, 0.6866547f, -0.8440441f, -3.4105279f, 1.5247352f, 0.5397177f, -0.5575525f, -0.4773433f, -0.8334278f, -1.8625259f, -0.5315101f, -0.1528733f, -0.1654517f, -0.1824549f, -0.9775680f, -2.9137912f, -0.4873172f },\n", + "{ -3.4014626f, -0.0843555f, 0.0550500f, 1.7411346f, -0.1923270f, 0.9657442f, -2.5730155f, -0.4918169f, 0.7225664f, -1.8889602f, 3.7809584f, 1.7930225f, -1.6396630f, -0.2156895f, -0.1968789f, -0.4836586f, -1.0472771f, 3.1005175f, -3.8917956f, -4.2640653f, 0.9576553f, -0.5442499f, -0.7966165f, -7.2177167f, 3.5426965f, 0.3307767f, -0.1710942f, -0.1729530f, -0.0668134f, -0.4916894f, 1.9444537f, 0.5960958f },\n", + "{ -0.9639474f, -0.0439117f, 0.0067572f, -0.8821743f, 0.1859997f, -0.7608641f, 1.5826563f, -1.4137437f, -0.3524639f, 0.4026311f, -0.0760814f, 0.7603636f, 2.5665376f, 0.0416676f, -0.1360063f, -0.6073055f, -1.3086349f, 1.9021115f, 1.3700111f, -0.9890064f, -0.1628284f, -1.3746781f, 0.2909067f, 0.3049279f, 0.8926148f, -0.4370743f, -0.0742112f, -0.0174381f, 0.0485841f, -0.0802396f, 3.5397923f, -0.6531604f },\n", + "{ 0.0491318f, -0.0591713f, 0.0498613f, 1.1349666f, -0.0889358f, -1.5782114f, 0.4471801f, 1.2610462f, -0.4626705f, 0.9874090f, 1.3419105f, 0.9187180f, 0.3291495f, -0.0520998f, -0.0861611f, 0.0016949f, -0.3544937f, -0.4579405f, -0.3237021f, -0.2511895f, -3.9122045f, -0.4933105f, 1.1840253f, 0.0769721f, 1.6942110f, -0.0282394f, -0.0797413f, -0.0965752f, -0.1654065f, 0.6972140f, 0.0176173f, 0.2189070f },\n", + "{ -1.1536870f, -0.1555014f, -0.3485503f, -1.5355002f, -0.2638741f, 1.3639680f, 1.6288967f, 0.6296268f, -0.5865426f, -0.5947766f, -1.4163440f, -1.5454646f, -0.7648816f, -0.1768368f, -0.1375362f, 0.3310686f, -0.2621009f, 0.6404898f, 1.0589499f, 0.2490305f, 0.3255049f, -0.1451550f, -0.3024975f, 0.8173084f, -2.1893589f, -0.0099213f, -0.1596652f, 0.1068924f, 0.0260035f, 1.4544648f, -0.7779374f, -0.7673930f },\n", + "{ 0.0435760f, -0.0246287f, -0.1238125f, -0.0320143f, 0.1576336f, 0.0702628f, -0.1744860f, -0.1600130f, -0.0870369f, 0.0819791f, 0.0445500f, 0.0019862f, 0.1529912f, 0.0886548f, 0.0478378f, -0.0487068f, 0.0401030f, 0.0869702f, 0.1359502f, 0.1713400f, -0.1529109f, 0.1224933f, 0.0072967f, -0.0784628f, 0.1494704f, -0.1378658f, -0.0747428f, -0.0850826f, -0.0940957f, 0.0511818f, 0.1520686f, -0.0680635f },\n", + "{ 1.3802143f, 0.0130436f, -0.0118297f, 0.9667702f, -0.4034582f, 0.3238442f, 0.4695910f, -0.3288718f, 1.1822759f, 1.5390285f, 0.2593471f, -0.0801804f, 0.4593041f, 0.0215906f, -0.2340092f, 0.4232760f, -0.0674659f, 0.7093915f, -0.0001982f, 0.2270681f, -0.8337699f, -0.2296698f, 0.9161285f, 0.3906515f, -0.1553026f, 0.2775815f, -0.0620411f, -0.1116518f, -0.0008762f, -0.3016884f, 1.4699873f, 0.6260686f },\n", + "};\n", + "\n", + "ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_output_layer[1] = {\n", + "0.7275639f };\n", + "\n", + "ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_output_layer[32][1] = {\n", + "{ 1.4243358f },\n", + "{ 0.0335807f },\n", + "{ 0.0551641f },\n", + "{ -0.9836086f },\n", + "{ -0.0249541f },\n", + "{ -1.5375688f },\n", + "{ -0.7714168f },\n", + "{ -0.9649364f },\n", + "{ -1.1769278f },\n", + "{ 1.3249911f },\n", + "{ -1.6541473f },\n", + "{ 1.4079021f },\n", + "{ -0.8831168f },\n", + "{ 0.0122874f },\n", + "{ 0.0511134f },\n", + "{ -2.6734750f },\n", + "{ 2.8394303f },\n", + "{ 0.9675560f },\n", + "{ -1.4186903f },\n", + "{ -2.0796514f },\n", + "{ -1.7693948f },\n", + "{ -0.8502544f },\n", + "{ -1.5927037f },\n", + "{ -1.1028550f },\n", + "{ 0.8137528f },\n", + "{ 6.3073616f },\n", + "{ 0.1059108f },\n", + "{ -0.0468376f },\n", + "{ 0.1322162f },\n", + "{ 0.7481517f },\n", + "{ -1.2260461f },\n", + "{ -0.9095332f },\n", + "};\n", + "\n" + ] + } + ], + "source": [ + "def print_formatted_weights_biases(weights, biases, layer_name):\n", + " # Print biases\n", + " print(f\"ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_{layer_name}[{len(biases)}] = {{\")\n", + " print(\", \".join(f\"{b:.7f}f\" for b in biases) + \" };\")\n", + " print()\n", + "\n", + " # Print weights\n", + " print(f\"ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_{layer_name}[{len(weights[0])}][{len(weights)}] = {{\")\n", + " for row in weights.T:\n", + " formatted_row = \", \".join(f\"{w:.7f}f\" for w in row)\n", + " print(f\"{{ {formatted_row} }},\")\n", + " print(\"};\")\n", + " print()\n", + "\n", + "def print_model_weights_biases(model):\n", + " # Make sure the model is in evaluation mode\n", + " model.eval()\n", + "\n", + " # Iterate through all named modules in the model\n", + " for name, module in model.named_modules():\n", + " # Check if the module is a linear layer\n", + " if isinstance(module, nn.Linear):\n", + " # Get weights and biases\n", + " weights = module.weight.data.cpu().numpy()\n", + " biases = module.bias.data.cpu().numpy()\n", + "\n", + " # Print formatted weights and biases\n", + " print_formatted_weights_biases(weights, biases, name.replace('.', '_'))\n", + "\n", + "print_model_weights_biases(model)\n" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": {}, + "outputs": [], + "source": [ + "# Ensure input_features_tensor is moved to the appropriate device\n", + "input_features_tensor = input_features_tensor.to(device)\n", + "\n", + "# Make predictions\n", + "with torch.no_grad():\n", + " model.eval()\n", + " outputs = model(input_features_tensor)\n", + " predictions = outputs.squeeze().cpu().numpy()\n", + "\n", + "full_tracks = (np.concatenate(branches['t3_pMatched']) > 0.95)" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "pt: 0 to inf\n", + "93% Retention Cut: {0.3115, 0.3408, 0.4742, 0.569, 0.5812, 0.6033, 0.7483, 0.7834, 0.8547, 0.8994} Mean: 0.6166\n", + "98% Retention Cut: {0.0642, 0.0673, 0.1201, 0.1563, 0.2093, 0.2102, 0.3688, 0.4391, 0.6223, 0.7282} Mean: 0.2986\n", + "99% Retention Cut: {0.024, 0.0267, 0.052, 0.0658, 0.093, 0.0968, 0.1913, 0.2443, 0.4012, 0.5449} Mean: 0.174\n" + ] + } + ], + "source": [ + "import numpy as np\n", + "from matplotlib import pyplot as plt\n", + "from matplotlib.colors import LogNorm\n", + "\n", + "def plot_for_pt_bin(pt_min, pt_max, percentiles, eta_bin_edges, eta_list, predictions, full_tracks, branches):\n", + " \"\"\"\n", + " Calculate and plot cut values for specified percentiles in a given pt bin\n", + " \n", + " Parameters:\n", + " -----------\n", + " pt_min : float\n", + " Minimum pt value for the bin\n", + " pt_max : float\n", + " Maximum pt value for the bin\n", + " percentiles : list\n", + " List of percentiles to calculate (e.g., [92.5, 96.7, 99])\n", + " eta_bin_edges : array\n", + " Edges of the eta bins\n", + " eta_list : list\n", + " List of eta values\n", + " predictions : array\n", + " Array of DNN predictions\n", + " full_tracks : array\n", + " Boolean array for track selection\n", + " branches : dict\n", + " Dictionary containing branch data\n", + " \"\"\"\n", + " # Filter data based on pt bin\n", + " abs_eta = eta_list[0][full_tracks]\n", + " predictions_filtered = predictions[full_tracks]\n", + " \n", + " # Dictionary to store cut values for different percentiles\n", + " cut_values = {p: [] for p in percentiles}\n", + "\n", + " # Loop through each eta bin\n", + " for i in range(len(eta_bin_edges) - 1):\n", + " # Get indices of tracks within the current eta bin\n", + " bin_indices = (abs_eta >= eta_bin_edges[i]) & (abs_eta < eta_bin_edges[i + 1])\n", + " \n", + " # Get the corresponding DNN prediction scores\n", + " bin_predictions = predictions_filtered[bin_indices]\n", + " \n", + " # Calculate the percentile cut values for the current bin\n", + " for percentile in percentiles:\n", + " cut_value = np.percentile(bin_predictions, 100 - percentile) # Convert retention to percentile\n", + " cut_values[percentile].append(cut_value)\n", + "\n", + " # Plot 2D histogram\n", + " plt.figure(figsize=(10, 6))\n", + " plt.hist2d(abs_eta, predictions_filtered, bins=[eta_bin_edges, 50], norm=LogNorm())\n", + " plt.colorbar(label='Counts')\n", + " plt.xlabel(\"Absolute Eta\")\n", + " plt.ylabel(\"DNN Prediction Score\")\n", + " plt.title(f\"DNN Score vs. Abs Eta for 100% Matched Tracks (pt: {pt_min} to {pt_max})\")\n", + "\n", + " # Plot the cut values with different colors\n", + " cut_x = eta_bin_edges[:-1] + (eta_bin_edges[1] - eta_bin_edges[0]) / 2 # Mid-points of the bins\n", + " colors = plt.cm.rainbow(np.linspace(0, 1, len(percentiles))) # Generate distinct colors\n", + " \n", + " for percentile, color in zip(percentiles, colors):\n", + " plt.plot(cut_x, cut_values[percentile], '-', color=color, marker='o', \n", + " label=f'{percentile}% Retention Cut')\n", + " \n", + " plt.legend()\n", + " plt.grid(True, alpha=0.3)\n", + " plt.show()\n", + " \n", + " # Print the cut values\n", + " print(f\"\\npt: {pt_min} to {pt_max}\")\n", + " for percentile in percentiles:\n", + " values = cut_values[percentile]\n", + " print(f\"{percentile}% Retention Cut:\", \n", + " '{' + ', '.join(str(x) for x in np.round(values, 4)) + '}',\n", + " \"Mean:\", np.round(np.mean(values), 4))\n", + "\n", + "# Example usage:\n", + "def analyze_pt_bins(pt_bins, percentiles, eta_bin_edges, eta_list, predictions, full_tracks, branches):\n", + " \"\"\"\n", + " Analyze and plot for multiple pt bins and percentiles\n", + " \n", + " Parameters:\n", + " -----------\n", + " pt_bins : list\n", + " List of pt bin edges\n", + " percentiles : list\n", + " List of percentiles to calculate\n", + " Other parameters same as plot_for_pt_bin function\n", + " \"\"\"\n", + " for i in range(len(pt_bins) - 1):\n", + " plot_for_pt_bin(pt_bins[i], pt_bins[i + 1], percentiles, eta_bin_edges, \n", + " eta_list, predictions, full_tracks, branches)\n", + "\n", + "# Example call:\n", + "percentiles = [93, 98, 99]\n", + "pt_bins = [0, np.inf]\n", + "eta_bin_edges = np.arange(0, 2.75, 0.25)\n", + "analyze_pt_bins(pt_bins, percentiles, eta_bin_edges, eta_list, predictions, full_tracks, branches)" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "(2707725,)" + ] + }, + "execution_count": 12, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "np.shape(predictions[predictions > 0.215])" + ] + }, + { + "cell_type": "code", + "execution_count": 13, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "(15131951,)" + ] + }, + "execution_count": 13, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "np.shape(predictions)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "analysisenv", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.11.7" + } + }, + "nbformat": 4, + "nbformat_minor": 2 +} diff --git a/RecoTracker/LSTCore/standalone/code/core/AccessHelper.cc b/RecoTracker/LSTCore/standalone/code/core/AccessHelper.cc index ed0ff994f5941..b9327b738760f 100644 --- a/RecoTracker/LSTCore/standalone/code/core/AccessHelper.cc +++ b/RecoTracker/LSTCore/standalone/code/core/AccessHelper.cc @@ -207,10 +207,26 @@ std::vector getModuleIdxsFromT5(LSTEvent* event, unsigned int T5) } return module_idxs; } + +//____________________________________________________________________________________________ +std::vector getModuleIdxsFromT3(LSTEvent* event, unsigned int T3) { + std::vector hits = getHitsFromT3(event, T3); + std::vector module_idxs; + auto hitsEvt = event->getHits(); + for (auto& hitIdx : hits) { + module_idxs.push_back(hitsEvt.moduleIndices()[hitIdx]); + } + return module_idxs; +} + //____________________________________________________________________________________________ std::vector getHitTypesFromT5(LSTEvent* event, unsigned int T5) { return {4, 4, 4, 4, 4, 4, 4, 4, 4, 4}; - ; +} + +//____________________________________________________________________________________________ +std::vector getHitTypesFromT3(LSTEvent* event, unsigned int T5) { + return {4, 4, 4, 4, 4, 4}; } //____________________________________________________________________________________________ diff --git a/RecoTracker/LSTCore/standalone/code/core/AccessHelper.h b/RecoTracker/LSTCore/standalone/code/core/AccessHelper.h index dbec129a60562..5790f3131fc3d 100644 --- a/RecoTracker/LSTCore/standalone/code/core/AccessHelper.h +++ b/RecoTracker/LSTCore/standalone/code/core/AccessHelper.h @@ -33,6 +33,8 @@ std::tuple, std::vector> getHitIdxsAndHi std::vector getLSsFromT3(LSTEvent* event, unsigned int T3); std::vector getMDsFromT3(LSTEvent* event, unsigned int T3); std::vector getHitsFromT3(LSTEvent* event, unsigned int T3); +std::vector getHitTypesFromT3(LSTEvent* event, unsigned int T3); +std::vector getModuleIdxsFromT3(LSTEvent* event, unsigned int T3); std::tuple, std::vector> getHitIdxsAndHitTypesFromT3(LSTEvent* event, unsigned T3); diff --git a/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.cc b/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.cc index deed88f833a00..b7d325d75d106 100644 --- a/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.cc +++ b/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.cc @@ -161,6 +161,7 @@ void createOptionalOutputBranches() { // T5 DNN branches createT5DNNBranches(); + createT3DNNBranches(); #endif } @@ -192,6 +193,39 @@ void createT5DNNBranches() { } } + +//________________________________________________________________________________________________________________________________ +void createT3DNNBranches() { + // Common branches for T3 properties based on TripletsSoA fields + ana.tx->createBranch>("t3_betaIn"); + ana.tx->createBranch>("t3_centerX"); + ana.tx->createBranch>("t3_centerY"); + ana.tx->createBranch>("t3_radius"); + ana.tx->createBranch>("t3_partOfPT5"); + ana.tx->createBranch>("t3_partOfT5"); + ana.tx->createBranch>("t3_partOfPT3"); + ana.tx->createBranch>("t3_pMatched"); + + // Hit-specific branches (T3 has 4 hits from two segments) + std::vector hitIndices = {"0", "1", "2", "3", "4", "5"}; + std::vector hitProperties = {"r", "x", "y", "z", "eta", "phi", "detId", "layer", "moduleType"}; + + for (const auto& idx : hitIndices) { + for (const auto& prop : hitProperties) { + std::string branchName = "t3_hit_" + idx + "_" + prop; + if (prop == "detId" || prop == "layer" || prop == "moduleType") { + ana.tx->createBranch>(branchName); + } else { + ana.tx->createBranch>(branchName); + } + } + } + + // Additional metadata branches + ana.tx->createBranch>("t3_layer_binary"); + ana.tx->createBranch>>("t3_matched_simIdx"); +} + //________________________________________________________________________________________________________________________________ void createGnnNtupleBranches() { // Mini Doublets @@ -340,6 +374,7 @@ void setOptionalOutputBranches(LSTEvent* event) { setQuintupletOutputBranches(event); setPixelTripletOutputBranches(event); setOccupancyBranches(event); + setT3DNNBranches(event); setT5DNNBranches(event); #endif @@ -638,6 +673,42 @@ void setPixelTripletOutputBranches(LSTEvent* event) { ana.tx->setBranch>("pT3_isDuplicate", pT3_isDuplicate); } +//________________________________________________________________________________________________________________________________ +void fillT3DNNBranches(LSTEvent* event, unsigned int iT3) { + auto hits = event->getHits(); + auto modules = event->getModules(); + + std::vector hitIdx = getHitsFromT3(event, iT3); + std::vector hitObjects; + + for (int i = 0; i < hitIdx.size(); ++i) { + unsigned int hit = hitIdx[i]; + float x = hits.xs()[hit]; + float y = hits.ys()[hit]; + float z = hits.zs()[hit]; + lst_math::Hit hitObj(x, y, z); + hitObjects.push_back(hitObj); + + std::string idx = std::to_string(i); + ana.tx->pushbackToBranch("t3_hit_" + idx + "_r", sqrt(x * x + y * y)); + ana.tx->pushbackToBranch("t3_hit_" + idx + "_x", x); + ana.tx->pushbackToBranch("t3_hit_" + idx + "_y", y); + ana.tx->pushbackToBranch("t3_hit_" + idx + "_z", z); + ana.tx->pushbackToBranch("t3_hit_" + idx + "_eta", hitObj.eta()); + ana.tx->pushbackToBranch("t3_hit_" + idx + "_phi", hitObj.phi()); + + int subdet = trk.ph2_subdet()[hits.idxs()[hit]]; + int is_endcap = subdet == 4; + int layer = trk.ph2_layer()[hits.idxs()[hit]] + 6 * is_endcap; + int detId = trk.ph2_detId()[hits.idxs()[hit]]; + unsigned int module = hits.moduleIndices()[hit]; + + ana.tx->pushbackToBranch("t3_hit_" + idx + "_detId", detId); + ana.tx->pushbackToBranch("t3_hit_" + idx + "_layer", layer); + ana.tx->pushbackToBranch("t3_hit_" + idx + "_moduleType", modules.moduleType()[module]); + } +} + //________________________________________________________________________________________________________________________________ void fillT5DNNBranches(LSTEvent* event, unsigned int iT3) { auto hits = event->getHits(); @@ -690,6 +761,50 @@ void fillT5DNNBranches(LSTEvent* event, unsigned int iT3) { ana.tx->pushbackToBranch("t5_t3_phi", hitObjects[0].phi()); } +void setT3DNNBranches(LSTEvent* event) { + auto const triplets = event->getTriplets(); + auto const tripletsOccupancy = event->getTriplets(); + auto modules = event->getModules(); + auto ranges = event->getRanges(); + + for (unsigned int lowerModuleIdx = 0; lowerModuleIdx < modules.nLowerModules(); ++lowerModuleIdx) { + int nTriplets = tripletsOccupancy.nTriplets()[lowerModuleIdx]; + for (unsigned int idx = 0; idx < nTriplets; idx++) { + unsigned int tripletIndex = ranges.tripletModuleIndices()[lowerModuleIdx] + idx; + + // Get hit indices and types + std::vector hit_idx = getHitsFromT3(event, tripletIndex); + std::vector hit_type = getHitTypesFromT3(event, tripletIndex); + std::vector module_idx = getModuleIdxsFromT3(event, tripletIndex); + + // Calculate layer binary representation + int layer_binary = 0; + for (size_t i = 0; i < module_idx.size(); i += 2) { + layer_binary |= (1 << (modules.layers()[module_idx[i]] + 6 * (modules.subdets()[module_idx[i]] == 4))); + } + + // Get matching information with percent matched + float percent_matched; + std::vector simidx = matchedSimTrkIdxs(hit_idx, hit_type, false, &percent_matched); + + // Fill the branches with T3-specific data + ana.tx->pushbackToBranch("t3_betaIn", triplets.betaIn()[tripletIndex]); + ana.tx->pushbackToBranch("t3_centerX", triplets.centerX()[tripletIndex]); + ana.tx->pushbackToBranch("t3_centerY", triplets.centerY()[tripletIndex]); + ana.tx->pushbackToBranch("t3_radius", triplets.radius()[tripletIndex]); + ana.tx->pushbackToBranch("t3_partOfPT5", triplets.partOfPT5()[tripletIndex]); + ana.tx->pushbackToBranch("t3_partOfT5", triplets.partOfT5()[tripletIndex]); // Fixed the typo here + ana.tx->pushbackToBranch("t3_partOfPT3", triplets.partOfPT3()[tripletIndex]); + ana.tx->pushbackToBranch("t3_layer_binary", layer_binary); + ana.tx->pushbackToBranch>("t3_matched_simIdx", simidx); + ana.tx->pushbackToBranch("t3_pMatched", percent_matched); + + // Fill hit-specific information + fillT3DNNBranches(event, tripletIndex); + } + } +} + //________________________________________________________________________________________________________________________________ void setT5DNNBranches(LSTEvent* event) { auto triplets = event->getTriplets(); diff --git a/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.h b/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.h index 5bfe439fadbb3..08cb80ad85be4 100644 --- a/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.h +++ b/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.h @@ -19,6 +19,7 @@ void createRequiredOutputBranches(); void createOptionalOutputBranches(); void createGnnNtupleBranches(); void createT5DNNBranches(); +void createT3DNNBranches(); void fillOutputBranches(LSTEvent* event); void setOutputBranches(LSTEvent* event); @@ -30,7 +31,9 @@ void setPixelTripletOutputBranches(LSTEvent* event); void setGnnNtupleBranches(LSTEvent* event); void setGnnNtupleMiniDoublet(LSTEvent* event, unsigned int MD); void fillT5DNNBranches(LSTEvent* event, unsigned int T3); +void fillT3DNNBranches(LSTEvent* event, unsigned int iT3); void setT5DNNBranches(LSTEvent* event); +void setT3DNNBranches(LSTEvent* event); std::tuple> parseTrackCandidate(LSTEvent* event, unsigned int); std::tuple, std::vector> parsepT5(LSTEvent* event, From 05442da56f839cd0352f2b6453dc1fc245c9f049 Mon Sep 17 00:00:00 2001 From: GNiendorf Date: Fri, 17 Jan 2025 12:59:58 -0500 Subject: [PATCH 2/3] add sim vertex to output --- .../standalone/code/core/write_lst_ntuple.cc | 23 ++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.cc b/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.cc index b7d325d75d106..6ae7b391e00b9 100644 --- a/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.cc +++ b/RecoTracker/LSTCore/standalone/code/core/write_lst_ntuple.cc @@ -205,6 +205,8 @@ void createT3DNNBranches() { ana.tx->createBranch>("t3_partOfT5"); ana.tx->createBranch>("t3_partOfPT3"); ana.tx->createBranch>("t3_pMatched"); + ana.tx->createBranch>("t3_sim_vxy"); + ana.tx->createBranch>("t3_sim_vz"); // Hit-specific branches (T3 has 4 hits from two segments) std::vector hitIndices = {"0", "1", "2", "3", "4", "5"}; @@ -793,12 +795,31 @@ void setT3DNNBranches(LSTEvent* event) { ana.tx->pushbackToBranch("t3_centerY", triplets.centerY()[tripletIndex]); ana.tx->pushbackToBranch("t3_radius", triplets.radius()[tripletIndex]); ana.tx->pushbackToBranch("t3_partOfPT5", triplets.partOfPT5()[tripletIndex]); - ana.tx->pushbackToBranch("t3_partOfT5", triplets.partOfT5()[tripletIndex]); // Fixed the typo here + ana.tx->pushbackToBranch("t3_partOfT5", triplets.partOfT5()[tripletIndex]); ana.tx->pushbackToBranch("t3_partOfPT3", triplets.partOfPT3()[tripletIndex]); ana.tx->pushbackToBranch("t3_layer_binary", layer_binary); ana.tx->pushbackToBranch>("t3_matched_simIdx", simidx); ana.tx->pushbackToBranch("t3_pMatched", percent_matched); + // Add vertex information for matched sim tracks + if (simidx.size() == 0) { + // No matched sim track - set default values + ana.tx->pushbackToBranch("t3_sim_vxy", 0.0); + ana.tx->pushbackToBranch("t3_sim_vz", 0.0); + } else { + // Get vertex information from the first matched sim track + int vtxidx = trk.sim_parentVtxIdx()[simidx[0]]; + float vtx_x = trk.simvtx_x()[vtxidx]; + float vtx_y = trk.simvtx_y()[vtxidx]; + float vtx_z = trk.simvtx_z()[vtxidx]; + + // Calculate transverse distance from origin + float vxy = sqrt(vtx_x * vtx_x + vtx_y * vtx_y); + + ana.tx->pushbackToBranch("t3_sim_vxy", vxy); + ana.tx->pushbackToBranch("t3_sim_vz", vtx_z); + } + // Fill hit-specific information fillT3DNNBranches(event, tripletIndex); } From 0a8572c6cb8d00cf9754f543a414cba4f88519df Mon Sep 17 00:00:00 2001 From: GNiendorf Date: Tue, 21 Jan 2025 16:12:04 -0500 Subject: [PATCH 3/3] Add displaced upweighting to t3dnn code --- RecoTracker/LSTCore/interface/alpaka/Common.h | 6 +- .../LSTCore/src/alpaka/NeuralNetwork.h | 10 +- .../src/alpaka/T3NeuralNetworkWeights.h | 162 ++-- ...workWeights.h => T5NeuralNetworkWeights.h} | 4 +- .../analysis/DNN/train_T3_DNN.ipynb | 709 ++++++++++-------- 5 files changed, 487 insertions(+), 404 deletions(-) rename RecoTracker/LSTCore/src/alpaka/{NeuralNetworkWeights.h => T5NeuralNetworkWeights.h} (99%) diff --git a/RecoTracker/LSTCore/interface/alpaka/Common.h b/RecoTracker/LSTCore/interface/alpaka/Common.h index 77ab860a46188..b03c58b175ef2 100644 --- a/RecoTracker/LSTCore/interface/alpaka/Common.h +++ b/RecoTracker/LSTCore/interface/alpaka/Common.h @@ -78,9 +78,9 @@ namespace ALPAKA_ACCELERATOR_NAMESPACE::lst { namespace t3dnn { ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kZ_max = 224.149505f; ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kR_max = 98.932365f; - // No pt binning for T3 - ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kWp[kEtaBins] = - {0.024, 0.0267, 0.052, 0.0658, 0.093, 0.0968, 0.1913, 0.2443, 0.4012, 0.5449}; + ALPAKA_STATIC_ACC_MEM_GLOBAL constexpr float kWp[kPtBins][kEtaBins] = { + {0.0387, 0.0395, 0.0599, 0.083, 0.1433, 0.163, 0.2927, 0.3243, 0.638, 0.7512}, + {0.0019, 0.0024, 0.0047, 0.0107, 0.0107, 0.0316, 0.2533, 0.1114, 0.0501, 0.0407}}; } // namespace t3dnn namespace t5dnn { diff --git a/RecoTracker/LSTCore/src/alpaka/NeuralNetwork.h b/RecoTracker/LSTCore/src/alpaka/NeuralNetwork.h index 4c81c137ee089..651f8dffec8cd 100644 --- a/RecoTracker/LSTCore/src/alpaka/NeuralNetwork.h +++ b/RecoTracker/LSTCore/src/alpaka/NeuralNetwork.h @@ -6,7 +6,7 @@ #include "RecoTracker/LSTCore/interface/alpaka/Common.h" #include "RecoTracker/LSTCore/interface/MiniDoubletsSoA.h" -#include "NeuralNetworkWeights.h" +#include "T5NeuralNetworkWeights.h" #include "T3NeuralNetworkWeights.h" namespace ALPAKA_ACCELERATOR_NAMESPACE::lst { @@ -117,10 +117,14 @@ namespace ALPAKA_ACCELERATOR_NAMESPACE::lst { linear_layer(x_2, x_3, t3dnn::wgtT_output_layer, t3dnn::bias_output_layer); float x_5 = sigmoid_activation(acc, x_3[0]); + // Get the bin index based on abs(eta) of first hit and t5_pt + float t3_pt = radius * lst::k2Rinv1GeVf * 2; + + uint8_t pt_index = (t3_pt > 5); uint8_t bin_index = (eta1 > 2.5f) ? (kEtaBins - 1) : static_cast(eta1 / 0.25f); - // Compare to cut value for relevant bin - return x_5 > kWp[bin_index]; + // Compare x_5 to the cut value for the relevant bin + return x_5 > kWp[pt_index][bin_index]; } } // namespace t3dnn diff --git a/RecoTracker/LSTCore/src/alpaka/T3NeuralNetworkWeights.h b/RecoTracker/LSTCore/src/alpaka/T3NeuralNetworkWeights.h index 59a9852fd99c7..20efefa1aa26d 100644 --- a/RecoTracker/LSTCore/src/alpaka/T3NeuralNetworkWeights.h +++ b/RecoTracker/LSTCore/src/alpaka/T3NeuralNetworkWeights.h @@ -6,99 +6,99 @@ namespace ALPAKA_ACCELERATOR_NAMESPACE::lst::t3dnn { ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_layer1[32] = { --0.1918962f, -1.0575372f, -0.8276564f, -0.0243965f, -0.1577621f, 1.0067693f, 1.5348158f, 0.4439710f, 0.0041234f, -1.1558943f, -1.4180470f, 1.0221841f, -0.0592227f, -1.2107433f, -0.2100758f, 1.2193928f, -0.3124787f, -1.9197327f, -0.8064887f, -0.2178766f, -0.0111392f, -0.1638742f, 0.0029338f, -0.0157688f, 0.2662797f, 1.8194629f, 0.8465537f, -0.7592145f, -0.8783396f, 0.5602613f, -0.0764334f, -0.8502049f }; +-0.2400621f, 1.3418640f, -0.1700544f, -0.1809052f, -0.0194816f, -1.6636838f, -0.2450859f, -1.1127867f, 0.1477622f, -0.0740084f, -0.8851718f, -0.2633939f, 0.1941925f, 0.9132360f, 1.5794699f, -0.7953343f, -0.2615424f, 0.2003026f, -2.2417912f, -0.8144404f, -1.2980548f, -0.1407361f, -0.0966190f, -0.2355828f, -0.3362100f, 0.8529679f, -0.0075295f, -0.0501248f, -0.8618563f, -0.7658819f, 0.8201296f, -0.0064448f }; ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_layer1[14][32] = { -{ -0.2258795f, 0.4783621f, 0.4048500f, -0.2535419f, -0.2204617f, 0.2292106f, 0.5185162f, 1.0356801f, -0.1940614f, 0.7601448f, -0.2762348f, 0.4394473f, -0.2472922f, 0.6392686f, 0.0174135f, 1.1337029f, 0.5831500f, -0.5536784f, 0.1608927f, 0.3617360f, -0.0014275f, 0.1839313f, -0.1858799f, 0.0080405f, 0.1321980f, 0.4460649f, 0.4296648f, 0.4569599f, -0.2347775f, -0.3548780f, 0.1986685f, 0.2211701f }, -{ 0.1509047f, 0.0057684f, -0.0019561f, -0.1210250f, -0.2612511f, -0.0098326f, 0.0019430f, 0.0152595f, -0.2313585f, 0.0117159f, -0.0087940f, -0.0000195f, 0.2194081f, -0.0033250f, 0.1478879f, 0.0097880f, -0.0041943f, 0.0362815f, -0.0197458f, -0.0063192f, -0.0004957f, -0.0104775f, -0.2756116f, -0.0049759f, -0.0015340f, -0.0111502f, 0.0034782f, -0.0100520f, 0.0124440f, 0.0076567f, -0.0263710f, 0.0381163f }, -{ 0.1448244f, 0.0803795f, 1.6524559f, 0.1457684f, -0.1350329f, -3.6343203f, -1.6798589f, -0.0142065f, 0.0835910f, 0.0497492f, 0.6378320f, -0.4540107f, -0.2349942f, -0.0257038f, -0.2397820f, -4.3412380f, 0.1279643f, 0.9793525f, -0.5925660f, -0.8705363f, 0.0134403f, -0.1410540f, -0.0032170f, 0.0107955f, -1.5223076f, -0.1977515f, -0.9901226f, -0.5315630f, 0.5766137f, 0.8363163f, -0.2219186f, -1.3835622f }, -{ -0.1057612f, 0.8005342f, -1.4052893f, -0.1196175f, 0.1360446f, 1.2852736f, -4.4616752f, 0.3914140f, -0.2356829f, 1.0064709f, 1.5389245f, -3.2000272f, 0.0828165f, 0.8944567f, -0.1592458f, 0.5353487f, 0.8051971f, 0.2057788f, 1.7512299f, 0.9215795f, -0.0036782f, 2.1131773f, -0.3204709f, -0.0044941f, 0.0865040f, 0.4481153f, -1.2256490f, 1.7857696f, -0.6445597f, -0.7477201f, 0.2316555f, -0.7155212f }, -{ 0.0901890f, -1.1808448f, 1.4875709f, -0.0262624f, -0.1323451f, -0.8950851f, 0.7974008f, -2.7021067f, 0.0281707f, -1.3674084f, -4.0485940f, 5.8682752f, -0.1891649f, 7.2682476f, 0.1300428f, -2.2845411f, -8.7903214f, 0.6268425f, -1.9901284f, -2.5285044f, 0.0066100f, 7.4012928f, -0.0831600f, 0.0095476f, 6.1734300f, 0.8520991f, -8.3464308f, 2.4261341f, 3.6403832f, 5.9824433f, -0.0999645f, -0.2824311f }, -{ -0.0335586f, 16.4833145f, 1.4688981f, -0.1789742f, 0.2243387f, 0.1425887f, 0.1262731f, -4.6265879f, -0.0683203f, -12.3650198f, 1.3732860f, -0.2777069f, -0.0603657f, 0.8076705f, 0.0482012f, -0.7798629f, -1.6209395f, -0.0720773f, -0.3990867f, -1.0641636f, -0.0139909f, -0.5711431f, -0.0209516f, 0.0946307f, 0.0096234f, 0.7337275f, 0.4466013f, 1.5696099f, -5.6807351f, -1.6136186f, -0.2170101f, -0.9905877f }, -{ -0.0965901f, 0.4846557f, -0.0652100f, -0.2353250f, -0.1215300f, -3.5360579f, -2.0285676f, -2.8374794f, -0.1559567f, 0.7083026f, -0.1030692f, -1.1805813f, 0.0706595f, -0.2951699f, -0.0989490f, -2.8418458f, 0.8427116f, -2.2046885f, -1.3581268f, 1.5271128f, -0.0045855f, 0.1438956f, -0.2343508f, -0.0425132f, 1.0763166f, 1.0795754f, 0.4891663f, 2.6029303f, -1.0363307f, 2.2133234f, -0.2510982f, 1.1560025f }, -{ 0.0074010f, -0.4152716f, 3.4255989f, -0.1664258f, 0.2028382f, 1.2840286f, 1.3538148f, -1.5683322f, 0.0429628f, -0.2171310f, 2.8748064f, -0.5141454f, 0.1099870f, -0.4614673f, -0.2515875f, -0.3862101f, 1.6164609f, -1.3735241f, -0.8896182f, 1.8310896f, 0.0188789f, -0.8612124f, -0.2143756f, 0.0385537f, -8.3583250f, 1.3280702f, -7.0806746f, -0.2001765f, 0.8213225f, -2.8566475f, 0.1479015f, 0.5155560f }, -{ 0.0511283f, -1.9676421f, 4.3773866f, -0.1648336f, 0.0505374f, -3.0326672f, 2.5867159f, 1.8804691f, -0.1412510f, -2.3595653f, 3.9877729f, -14.6059513f, -0.0444889f, 7.5017557f, -0.0684788f, 1.0775388f, -8.0584450f, -0.2013538f, -12.5508022f, 9.2100115f, -0.0011293f, -7.0289955f, 0.0521985f, 0.0155513f, -1.6320782f, -4.0694451f, 7.8784938f, 8.9188061f, 10.7590771f, -9.3559170f, -0.1319706f, -0.0336969f }, -{ -0.1288323f, 7.9389300f, -0.4022054f, 0.0851088f, -0.2593997f, -0.3426172f, 0.3062155f, 1.9866818f, -0.1748946f, -9.7667055f, -0.6524503f, 0.3385161f, 0.1884300f, 9.0051126f, -0.2027990f, 0.3781536f, -0.1085841f, -5.2012277f, -0.5169301f, 2.4512987f, -14.8858967f, 0.8265918f, 0.1330098f, 14.9286051f, 0.4946520f, 1.2087970f, -1.8545671f, 0.7865057f, -3.1178455f, 2.2735860f, 0.1769712f, 0.7800660f }, -{ -0.0041451f, -0.3216657f, 0.2317746f, -0.0770410f, -0.1181624f, -1.6830167f, 1.3180428f, 1.7278676f, -0.1011897f, -0.8716651f, 0.8045275f, 1.0895320f, 0.0222730f, -1.0280910f, 0.1943782f, -3.5416641f, -1.8209232f, -0.6835734f, 1.1706284f, -1.0224590f, -0.0421350f, 0.6345906f, -0.2030822f, -0.0540403f, 1.6704807f, 0.2792225f, 0.5046398f, 1.1982751f, 0.1087174f, -0.6976060f, 0.1717374f, 1.3827170f }, -{ 0.0014786f, -1.5822506f, -1.4512368f, 0.1379846f, -0.0946356f, 1.0877693f, -1.4706703f, 0.5806997f, 0.1940563f, -1.1956979f, 0.9592837f, -0.8354632f, 0.2176267f, -0.7358339f, -0.0474411f, 0.2848974f, -0.6754610f, -0.9115687f, 2.1276686f, -1.0713644f, -0.0001056f, 1.8921490f, -0.0347501f, 0.0047789f, 1.3341833f, 1.8055003f, 1.1767987f, -3.9586561f, -0.6598469f, -0.3037908f, -0.1673333f, 0.2427263f }, -{ -0.2013803f, 0.1559301f, 0.2713688f, -0.0187786f, 0.0075337f, -0.0453327f, 0.0510727f, -0.2533994f, -0.0093928f, 0.0939973f, 0.0683563f, 0.0257523f, -0.1638049f, 0.2167263f, 0.0139614f, -0.0689526f, -0.2007709f, 0.8788205f, 0.1043992f, 0.0529033f, 0.0041733f, -0.2248188f, 0.0029659f, 0.0044919f, 0.1728916f, -1.2823037f, 0.0284686f, -0.0879781f, 0.6332331f, 0.0599467f, -0.2467749f, 0.7796255f }, -{ 0.1541705f, -2.8967228f, 0.2088300f, 0.0289306f, -0.1897649f, -0.1835614f, 0.1872510f, -0.5846522f, -0.0145777f, 2.2226386f, 0.0885817f, 0.0293056f, -0.0056043f, 2.9454181f, -0.0623621f, 0.0230481f, -0.8140234f, -4.3990140f, -0.2562745f, 0.6827632f, -4.7245188f, 0.1150251f, 0.0615204f, 4.7553473f, 0.1170709f, 0.0822542f, -0.6365855f, 0.3014538f, -2.6740234f, 0.1919117f, -0.0003937f, -0.0227543f }, +{ 1.2430706f, 0.2385966f, -0.1307205f, 0.0325143f, -0.1869316f, 0.3236946f, -0.0250754f, 0.5280194f, -0.1030632f, -0.1919166f, 0.5273553f, 0.0652898f, -0.3363385f, 1.7545956f, -1.3353617f, -0.5440192f, -2.4663746f, -0.2166514f, 0.5136666f, -0.3368240f, -0.0757182f, 0.0920149f, -0.5879773f, -0.0225713f, 0.0697212f, 0.8563508f, -0.0126836f, -0.1828795f, -1.0318376f, -2.1025839f, 0.2279171f, 0.0045987f }, +{ -0.0346952f, -0.0168656f, 0.0407209f, -0.1482209f, 0.0585610f, -0.0068832f, -0.1967536f, 0.0088515f, 0.0117921f, -0.1666735f, 0.0043757f, 0.0192000f, 0.0101610f, -0.0203226f, -0.0722811f, -0.0180960f, 0.0279165f, -0.2452999f, -0.0029923f, -0.0095144f, 0.0082529f, 0.0223957f, 0.0012511f, -0.0040226f, -0.0002326f, 0.0050690f, -0.0015876f, -0.2731955f, -0.0103268f, -0.0168681f, 0.0081678f, -0.0030315f }, +{ -1.8121948f, -0.5878808f, -0.0582871f, 0.0676443f, -0.2566810f, 1.6678520f, 0.0751771f, -0.5183609f, 0.1661583f, -0.2097989f, -0.1213966f, 0.0194388f, 0.1857277f, -0.9746100f, -0.3444054f, -0.0763516f, 1.0879602f, 0.0184965f, 1.1060841f, 0.1923849f, -0.0412517f, -0.2354108f, 1.4612926f, -0.0095714f, 0.4891190f, -2.5078576f, 0.0020186f, 0.2102930f, 4.3355784f, 1.3152916f, -0.0975863f, 0.0044517f }, +{ 1.5896463f, -2.0003557f, -0.1210541f, -0.1304628f, 0.0131182f, -0.6161298f, -0.1486594f, 1.2760408f, -0.1021480f, 0.0877840f, 0.6571876f, 0.0804771f, 0.0627130f, 1.8012530f, -4.5879707f, 1.4820974f, 0.1388381f, -0.2526270f, 0.8011180f, 0.9165006f, 0.5784082f, 0.1881581f, -0.6235034f, -0.0051164f, -1.4052374f, 0.0785857f, -0.0013880f, 0.1095322f, -0.7017108f, -0.5751966f, -0.3369588f, -0.0015683f }, +{ 4.6118355f, -3.6102943f, 0.3775424f, 0.2372299f, 0.0598787f, 0.5582290f, 0.2304319f, 0.9093012f, -0.3773867f, 0.0047977f, 2.6425505f, -0.1439949f, 0.3935693f, 1.3001438f, 1.4970073f, -5.4273233f, -0.0687990f, 0.0670471f, 0.6703949f, 11.1249809f, 5.9740663f, 0.2207747f, -7.4394078f, -0.2285454f, 0.2585711f, 0.1874733f, -0.0016656f, 0.0217293f, 0.9555681f, 3.0554528f, -0.5727979f, -0.0026786f }, +{ -1.7840524f, -0.4664120f, -0.1709528f, -0.1229654f, 0.2472902f, 16.6959667f, 0.1021524f, 23.0225563f, -8.8298550f, 0.1945520f, 1.4932841f, -0.1441254f, 17.2260818f, 2.2682095f, -0.2566192f, -0.3763468f, -2.6818073f, -0.1413464f, -27.9071026f, 3.8443294f, 1.4446557f, -0.0005846f, 0.8942884f, 0.2102544f, -3.4054801f, 2.1674650f, 0.1294054f, 0.1876321f, 1.8387170f, 2.3066981f, -4.6334472f, 0.0203269f }, +{ 0.1221364f, -0.9206535f, -0.0578602f, 0.0086757f, -0.2020053f, 1.2470760f, 0.0963023f, 0.8218359f, -0.3108872f, -0.0679653f, 0.8242676f, -0.0567979f, -0.4664726f, 3.0569484f, 2.4042845f, -0.8774980f, 1.5614197f, -0.0359291f, 1.8720486f, 0.0217175f, -0.7438901f, -0.1705937f, 0.6922700f, -0.0708772f, 2.1580865f, -4.4101124f, 0.0070985f, 0.2220936f, 3.2375884f, -1.9589092f, -0.6771650f, -0.0064689f }, +{ -1.5325402f, -1.3310759f, 0.1091106f, -0.0624776f, -0.1820637f, 1.2152895f, 0.0591179f, -2.6746740f, -0.1679525f, -0.2618925f, -0.0701432f, -0.0717427f, -0.3926733f, 1.0515925f, 1.2097669f, -0.8023561f, 0.8519148f, 0.2428202f, -2.4016669f, -0.6443134f, -0.8455796f, 0.0883630f, -0.5246723f, -0.2599009f, 2.5098257f, 2.5735190f, -0.0004903f, 0.0228214f, 0.3101828f, -2.6110868f, -0.1449447f, 0.0112252f }, +{ -2.4570007f, 4.1306930f, -0.0709054f, -0.2456777f, -0.0434773f, 0.2500325f, -0.2469379f, 6.4086423f, 0.9810975f, 0.2620402f, 1.1104311f, -0.0850174f, 1.4035217f, -10.0576115f, -2.8340843f, -10.1643343f, -1.2882394f, -0.1628990f, 8.2278786f, 2.6009655f, 10.5084915f, 0.1247435f, 14.2758970f, -0.2018164f, -5.3260231f, -5.9585323f, 0.0054128f, 0.1511814f, -0.6517492f, 2.2414865f, -0.1060251f, -0.0213229f }, +{ -3.5239015f, 0.4428157f, 0.0098384f, -0.1035724f, 0.2192848f, 10.4552908f, -0.1394745f, 3.0283575f, -15.7956038f, 0.2377471f, -10.4040318f, 0.0577844f, 9.0448751f, 0.9642384f, 2.6048126f, 0.5453808f, 1.8676072f, -0.2262070f, -0.2392199f, -1.4581424f, 2.0151219f, 0.0250153f, 0.2735855f, 0.1904391f, -2.4851513f, -0.4783628f, 17.5952053f, -0.2103962f, -2.4205782f, -2.6922755f, -2.0195608f, -17.3869209f }, +{ 1.5699111f, 1.6051487f, -0.1158549f, -0.0307793f, 0.1400554f, 1.2221471f, 0.2116016f, 1.9285921f, 0.0042934f, 0.2085318f, 0.4481271f, 0.1907820f, 0.6049834f, -0.0403695f, -1.6741180f, 1.4565691f, 0.8936112f, -0.1921556f, 2.4673624f, 2.3142762f, 0.8624949f, 0.1622195f, -0.0171864f, 0.0495124f, 1.2759757f, -0.5775546f, -0.0131285f, 0.0019629f, 1.2494990f, -0.0695108f, 2.5645378f, -0.0251958f }, +{ 0.8203233f, -0.7310246f, 0.0924949f, 0.0314029f, 0.0004415f, 1.4434688f, 0.0189398f, -3.1009448f, -0.2930109f, -0.2146460f, -0.7676558f, -0.2426467f, -1.5872091f, 0.4062420f, 0.6988549f, 1.1018623f, 1.6228473f, 0.0036644f, -1.7428403f, -0.6364639f, -0.4087792f, -0.0277658f, 0.5860791f, -0.2002079f, 1.7377874f, 1.4471169f, -0.0070140f, -0.0075029f, 0.7090668f, 0.5809853f, 1.8224814f, 0.0000271f }, +{ -0.0271162f, 0.0158998f, 0.0030754f, -0.1410241f, -0.2236365f, 0.2567714f, -0.0673556f, 0.2602976f, -0.0035925f, -0.0680699f, 0.1860849f, -0.0224016f, 0.0799600f, -1.4051689f, 0.1068115f, 0.3181221f, 0.2607351f, -0.2092566f, 0.4402137f, 0.1669762f, 0.7319549f, -0.0637127f, 0.0924880f, -0.1975695f, 0.1949479f, -0.1818466f, 0.0019299f, -0.2326310f, -0.1074452f, 0.7910834f, -0.4338400f, 0.0037428f }, +{ -0.8883290f, -0.0982433f, -0.1428742f, -0.1160625f, 0.2654903f, 6.9522190f, 0.1999476f, 4.6878695f, 4.9569860f, -0.1421416f, 6.2829461f, 0.1076571f, -2.4888971f, 0.2070822f, 1.0991781f, -0.1491540f, -0.2840673f, 0.1541903f, -3.4112997f, 0.9145759f, 1.2377257f, 0.1603911f, 0.3144833f, -0.2557842f, -1.3645093f, 0.6859801f, 5.6290560f, 0.0619957f, -0.5346277f, -0.8084556f, -4.3686385f, -5.5265436f }, }; ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_layer2[32] = { -0.5612384f, -0.0400684f, -0.1890610f, -0.8038151f, -0.0184527f, -0.8351601f, -1.7656847f, -0.4417709f, -0.3462953f, 1.0924704f, 0.1019267f, 0.0497286f, 0.3448936f, -0.0442495f, -0.1294845f, -0.1740453f, -0.6256254f, 1.0588725f, 0.1306455f, 0.0451779f, 1.2896419f, -0.0145429f, 0.2775581f, -0.6205941f, 0.0369313f, -0.0632537f, -0.1257888f, -0.2130138f, 0.0593540f, 0.8294140f, -0.5174863f, -0.0208223f }; +-1.2246032f, -0.1255217f, -1.0217633f, -0.1814615f, -0.2436912f, -0.9311994f, -0.2956236f, -0.0172577f, -0.2968712f, 0.0741151f, -0.0555606f, 0.1839059f, -0.4699410f, -1.0013667f, 0.5347722f, 1.2293311f, 0.5103592f, 1.4865664f, -0.1879897f, -0.1940634f, -0.0728773f, -0.1775148f, -0.2175526f, 0.7983139f, 1.2735363f, -0.0693470f, 0.0497536f, -0.1538924f, 0.4001399f, -0.2371438f, -0.1275322f, 0.8652419f }; ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_layer2[32][32] = { -{ 0.0802436f, -0.1068745f, -0.1320603f, 0.0222937f, 0.0260173f, -0.0999878f, 0.0045460f, 0.0257662f, -0.1026595f, 0.2246336f, -0.0252555f, 0.0313543f, -0.2062458f, 0.1684278f, -0.0972477f, -0.2176001f, -0.1525148f, -0.0692302f, 0.1213232f, -0.1288088f, -0.1729289f, -0.0895176f, -0.0199082f, -0.0172326f, 0.0126052f, 0.0598741f, 0.0853796f, -0.1326686f, 0.0753646f, -0.0108036f, -0.1034372f, -0.0026847f }, -{ -0.5740064f, -0.1143318f, 0.1388070f, 1.3980483f, 0.0779263f, -1.2151324f, 2.2568495f, -0.5645000f, 2.2761426f, -0.2111119f, -0.3852313f, -1.2643801f, -0.8515005f, -0.1352748f, -0.1276971f, -0.9761337f, -1.1221037f, -0.1373484f, -6.1303482f, -0.6726473f, -3.5227783f, 0.3276748f, -0.8825266f, -0.0658490f, -0.1995521f, 0.4330961f, -0.0200665f, 0.1250316f, -0.1700243f, 1.1084150f, 1.6254629f, 0.3429218f }, -{ 1.1285430f, -0.2066509f, -0.0835447f, 1.6354681f, -0.1452742f, -0.4063630f, 0.9329628f, -0.3926201f, 0.1437214f, 0.2107810f, 1.6312121f, 1.3499900f, 0.3157536f, -0.0890818f, 0.1018713f, 0.1947485f, 0.2054133f, -0.2302379f, -0.1797780f, -0.5907550f, 0.9738173f, 0.6502790f, 0.2307118f, -0.2878186f, 0.7077893f, -0.2159140f, -0.0506834f, -0.1203295f, 0.0040051f, 0.9362236f, -0.8645781f, -5.5910249f }, -{ 0.1194026f, -0.0770950f, 0.0977710f, -0.0434290f, -0.0754589f, -0.0162118f, 0.1764810f, -0.0389150f, 0.1742229f, -0.0958543f, -0.1752356f, -0.0456456f, 0.1090995f, -0.0889283f, 0.0165932f, 0.0426875f, 0.0146571f, 0.0522842f, -0.0530897f, 0.0810161f, 0.0913054f, 0.0682782f, 0.0375430f, 0.0441132f, 0.0616512f, 0.1282314f, -0.0735286f, 0.0566166f, 0.1289222f, 0.0390186f, 0.0632017f, 0.0279914f }, -{ -0.1397050f, 0.1469029f, -0.1035157f, -0.1537557f, 0.0723362f, 0.1360317f, 0.1632015f, -0.0519030f, 0.1664423f, -0.1674401f, 0.1221813f, 0.1218546f, 0.0581016f, -0.0302506f, 0.1745771f, 0.0057055f, -0.0728391f, -0.1049056f, -0.1592961f, 0.0329875f, 0.0789358f, -0.0462632f, 0.0748179f, 0.0180050f, 0.0803397f, -0.1458464f, -0.1164891f, -0.0082622f, 0.0343072f, 0.0366396f, -0.1715891f, -0.1089048f }, -{ 0.5968121f, -0.1619606f, -0.1518290f, 0.7109355f, -0.2739125f, -1.2199029f, -0.0848204f, -0.9633313f, 1.5520710f, 0.9341283f, 0.2848818f, 1.7079298f, -0.0926562f, 0.0886664f, 0.0828461f, 2.1451118f, 1.2788274f, -0.1305077f, -0.1538649f, 0.6987577f, -1.6689410f, -0.0606114f, 0.8084964f, -0.4838576f, 0.0563342f, 0.7853296f, -0.0552194f, -0.1284064f, -0.1123770f, 1.0149344f, 0.1754804f, 0.2576492f }, -{ -2.3144670f, -0.0794675f, -0.1847414f, -0.1791797f, -0.0605982f, -3.8482890f, 0.9826989f, -1.9203123f, -0.8249935f, -0.6718075f, -1.6555610f, 2.0798690f, 1.5665153f, -0.2050105f, -0.1993937f, 1.1070000f, 0.0836252f, 1.6307852f, 0.2448520f, -2.1885965f, -0.2890749f, 2.9986718f, 2.8526762f, -0.7994567f, -0.3042918f, 0.0960785f, 0.0377693f, -0.0463109f, -0.0870433f, -0.8293707f, -1.3597184f, 1.3583397f }, -{ -0.3422491f, -0.0783274f, -0.1543446f, 0.0939283f, -0.3309467f, 0.6933140f, 0.6562154f, -0.6217518f, 0.7983661f, -0.1371530f, 0.5118276f, 0.7320337f, 0.5217202f, -0.0545936f, -0.1052059f, -0.1444394f, 0.3567903f, 1.2509772f, -0.6311634f, -0.7454629f, -1.0077031f, 0.8453025f, -0.2257671f, -0.2347097f, -0.3497045f, -0.1478627f, -0.0333489f, -0.2663794f, -0.0317280f, -0.3712497f, 0.8136677f, -2.5842018f }, -{ 0.0997253f, 0.0620212f, 0.0460688f, 0.1300038f, 0.1323504f, -0.1669361f, 0.0732264f, -0.0860083f, -0.0762404f, -0.1628099f, -0.0881021f, 0.1323460f, -0.0460503f, 0.1401906f, 0.0602387f, 0.1474468f, 0.0935555f, -0.1531849f, 0.1754870f, 0.1422898f, 0.1418013f, -0.0302381f, -0.1058683f, -0.0258421f, -0.0110884f, 0.1016085f, 0.0643939f, -0.0729882f, -0.0572575f, -0.0603358f, 0.0267279f, 0.0611828f }, -{ -0.6703482f, -0.0389322f, -0.1896956f, 1.0978996f, -0.1892595f, -2.0757272f, 1.7673731f, -0.7337908f, 2.1260054f, -0.6923556f, -0.6350248f, -1.8122731f, -1.2794832f, 0.0727381f, -0.1702943f, -1.0052446f, -1.1308233f, -0.1550428f, -5.3850775f, -0.5972114f, -0.3461530f, -0.4958812f, -1.2339170f, 0.0774786f, -0.8347382f, 0.2885106f, -0.1653138f, -0.0159324f, -0.1990542f, 1.6405339f, 1.7268503f, 0.8528640f }, -{ 0.0038449f, -0.2388104f, -0.1149939f, -1.2830256f, -0.0052918f, 0.3492746f, 0.3040406f, 0.6817812f, 0.8697712f, 0.4297135f, -1.2266355f, -0.6838435f, 0.4784633f, -0.2080981f, 0.1026195f, 0.4500216f, -0.4817615f, -1.5641849f, 1.6826199f, -0.6830738f, 0.3725868f, 0.4554783f, -0.0576175f, 1.0322572f, 1.9159367f, -0.3336473f, 0.1027016f, -0.2515334f, -0.1622610f, -1.8702512f, -0.7941946f, -2.9838107f }, -{ -0.2906766f, -0.0472568f, -0.2609593f, -1.5798510f, -0.1301123f, -1.1353090f, -2.5198724f, 1.5027611f, 1.2625716f, 3.2891662f, -2.3402910f, -0.0245398f, -9.8846655f, -0.2448200f, -0.0981539f, -0.2132508f, 0.7027491f, -3.4207478f, 1.3422097f, 4.3238688f, -1.1800685f, -2.7913725f, 1.8557802f, 5.5698090f, 0.2008359f, 0.2571939f, -0.0491005f, -0.1192926f, -0.0141392f, 0.1872108f, -0.8192848f, 0.6364858f }, -{ 0.0671812f, -0.0234023f, 0.1400131f, 0.0778011f, 0.1308578f, -0.1675161f, 0.0237332f, 0.0215410f, 0.1514422f, 0.0736446f, 0.0181612f, -0.0219220f, -0.0099684f, 0.0102909f, 0.1243076f, 0.0897413f, -0.0682666f, 0.0389046f, 0.1245468f, 0.0098897f, -0.0425716f, 0.1595688f, 0.0397469f, -0.0664724f, -0.1641733f, 0.0605745f, 0.1712325f, 0.1596854f, 0.0224220f, -0.1328268f, -0.1743169f, 0.0608113f }, -{ -1.1302176f, -0.0267920f, -0.2404846f, 0.6629997f, 0.1574058f, 4.7085104f, 0.5321816f, 0.4496275f, -0.2375129f, 1.2458097f, 1.8413728f, -1.8281459f, 2.3066695f, -0.2194766f, -0.0977243f, -0.4933192f, 0.5315795f, -0.8266373f, -1.6038543f, -0.6524141f, -0.2899630f, -4.3681431f, 0.2379541f, -0.3776518f, 1.5330435f, 0.3396196f, -0.0837295f, -0.0557159f, 0.1054505f, 1.1391810f, 0.6029354f, 0.9084895f }, -{ -0.0898137f, 0.0973598f, -0.1583033f, -0.1445000f, 0.0044857f, -0.1162652f, -0.1396598f, 0.1172944f, 0.0714815f, -0.1043701f, 0.0497041f, 0.0381792f, 0.1395592f, -0.0310079f, -0.1161861f, -0.0776435f, 0.0771079f, -0.0311096f, -0.1498033f, 0.0838249f, 0.1343088f, 0.1152903f, 0.0749620f, -0.1248982f, -0.0346379f, -0.0780947f, -0.0730843f, 0.1654648f, -0.1482577f, -0.0118278f, 0.1078758f, -0.1372479f }, -{ -0.7470447f, -0.0269854f, 0.0793572f, -0.7553226f, -0.0888210f, 1.8323143f, 0.2497575f, -10.9856701f, -0.7587091f, 1.1695130f, -0.0198075f, -0.8016124f, -0.2920889f, -0.1624553f, -0.0925370f, -1.3623806f, -1.3067284f, 0.2490501f, 0.0292220f, -0.2309522f, 0.0131977f, -0.7858045f, 0.1076498f, 0.6439033f, -0.1686209f, -0.8376603f, 0.0470587f, 0.0656591f, 0.0300643f, 1.4373403f, 0.2995886f, -1.3597747f }, -{ -0.8746193f, -0.0201447f, -0.1603909f, -0.9373384f, 0.0254792f, 0.7993639f, 0.1087815f, -0.4859151f, -1.9559643f, -2.9621441f, -1.7735658f, -2.2227323f, -0.3725011f, 0.0284650f, -0.2266676f, -0.7529051f, -0.4477961f, 0.2071678f, 2.2770953f, 1.1170574f, -0.7023426f, 0.6896675f, -1.2416189f, -1.0012786f, -1.8752691f, -0.0359559f, -0.0125555f, -0.0457818f, -0.0775177f, 0.6092747f, 0.6639680f, 1.0951738f }, -{ -1.7250717f, -0.0309622f, 0.1054536f, 2.0844676f, -0.2957073f, -0.0859962f, -1.5239947f, -0.2195731f, -1.5450290f, 0.4916542f, -1.3940116f, -0.5043938f, -0.7850562f, -0.1393226f, 0.0421535f, -0.2805571f, 0.5149463f, -0.4160199f, 1.3064783f, 0.3980424f, 0.1537250f, 0.2291165f, -1.5061877f, -0.0060016f, 0.1236818f, -0.1598873f, 0.0182953f, -0.0649871f, -0.1681922f, -0.8940877f, -0.6040494f, -0.5640311f }, -{ 0.1721584f, -0.0760296f, 0.0164177f, 1.6552234f, -0.2948481f, -0.8678465f, 0.9449416f, -2.5188146f, -0.3830298f, -0.9600880f, 1.3968240f, 0.4318309f, 1.2421557f, -0.2275233f, -0.2093997f, -0.4352153f, -0.2142241f, -1.6719555f, -1.5515612f, -1.5296252f, 0.6067497f, 2.3952096f, -1.4322679f, -1.7078539f, -0.0283693f, -0.0217880f, -0.0268028f, -0.0223932f, -0.0441584f, 0.2260311f, -0.3117388f, 0.0257804f }, -{ -0.3736052f, -0.1436419f, -0.1791924f, 0.6735403f, -0.1233307f, 0.9371016f, -0.5032559f, 0.0065924f, 0.2331814f, 0.5494700f, 1.4163370f, 0.7603047f, 1.8672758f, -0.1058595f, -0.0214494f, -0.1112118f, 0.4055682f, -0.2838995f, -1.8654461f, -0.6464235f, 0.3173830f, 0.1275925f, -0.0674378f, -0.1359007f, 0.8348094f, 0.2912215f, -0.1404384f, -0.2908648f, -0.1660577f, 0.4171410f, -0.5499387f, -0.9420084f }, -{ -10.1053333f, -0.1126838f, 0.0914457f, 2.1485925f, -0.2125181f, -0.6991695f, -1.5648814f, 3.5901635f, 0.4265155f, 2.2461650f, 1.5659872f, -4.5079947f, -1.8503289f, 0.0036438f, -0.0552862f, 6.9531665f, 7.5919185f, -5.2455883f, 1.9158086f, -0.4571260f, -8.5806217f, 3.1702795f, 4.6932554f, 0.9312032f, -0.2189541f, -35.8472443f, -0.0437394f, 0.0193345f, -0.0759754f, -0.8393922f, -8.7812824f, 1.2664827f }, -{ -0.4285544f, -0.0112712f, -0.1986143f, -0.6809276f, -0.2643681f, 0.8245937f, -0.1275320f, 1.8360085f, -1.0321223f, 1.0097107f, -0.4381949f, -0.0334159f, -0.1864567f, 0.0119804f, -0.1997918f, -0.3271705f, -0.0386158f, 1.0598557f, 0.9783336f, 1.9556490f, -0.2163550f, -2.8529193f, 0.6785525f, 1.3684491f, 1.2711524f, -0.1338721f, -0.0004762f, -0.0270106f, -0.1851630f, 0.1504625f, 0.1159483f, 1.3600112f }, -{ 0.1121258f, 0.0640089f, -0.1267886f, 0.0409490f, -0.1376833f, -0.0064835f, 0.1204074f, -0.0622866f, 0.1238085f, 0.0468351f, -0.0761505f, 0.0465574f, 0.0356306f, -0.0866522f, 0.1416768f, 0.1586308f, 0.1173862f, -0.0805042f, 0.1809647f, -0.1865381f, -0.1134762f, -0.1078163f, 0.1637889f, 0.3176169f, 0.1323416f, 0.0248524f, -0.0535714f, 0.1136851f, 0.0925486f, 0.0475878f, -0.0636055f, -0.0584198f }, -{ -14.6186295f, -0.1993816f, 0.1559632f, 2.3735945f, -0.0626924f, -8.2991180f, 0.2650338f, 3.5373747f, 1.6012716f, 2.8416808f, 1.1672604f, -4.3095527f, -1.1607343f, 0.0817207f, 0.1491408f, 7.4242001f, 7.2067833f, -4.3056359f, 0.8474119f, -0.9166479f, 0.6955321f, 2.0489492f, 4.0285969f, 0.0758955f, -1.0957255f, -38.1951637f, -0.0826333f, -0.1540980f, -0.2915384f, 1.0447723f, -9.4216700f, 0.7907819f }, -{ 1.0854393f, -0.2057757f, -0.2333630f, -2.8073475f, -0.0092103f, -0.0027702f, 0.0400727f, 1.1957222f, 2.1408370f, 2.4174137f, -4.4619761f, -2.1267519f, -2.3692110f, -0.0215195f, 0.0940003f, 1.5145874f, 0.4090779f, 1.0525370f, 0.7181456f, 2.2475765f, -4.0287180f, 0.2974668f, 1.9458562f, 0.4356304f, -1.8651968f, 0.1091178f, -0.0625651f, 0.0101182f, -0.1866876f, 2.0775406f, -0.0924411f, -0.9942529f }, -{ 0.6088188f, -0.1557965f, -0.1098123f, -2.0238423f, -0.0628158f, 1.8817046f, -2.8116503f, -1.9974719f, -2.8561919f, -1.5064508f, -0.5036830f, 0.0460033f, -1.2808908f, -0.2553190f, -0.2242925f, 0.3286708f, 0.6866547f, -0.8440441f, -3.4105279f, 1.5247352f, 0.5397177f, -0.5575525f, -0.4773433f, -0.8334278f, -1.8625259f, -0.5315101f, -0.1528733f, -0.1654517f, -0.1824549f, -0.9775680f, -2.9137912f, -0.4873172f }, -{ -3.4014626f, -0.0843555f, 0.0550500f, 1.7411346f, -0.1923270f, 0.9657442f, -2.5730155f, -0.4918169f, 0.7225664f, -1.8889602f, 3.7809584f, 1.7930225f, -1.6396630f, -0.2156895f, -0.1968789f, -0.4836586f, -1.0472771f, 3.1005175f, -3.8917956f, -4.2640653f, 0.9576553f, -0.5442499f, -0.7966165f, -7.2177167f, 3.5426965f, 0.3307767f, -0.1710942f, -0.1729530f, -0.0668134f, -0.4916894f, 1.9444537f, 0.5960958f }, -{ -0.9639474f, -0.0439117f, 0.0067572f, -0.8821743f, 0.1859997f, -0.7608641f, 1.5826563f, -1.4137437f, -0.3524639f, 0.4026311f, -0.0760814f, 0.7603636f, 2.5665376f, 0.0416676f, -0.1360063f, -0.6073055f, -1.3086349f, 1.9021115f, 1.3700111f, -0.9890064f, -0.1628284f, -1.3746781f, 0.2909067f, 0.3049279f, 0.8926148f, -0.4370743f, -0.0742112f, -0.0174381f, 0.0485841f, -0.0802396f, 3.5397923f, -0.6531604f }, -{ 0.0491318f, -0.0591713f, 0.0498613f, 1.1349666f, -0.0889358f, -1.5782114f, 0.4471801f, 1.2610462f, -0.4626705f, 0.9874090f, 1.3419105f, 0.9187180f, 0.3291495f, -0.0520998f, -0.0861611f, 0.0016949f, -0.3544937f, -0.4579405f, -0.3237021f, -0.2511895f, -3.9122045f, -0.4933105f, 1.1840253f, 0.0769721f, 1.6942110f, -0.0282394f, -0.0797413f, -0.0965752f, -0.1654065f, 0.6972140f, 0.0176173f, 0.2189070f }, -{ -1.1536870f, -0.1555014f, -0.3485503f, -1.5355002f, -0.2638741f, 1.3639680f, 1.6288967f, 0.6296268f, -0.5865426f, -0.5947766f, -1.4163440f, -1.5454646f, -0.7648816f, -0.1768368f, -0.1375362f, 0.3310686f, -0.2621009f, 0.6404898f, 1.0589499f, 0.2490305f, 0.3255049f, -0.1451550f, -0.3024975f, 0.8173084f, -2.1893589f, -0.0099213f, -0.1596652f, 0.1068924f, 0.0260035f, 1.4544648f, -0.7779374f, -0.7673930f }, -{ 0.0435760f, -0.0246287f, -0.1238125f, -0.0320143f, 0.1576336f, 0.0702628f, -0.1744860f, -0.1600130f, -0.0870369f, 0.0819791f, 0.0445500f, 0.0019862f, 0.1529912f, 0.0886548f, 0.0478378f, -0.0487068f, 0.0401030f, 0.0869702f, 0.1359502f, 0.1713400f, -0.1529109f, 0.1224933f, 0.0072967f, -0.0784628f, 0.1494704f, -0.1378658f, -0.0747428f, -0.0850826f, -0.0940957f, 0.0511818f, 0.1520686f, -0.0680635f }, -{ 1.3802143f, 0.0130436f, -0.0118297f, 0.9667702f, -0.4034582f, 0.3238442f, 0.4695910f, -0.3288718f, 1.1822759f, 1.5390285f, 0.2593471f, -0.0801804f, 0.4593041f, 0.0215906f, -0.2340092f, 0.4232760f, -0.0674659f, 0.7093915f, -0.0001982f, 0.2270681f, -0.8337699f, -0.2296698f, 0.9161285f, 0.3906515f, -0.1553026f, 0.2775815f, -0.0620411f, -0.1116518f, -0.0008762f, -0.3016884f, 1.4699873f, 0.6260686f }, +{ 0.5089083f, 0.0575017f, 0.5365443f, -0.0749042f, 0.0996821f, -0.0870541f, 1.4083110f, 0.0042451f, 0.0506413f, 0.1160910f, -0.0628334f, 0.5274143f, 0.4513455f, -0.9582062f, 0.0673435f, 0.6348773f, -1.9705150f, -0.2561726f, -0.1166724f, 0.0368685f, -0.1610932f, 0.0683969f, 0.0069520f, -0.1350110f, 0.9181439f, 0.1249722f, 1.0909826f, -0.0840018f, 0.9957948f, -0.0971043f, -0.0137849f, -2.2041860f }, +{ 0.6033787f, 0.0869169f, -0.0859117f, 0.0315756f, -0.1130356f, 0.7097771f, -0.1339651f, -0.0328042f, 1.0109031f, 0.2830297f, -0.0936190f, -0.6122728f, 0.3918095f, 1.2013817f, 0.4156620f, -0.6251741f, 1.1518294f, 1.6751471f, -0.2515613f, -0.0678125f, -0.1258993f, 0.0016276f, 0.0489553f, -0.3969306f, -1.0053991f, -0.1408774f, -0.5742593f, -0.3888345f, -1.0801376f, 0.1232338f, -0.2474749f, 1.3247567f }, +{ -0.0069317f, -0.0813263f, 0.0308648f, 0.0865811f, -0.1700420f, 0.0967914f, -0.0156711f, -0.1217743f, -0.0792132f, 0.0874347f, -0.0772177f, -0.0141818f, -0.0828527f, 0.1910101f, 0.0311023f, 0.0512783f, -0.1374564f, -0.1192217f, 0.0168840f, -0.0671096f, -0.1756282f, 0.0961502f, 0.1194936f, 0.1121689f, 0.0702861f, 0.0227058f, -0.1065478f, 0.0892736f, 0.1295921f, -0.0824243f, 0.1052793f, 0.1367913f }, +{ 0.0218436f, 0.0658335f, -0.1404412f, 0.1147386f, 0.1404460f, -0.1633583f, -0.1220772f, 0.1565208f, -0.0028298f, -0.0484391f, 0.1510092f, 0.0730970f, 0.0835523f, -0.1401114f, 0.1184470f, 0.1755763f, -0.0048258f, 0.0622622f, -0.0341085f, -0.0806764f, -0.1538521f, -0.0015015f, 0.1527237f, -0.1283317f, -0.1138941f, 0.0758406f, 0.1322748f, 0.0583293f, -0.1526688f, -0.0575901f, 0.0629805f, 0.1006109f }, +{ -0.0993613f, 0.1225954f, 0.0214997f, -0.1216850f, -0.1446034f, -0.1262178f, 0.1148179f, -0.0104865f, 0.1139638f, 0.1162356f, 0.0760315f, 0.0347920f, 0.1263798f, -0.0845871f, -0.0076240f, 0.0147776f, 0.0706741f, -0.0541775f, 0.1547825f, -0.0731929f, -0.1654783f, 0.1004926f, 0.0731769f, -0.0515466f, 0.0012489f, 0.1251491f, 0.0608163f, -0.1210160f, -0.0371475f, 0.0280545f, 0.0563840f, 0.0180449f }, +{ 0.8394753f, 0.0863981f, 0.6413645f, 0.0409160f, -0.0078189f, -3.7878563f, 0.2678362f, 0.0196527f, 1.7210206f, -1.6028817f, -0.2068004f, -2.3214684f, -1.0184679f, -1.5257946f, 0.2028781f, -2.7971203f, 0.5778140f, -0.2566824f, -0.0615155f, -0.1043012f, -0.1660616f, -0.1092668f, -0.0927776f, 1.8451736f, 0.4113030f, -0.1081991f, 0.9670511f, 0.0100726f, -2.8746226f, -0.1742988f, -0.1429935f, 1.5318521f }, +{ -0.0845136f, 0.0384511f, -0.1385615f, -0.0871150f, 0.0897845f, 0.0878664f, -0.0010738f, -0.1549659f, 0.0356569f, -0.1536330f, -0.0262665f, -0.1161826f, 0.1544266f, 0.0787109f, -0.0407829f, -0.0264572f, -0.0792578f, -0.0405977f, 0.1106679f, 0.1500933f, -0.0898741f, -0.1199529f, -0.1163880f, -0.1057636f, 0.0889505f, 0.0825375f, 0.1614031f, 0.0067508f, 0.0352729f, 0.0078936f, 0.1533775f, 0.0849639f }, +{ 2.7207191f, -0.0055598f, 2.5277357f, -1.4293425f, 0.1117166f, 0.8231006f, -2.0332973f, -0.0739002f, 1.8611740f, -8.6415014f, -0.0123378f, -0.3534637f, -1.7967031f, 2.9673553f, 0.2152573f, -4.9613218f, 0.3553480f, -0.1516881f, 0.0313580f, 0.0298279f, 0.1744945f, -0.0543368f, -0.0657357f, -0.3597513f, 1.8521340f, -0.0200091f, 1.6491756f, -0.1347934f, -8.2300911f, 0.0531232f, 0.0599637f, -0.1086621f }, +{ -0.7867953f, -0.0110963f, -0.4613801f, 1.0450928f, 0.1429149f, 1.4774159f, -1.1104021f, 0.0747843f, -0.8985512f, -1.1148651f, 0.0115854f, 0.1383941f, 0.1892344f, -0.4564212f, -0.1128876f, -6.8694520f, -0.1744848f, 1.5719544f, -0.1184832f, 0.0280010f, -0.0129338f, -0.0419609f, -0.1565810f, -1.3889359f, -2.2191308f, 0.0424039f, -1.2463529f, -0.0588384f, -0.4484319f, -0.1165283f, 0.0796688f, -0.5284064f }, +{ 0.0652653f, 0.0615156f, -0.0795163f, 0.0001713f, 0.1648186f, -0.0528994f, -0.0847025f, -0.0793956f, 0.1112951f, -0.1075590f, -0.0029891f, -0.1321859f, -0.0573127f, -0.1392062f, 0.0507489f, -0.0275592f, 0.1043706f, 0.1051407f, 0.1235337f, -0.0951271f, 0.0715892f, -0.0868704f, -0.0582845f, -0.0601897f, 0.1471449f, 0.0802297f, 0.1092225f, -0.1009769f, 0.1374661f, -0.1123024f, -0.1594280f, 0.1239780f }, +{ 0.4587068f, -0.0936088f, 0.1091407f, -1.4789468f, -0.2378854f, -0.7586868f, 1.2026873f, -0.1347813f, 0.6108205f, -0.9265009f, -0.1364171f, 0.4698360f, 0.1535055f, -0.2266157f, 0.0333197f, 1.4881195f, -0.1966134f, -0.9363742f, -0.2550847f, 0.0512618f, 0.0478554f, -0.0761714f, 0.0852998f, 0.7968215f, 0.3493969f, -0.2375832f, 0.6220873f, -0.1845089f, 1.1993877f, -0.0197335f, -0.0138845f, 1.9340528f }, +{ -0.1209233f, 0.1684445f, -0.1172153f, 0.0205280f, -0.1180427f, -0.0904171f, -0.1664954f, -0.0941068f, 0.0728728f, 0.1295698f, -0.1708585f, -0.1739318f, 0.0748263f, -0.1401377f, 0.1161512f, 0.1253984f, 0.1506475f, 0.0483222f, -0.0940781f, -0.1165934f, 0.1754140f, 0.1053056f, -0.1382262f, 0.1219736f, -0.1636132f, 0.1426921f, -0.0591449f, 0.0103367f, -0.1190467f, -0.0959214f, -0.0777284f, -0.1641069f }, +{ -0.6834289f, -0.0725848f, -0.2700611f, 1.0437120f, 0.0824887f, -2.7070620f, 0.4225597f, -0.0604267f, -0.6852378f, 0.1944695f, 0.0911727f, 0.1900434f, -0.0984350f, 1.9884139f, 0.4437184f, 0.7536621f, 0.0569126f, 0.6176490f, -0.0592906f, -0.0017330f, 0.0474772f, 0.0168128f, 0.0533080f, -1.2100418f, -2.4400961f, -0.0830669f, -1.3237737f, 0.1108096f, -1.6645818f, -0.1670439f, -0.0161379f, -0.3699313f }, +{ -0.2601279f, -0.0627888f, -2.3651617f, -1.1256676f, -0.0283316f, -6.7830796f, -3.6895976f, -0.1580759f, -1.4195836f, -6.0885744f, 0.0452110f, -5.3680663f, -2.2806647f, -5.1564946f, 0.0099778f, 2.9737246f, -3.1940920f, -0.6476036f, 0.1829322f, 0.0246909f, -0.1631330f, -0.0444936f, -0.2123401f, -4.5203066f, -0.0740298f, 0.0058150f, -0.7009476f, 0.1273159f, 2.3706875f, 0.0853582f, -0.2045061f, 2.0988390f }, +{ 0.0851754f, -0.1730167f, -2.2383544f, -1.4635299f, -0.0050372f, -1.0768518f, 1.2955434f, -0.1811388f, 2.4874403f, -0.0573546f, -0.0442298f, -1.1430494f, -0.1609198f, 0.5267517f, 0.1293965f, -0.7079992f, 0.9464460f, 1.9218533f, -0.1341864f, -0.0830647f, -0.0308418f, -0.0314987f, -0.1568387f, -0.5759755f, -3.5660131f, -0.2200986f, 1.9044633f, -0.0455861f, 1.2135637f, -0.0406553f, -0.0549359f, -3.2340167f }, +{ -0.3601795f, -0.0950057f, 1.2506303f, 0.0705646f, 0.0159455f, 0.1998335f, -0.9806644f, -0.0993264f, 0.2361836f, 0.2085727f, -0.1145219f, -0.4826447f, 0.0773994f, 0.0078874f, -0.0499348f, 0.0593346f, 2.1682246f, 0.8369263f, -0.2426349f, 0.0915828f, 0.0177455f, -0.1256301f, -0.1294592f, -0.7195153f, -2.4313867f, -0.2160123f, -1.9654233f, -0.3661108f, 0.0221972f, 0.0420728f, -0.1662701f, 2.2909124f }, +{ 0.5229083f, -0.0859756f, -1.2948138f, 0.2981660f, -0.0616795f, 0.5993323f, 1.2229108f, -0.1168597f, -0.1483506f, 0.2205500f, -0.1778468f, 0.7693940f, -0.1189292f, 0.4196700f, 0.5927172f, 0.8341553f, 1.0524089f, -1.6661072f, -0.1706631f, -0.1182255f, -0.1298270f, 0.1067912f, -0.0550823f, -0.3989949f, 1.0862428f, 0.0692725f, 0.0220655f, -0.0182227f, -0.3141328f, 0.0763854f, -0.2399873f, 0.3735673f }, +{ -0.0493408f, -0.0620185f, -0.1440084f, 0.1366522f, -0.1421596f, -0.1071353f, 0.1569152f, -0.1714708f, -0.0085891f, -0.1523663f, -0.0666709f, 0.0920407f, -0.0104404f, 0.1303387f, -0.1666427f, -0.0643643f, -0.1648330f, -0.0403286f, 0.0011041f, 0.1209591f, 0.0040688f, -0.0941681f, -0.0029900f, -0.0372638f, 0.0528831f, -0.0957583f, -0.0537955f, -0.0797893f, -0.1704947f, -0.0559995f, 0.0740464f, -0.0752997f }, +{ 1.5719815f, -0.1644378f, 1.8601693f, 5.5668306f, -0.1303247f, 5.0759544f, -1.4812335f, 0.0062226f, 0.5767426f, 5.0568337f, -0.1407256f, 4.7255578f, -0.6883666f, 2.1262701f, 0.3128934f, 2.2308426f, -1.1327085f, 3.4682004f, -0.2262406f, -0.1728720f, 0.1500425f, -0.1453467f, 0.0638414f, 0.0793114f, -0.3400198f, -0.2237183f, 0.1568154f, -0.0257598f, -0.4672260f, -0.1525187f, -0.2054953f, 0.1591575f }, +{ -0.4987546f, -0.0995077f, 2.7520263f, 0.4053383f, -0.1276614f, 0.0818262f, -0.9757426f, -0.0043751f, -1.8990395f, -1.7651865f, -0.0606477f, -0.6438876f, -0.1125105f, -0.6256728f, -0.0335146f, 0.7741688f, 0.3912064f, -0.9055670f, -0.0045660f, 0.0077818f, -0.1204589f, 0.0773467f, 0.0537127f, -1.7081609f, 0.7387068f, -0.0725944f, -0.1396644f, 0.1628178f, -0.7278796f, -0.1286270f, -0.1079767f, -0.2916186f }, +{ 0.1261214f, -0.0106476f, -1.0569049f, 0.2244065f, -0.1248134f, 0.9224349f, 0.1777439f, -0.1238517f, 0.3914232f, 0.1859665f, -0.1939506f, -0.4722152f, 0.2143378f, 0.8231044f, -0.2002590f, -1.3563371f, -0.2053423f, 1.3496532f, -0.1801796f, -0.0217177f, 0.0810915f, -0.1319678f, -0.1299806f, 0.9309435f, 1.2306161f, -0.1924531f, 1.8588091f, -0.2374240f, -1.2482435f, 0.0170704f, 0.0808160f, -1.6042936f }, +{ 0.0353613f, -0.0415041f, 0.0430906f, 0.0391508f, 0.1654718f, 0.1647980f, 0.1752385f, 0.0376194f, 0.0297110f, -0.1648499f, -0.1170499f, 0.0424385f, -0.1728038f, 0.0867688f, 0.1044088f, -0.0208259f, -0.0972596f, 0.0016974f, -0.0276596f, -0.1079382f, -0.1497553f, 0.0368193f, -0.1487449f, -0.0142207f, -0.1384017f, -0.1133966f, 0.1205297f, -0.1694416f, -0.0848187f, 0.0165042f, -0.0311819f, -0.0107586f }, +{ 0.3395178f, -0.0544951f, 0.4140292f, 0.4284469f, -0.1532507f, -0.8734673f, 1.3085660f, -0.0968414f, -0.9431682f, 0.0212970f, -0.0573725f, 0.2512161f, 0.1806573f, 1.9569112f, -0.2206540f, -1.4288682f, 2.3636749f, -2.0449739f, -0.0088101f, -0.0307277f, -0.1746318f, 0.0359623f, -0.1769885f, -0.1383351f, -3.5865929f, 0.0237848f, 0.2055962f, -0.2543508f, 0.9893691f, -0.1158237f, -0.1074060f, -0.1367110f }, +{ -0.0818370f, -0.0691959f, 0.1008360f, -0.0867900f, 0.0706240f, 0.0653615f, -0.1036123f, -0.0565025f, 0.0877569f, -0.1702891f, -0.0305304f, 0.0311808f, -0.0784205f, 0.1343800f, 0.0457310f, 0.0330426f, -0.1380319f, -0.1204205f, 0.0351189f, 0.1660071f, -0.1655417f, -0.1318305f, -0.1387838f, -0.1377107f, -0.1507244f, -0.0132244f, -0.0144924f, 0.0727539f, -0.0959819f, -0.0384600f, 0.0876117f, 0.0915885f }, +{ -0.2913065f, -0.1268295f, 0.0402893f, 0.5167278f, -0.0104059f, 0.2077989f, 0.3679934f, -0.2081723f, 0.1353243f, 0.8197210f, -0.1013957f, 1.1287574f, 0.1784925f, 1.4747944f, 0.0095530f, -0.2737869f, 0.7512702f, -0.4536139f, -0.2225167f, -0.1508588f, -0.1516107f, -0.0296172f, 0.0093031f, -0.0913578f, -0.1219817f, -0.0470407f, 0.2288801f, -0.1066827f, 0.9788207f, -0.1060857f, -0.2672491f, -1.1543201f }, +{ 0.8206128f, 0.0830891f, 0.6098021f, 0.8114421f, 0.0100041f, 0.1802981f, 1.4186163f, -0.0978715f, 1.4950410f, -0.0608325f, -0.0877750f, -2.8319721f, 0.4774732f, -0.0360274f, -0.4008928f, -1.3351992f, -0.6958030f, 0.4174623f, -0.1046506f, -0.0839466f, -0.0734599f, -0.0768588f, -0.1111481f, 1.5469090f, -2.7853961f, -0.1235103f, 2.0516131f, -0.1304165f, -0.6423004f, -0.0017117f, -0.0383703f, -1.2066915f }, +{ 5.9012880f, -0.1430433f, -4.8122344f, 1.3565518f, 0.0676391f, -6.5592794f, -10.6126184f, -0.0076572f, 3.2418084f, 0.8378894f, -0.0120687f, -2.6607361f, -32.6262512f, 0.5281777f, -40.5600204f, 0.2489026f, -1.4531211f, -3.3126616f, -0.2288216f, 0.0087201f, 0.1374615f, 0.1106128f, 0.0246236f, -22.0268345f, 0.0629736f, 0.1088031f, 4.0246558f, 0.0067829f, -3.1944098f, -0.0843966f, -0.1402911f, 0.2893855f }, +{ 0.0687621f, -0.0233423f, -0.0049267f, 0.0329577f, -0.0040116f, -0.0038231f, -0.1479600f, 0.0412774f, -0.1182764f, -0.0800665f, 0.0088746f, -0.0386593f, 0.1165797f, -0.0171098f, 0.0882285f, -0.0606064f, -0.0594507f, -0.0446539f, -0.1713613f, -0.0783319f, -0.1286740f, 0.0856650f, -0.1753173f, -0.0516984f, -0.1444318f, 0.1162534f, 0.0405000f, -0.0671118f, -0.0576360f, -0.1409870f, -0.1164895f, -0.0137712f }, +{ -1.3168329f, -0.0033607f, 1.7238024f, -2.6184859f, -0.1436181f, -0.4960048f, 0.3308013f, 0.0842941f, -0.8149745f, -0.5922119f, 0.1075923f, -0.8475354f, -4.2522564f, -0.6004850f, -1.7454799f, 0.7431618f, -1.3471994f, 1.8992596f, 0.0583799f, -0.0095420f, -0.1535583f, -0.0602523f, -0.1145862f, -0.9051632f, 0.4375299f, -0.0222844f, -0.7205035f, -0.3361300f, 1.7331023f, -0.1060063f, -0.1233467f, -0.9474893f }, +{ 0.0834905f, -0.0987103f, -1.4479494f, 1.1476132f, -0.1930190f, -0.8689866f, -0.8929498f, -0.1202456f, -0.4120998f, -0.3361138f, -0.1739257f, 0.7572783f, 0.2644251f, -1.1061769f, -0.0444610f, 0.5947868f, -1.8700641f, -0.8042451f, 0.0032584f, -0.0918714f, -0.0878237f, -0.0561467f, -0.0953190f, -1.0795568f, -1.2196006f, -0.1444945f, -1.2117988f, -0.1134432f, 0.3373002f, -0.0059311f, 0.0073372f, -0.0959612f }, +{ 0.4720845f, -0.1032461f, -0.2123445f, -1.0046439f, -0.1504408f, 0.7121164f, 0.3517790f, -0.1829089f, 0.8960306f, -0.2978761f, -0.1084950f, -0.3584599f, -0.1990081f, 0.5987743f, -0.5727913f, 0.0202107f, -0.9536695f, 0.3441935f, 0.0361146f, -0.2192044f, -0.0706470f, 0.0332405f, -0.0892546f, 1.4020443f, 0.8471918f, -0.1363626f, 0.3598230f, 0.0537083f, -0.2526802f, 0.0390054f, -0.0971618f, -0.1145800f }, +{ 6.6881781f, 0.0808753f, -4.0573001f, 5.1219244f, -0.1900990f, -0.9485528f, -8.2378635f, -0.0019337f, 8.3292599f, -5.9191675f, 0.0768199f, -1.6080343f, -38.4854393f, 2.5592017f, -35.8240967f, -2.1315293f, 1.3917850f, 3.9621065f, -0.0458908f, 0.1264065f, 0.0298109f, 0.0203746f, 0.0184819f, -14.0934982f, 1.2857219f, 0.0492928f, 4.1220269f, -0.3709270f, 2.5108135f, -0.0593165f, 0.0093613f, -1.2152934f }, }; ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_output_layer[1] = { -0.7275639f }; +1.4821327f }; ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_output_layer[32][1] = { -{ 1.4243358f }, -{ 0.0335807f }, -{ 0.0551641f }, -{ -0.9836086f }, -{ -0.0249541f }, -{ -1.5375688f }, -{ -0.7714168f }, -{ -0.9649364f }, -{ -1.1769278f }, -{ 1.3249911f }, -{ -1.6541473f }, -{ 1.4079021f }, -{ -0.8831168f }, -{ 0.0122874f }, -{ 0.0511134f }, -{ -2.6734750f }, -{ 2.8394303f }, -{ 0.9675560f }, -{ -1.4186903f }, -{ -2.0796514f }, -{ -1.7693948f }, -{ -0.8502544f }, -{ -1.5927037f }, -{ -1.1028550f }, -{ 0.8137528f }, -{ 6.3073616f }, -{ 0.1059108f }, -{ -0.0468376f }, -{ 0.1322162f }, -{ 0.7481517f }, -{ -1.2260461f }, -{ -0.9095332f }, +{ 2.0100033f }, +{ 0.1341594f }, +{ -1.1707165f }, +{ -1.1725436f }, +{ -0.1255062f }, +{ -1.1738188f }, +{ 1.0822003f }, +{ 0.1121197f }, +{ -1.0663537f }, +{ -0.7722268f }, +{ 0.0201395f }, +{ 1.4703517f }, +{ 2.3982928f }, +{ -0.5911243f }, +{ 3.9915426f }, +{ -0.9631299f }, +{ -0.6016800f }, +{ 1.3677936f }, +{ 0.0113799f }, +{ 0.0772981f }, +{ 0.0646394f }, +{ 0.0517149f }, +{ 0.0408024f }, +{ 1.2203258f }, +{ -0.6350328f }, +{ -0.0750008f }, +{ -1.5730240f }, +{ -0.0382495f }, +{ -1.2106268f }, +{ 0.0631248f }, +{ 0.0212301f }, +{ -0.8234047f }, }; } diff --git a/RecoTracker/LSTCore/src/alpaka/NeuralNetworkWeights.h b/RecoTracker/LSTCore/src/alpaka/T5NeuralNetworkWeights.h similarity index 99% rename from RecoTracker/LSTCore/src/alpaka/NeuralNetworkWeights.h rename to RecoTracker/LSTCore/src/alpaka/T5NeuralNetworkWeights.h index 42f7b19f33898..9938a16cb03fd 100644 --- a/RecoTracker/LSTCore/src/alpaka/NeuralNetworkWeights.h +++ b/RecoTracker/LSTCore/src/alpaka/T5NeuralNetworkWeights.h @@ -1,5 +1,5 @@ -#ifndef RecoTracker_LSTCore_src_alpaka_NeuralNetworkWeights_h -#define RecoTracker_LSTCore_src_alpaka_NeuralNetworkWeights_h +#ifndef RecoTracker_LSTCore_src_alpaka_T5NeuralNetworkWeights_h +#define RecoTracker_LSTCore_src_alpaka_T5NeuralNetworkWeights_h #include diff --git a/RecoTracker/LSTCore/standalone/analysis/DNN/train_T3_DNN.ipynb b/RecoTracker/LSTCore/standalone/analysis/DNN/train_T3_DNN.ipynb index b3bf66c2140ea..d6d51ba5bf9c5 100644 --- a/RecoTracker/LSTCore/standalone/analysis/DNN/train_T3_DNN.ipynb +++ b/RecoTracker/LSTCore/standalone/analysis/DNN/train_T3_DNN.ipynb @@ -42,13 +42,15 @@ " 't3_layer_binary',\n", " 't3_pMatched',\n", " 't3_matched_simIdx',\n", + " 't3_sim_vxy',\n", + " 't3_sim_vz'\n", "]\n", "\n", "# Hit-dependent branches\n", "suffixes = ['r', 'z', 'eta', 'phi', 'layer']\n", "branches_list += [f't3_hit_{i}_{suffix}' for i in [0, 1, 2, 3, 4, 5] for suffix in suffixes]\n", "\n", - "file_path = \"t3_dnn_train_175.root\"\n", + "file_path = \"200_t3_dnn_relval_fix.root\"\n", "branches = load_root_file(file_path, branches_list)" ] }, @@ -61,7 +63,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "Z max: 224.14950561523438, R max: 98.93236541748047, Eta max: 2.5\n" + "Z max: 224.14950561523438, R max: 98.93299102783203, Eta max: 2.5\n" ] } ], @@ -165,7 +167,7 @@ "\n", "# Apply mask across all columns: retain a row only if all its entries are neither NaN nor Inf\n", "filtered_input_features_numpy = input_features_numpy[np.all(mask, axis=1)]\n", - "t5_isFake_filtered = (np.concatenate(branches['t3_pMatched']) < 0.9)[np.all(mask, axis=1)]\n", + "t3_isFake_filtered = (np.concatenate(branches['t3_pMatched']) < 0.75)[np.all(mask, axis=1)]\n", "\n", "# Convert to PyTorch tensor when ready to use with NN\n", "input_features_tensor = torch.tensor(filtered_input_features_numpy, dtype=torch.float32)" @@ -180,161 +182,161 @@ "name": "stdout", "output_type": "stream", "text": [ - "Using device: cuda\n", - "Initial dataset size: 15131951\n", - "Dataset size after initial 100.0% downsampling: 15131951\n", - "Class distribution after initial downsampling - Class 0: 13896802, Class 1: 1235149\n", - "Final class distribution after balancing - Class 0: 1235149, Class 1: 1235149\n", - "Epoch [1/150], Loss: 0.2612, Test Acc: 89.49%\n", - "Epoch [2/150], Loss: 0.2496, Test Acc: 91.23%\n", - "Epoch [3/150], Loss: 0.2351, Test Acc: 92.18%\n", - "Epoch [4/150], Loss: 0.1740, Test Acc: 92.70%\n", - "Epoch [5/150], Loss: 0.2017, Test Acc: 93.42%\n", - "Epoch [6/150], Loss: 0.1938, Test Acc: 93.49%\n", - "Epoch [7/150], Loss: 0.1705, Test Acc: 93.64%\n", - "Epoch [8/150], Loss: 0.1661, Test Acc: 93.77%\n", - "Epoch [9/150], Loss: 0.1995, Test Acc: 93.77%\n", - "Epoch [10/150], Loss: 0.1610, Test Acc: 93.71%\n", - "Epoch [11/150], Loss: 0.1647, Test Acc: 93.94%\n", - "Epoch [12/150], Loss: 0.1356, Test Acc: 93.88%\n", - "Epoch [13/150], Loss: 0.1737, Test Acc: 93.91%\n", - "Epoch [14/150], Loss: 0.1596, Test Acc: 93.91%\n", - "Epoch [15/150], Loss: 0.1907, Test Acc: 93.97%\n", - "Epoch [16/150], Loss: 0.2036, Test Acc: 94.02%\n", - "Epoch [17/150], Loss: 0.1486, Test Acc: 93.96%\n", - "Epoch [18/150], Loss: 0.1830, Test Acc: 94.00%\n", - "Epoch [19/150], Loss: 0.1912, Test Acc: 93.86%\n", - "Epoch [20/150], Loss: 0.1674, Test Acc: 94.03%\n", - "Epoch [21/150], Loss: 0.1751, Test Acc: 93.84%\n", - "Epoch [22/150], Loss: 0.1354, Test Acc: 93.87%\n", - "Epoch [23/150], Loss: 0.1845, Test Acc: 93.99%\n", - "Epoch [24/150], Loss: 0.1562, Test Acc: 94.14%\n", - "Epoch [25/150], Loss: 0.1947, Test Acc: 93.74%\n", - "Epoch [26/150], Loss: 0.1526, Test Acc: 94.03%\n", - "Epoch [27/150], Loss: 0.1891, Test Acc: 94.14%\n", - "Epoch [28/150], Loss: 0.1824, Test Acc: 94.09%\n", - "Epoch [29/150], Loss: 0.1931, Test Acc: 94.13%\n", - "Epoch [30/150], Loss: 0.1595, Test Acc: 93.95%\n", - "Epoch [31/150], Loss: 0.1625, Test Acc: 94.10%\n", - "Epoch [32/150], Loss: 0.1604, Test Acc: 94.07%\n", - "Epoch [33/150], Loss: 0.1826, Test Acc: 94.03%\n", - "Epoch [34/150], Loss: 0.1632, Test Acc: 93.78%\n", - "Epoch [35/150], Loss: 0.1870, Test Acc: 93.95%\n", - "Epoch [36/150], Loss: 0.1606, Test Acc: 94.01%\n", - "Epoch [37/150], Loss: 0.1368, Test Acc: 94.04%\n", - "Epoch [38/150], Loss: 0.1544, Test Acc: 94.15%\n", - "Epoch [39/150], Loss: 0.1447, Test Acc: 94.14%\n", - "Epoch [40/150], Loss: 0.1664, Test Acc: 94.16%\n", - "Epoch [41/150], Loss: 0.1780, Test Acc: 93.99%\n", - "Epoch [42/150], Loss: 0.1573, Test Acc: 94.16%\n", - "Epoch [43/150], Loss: 0.1633, Test Acc: 94.08%\n", - "Epoch [44/150], Loss: 0.1728, Test Acc: 94.17%\n", - "Epoch [45/150], Loss: 0.1685, Test Acc: 94.20%\n", - "Epoch [46/150], Loss: 0.2017, Test Acc: 93.97%\n", - "Epoch [47/150], Loss: 0.1557, Test Acc: 94.03%\n", - "Epoch [48/150], Loss: 0.1593, Test Acc: 94.23%\n", - "Epoch [49/150], Loss: 0.1554, Test Acc: 94.16%\n", - "Epoch [50/150], Loss: 0.1530, Test Acc: 94.16%\n", - "Epoch [51/150], Loss: 0.1641, Test Acc: 94.25%\n", - "Epoch [52/150], Loss: 0.1831, Test Acc: 94.24%\n", - "Epoch [53/150], Loss: 0.1391, Test Acc: 94.17%\n", - "Epoch [54/150], Loss: 0.1704, Test Acc: 94.19%\n", - "Epoch [55/150], Loss: 0.1355, Test Acc: 94.18%\n", - "Epoch [56/150], Loss: 0.1590, Test Acc: 94.21%\n", - "Epoch [57/150], Loss: 0.1926, Test Acc: 94.27%\n", - "Epoch [58/150], Loss: 0.1704, Test Acc: 94.28%\n", - "Epoch [59/150], Loss: 0.1541, Test Acc: 94.08%\n", - "Epoch [60/150], Loss: 0.1755, Test Acc: 94.31%\n", - "Epoch [61/150], Loss: 0.1585, Test Acc: 94.16%\n", - "Epoch [62/150], Loss: 0.1579, Test Acc: 94.21%\n", - "Epoch [63/150], Loss: 0.1508, Test Acc: 94.12%\n", - "Epoch [64/150], Loss: 0.1600, Test Acc: 94.31%\n", - "Epoch [65/150], Loss: 0.1678, Test Acc: 94.32%\n", - "Epoch [66/150], Loss: 0.1572, Test Acc: 94.30%\n", - "Epoch [67/150], Loss: 0.1731, Test Acc: 94.24%\n", - "Epoch [68/150], Loss: 0.1651, Test Acc: 93.88%\n", - "Epoch [69/150], Loss: 0.1626, Test Acc: 94.10%\n", - "Epoch [70/150], Loss: 0.1486, Test Acc: 94.13%\n", - "Epoch [71/150], Loss: 0.1659, Test Acc: 93.81%\n", - "Epoch [72/150], Loss: 0.1617, Test Acc: 93.97%\n", - "Epoch [73/150], Loss: 0.1658, Test Acc: 94.17%\n", - "Epoch [74/150], Loss: 0.1795, Test Acc: 93.94%\n", - "Epoch [75/150], Loss: 0.1849, Test Acc: 94.26%\n", - "Epoch [76/150], Loss: 0.1364, Test Acc: 94.30%\n", - "Epoch [77/150], Loss: 0.1865, Test Acc: 94.08%\n", - "Epoch [78/150], Loss: 0.1890, Test Acc: 94.24%\n", - "Epoch [79/150], Loss: 0.1618, Test Acc: 94.24%\n", - "Epoch [80/150], Loss: 0.1703, Test Acc: 94.31%\n", - "Epoch [81/150], Loss: 0.1239, Test Acc: 94.28%\n", - "Epoch [82/150], Loss: 0.1638, Test Acc: 94.38%\n", - "Epoch [83/150], Loss: 0.1834, Test Acc: 94.37%\n", - "Epoch [84/150], Loss: 0.1581, Test Acc: 94.19%\n", - "Epoch [85/150], Loss: 0.1982, Test Acc: 94.29%\n", - "Epoch [86/150], Loss: 0.1707, Test Acc: 94.31%\n", - "Epoch [87/150], Loss: 0.1748, Test Acc: 94.28%\n", - "Epoch [88/150], Loss: 0.1615, Test Acc: 94.27%\n", - "Epoch [89/150], Loss: 0.1569, Test Acc: 94.28%\n", - "Epoch [90/150], Loss: 0.1377, Test Acc: 94.29%\n", - "Epoch [91/150], Loss: 0.1430, Test Acc: 94.33%\n", - "Epoch [92/150], Loss: 0.1575, Test Acc: 93.93%\n", - "Epoch [93/150], Loss: 0.1722, Test Acc: 94.36%\n", - "Epoch [94/150], Loss: 0.1522, Test Acc: 94.36%\n", - "Epoch [95/150], Loss: 0.1474, Test Acc: 94.39%\n", - "Epoch [96/150], Loss: 0.1479, Test Acc: 94.25%\n", - "Epoch [97/150], Loss: 0.1311, Test Acc: 94.31%\n", - "Epoch [98/150], Loss: 0.1569, Test Acc: 94.31%\n", - "Epoch [99/150], Loss: 0.1634, Test Acc: 94.35%\n", - "Epoch [100/150], Loss: 0.1541, Test Acc: 94.38%\n", - "Epoch [101/150], Loss: 0.1606, Test Acc: 94.30%\n", - "Epoch [102/150], Loss: 0.1845, Test Acc: 94.26%\n", - "Epoch [103/150], Loss: 0.1361, Test Acc: 94.26%\n", - "Epoch [104/150], Loss: 0.1923, Test Acc: 94.40%\n", - "Epoch [105/150], Loss: 0.1552, Test Acc: 94.27%\n", - "Epoch [106/150], Loss: 0.1352, Test Acc: 94.36%\n", - "Epoch [107/150], Loss: 0.1963, Test Acc: 94.21%\n", - "Epoch [108/150], Loss: 0.1819, Test Acc: 94.41%\n", - "Epoch [109/150], Loss: 0.1593, Test Acc: 94.30%\n", - "Epoch [110/150], Loss: 0.1741, Test Acc: 94.16%\n", - "Epoch [111/150], Loss: 0.1899, Test Acc: 94.35%\n", - "Epoch [112/150], Loss: 0.1653, Test Acc: 94.32%\n", - "Epoch [113/150], Loss: 0.1597, Test Acc: 94.34%\n", - "Epoch [114/150], Loss: 0.1502, Test Acc: 94.40%\n", - "Epoch [115/150], Loss: 0.1510, Test Acc: 94.36%\n", - "Epoch [116/150], Loss: 0.1306, Test Acc: 94.41%\n", - "Epoch [117/150], Loss: 0.1805, Test Acc: 94.38%\n", - "Epoch [118/150], Loss: 0.1518, Test Acc: 94.15%\n", - "Epoch [119/150], Loss: 0.2202, Test Acc: 94.32%\n", - "Epoch [120/150], Loss: 0.1514, Test Acc: 94.27%\n", - "Epoch [121/150], Loss: 0.1433, Test Acc: 94.35%\n", - "Epoch [122/150], Loss: 0.1583, Test Acc: 94.41%\n", - "Epoch [123/150], Loss: 0.1809, Test Acc: 94.36%\n", - "Epoch [124/150], Loss: 0.1712, Test Acc: 94.40%\n", - "Epoch [125/150], Loss: 0.1765, Test Acc: 94.41%\n", - "Epoch [126/150], Loss: 0.1582, Test Acc: 94.38%\n", - "Epoch [127/150], Loss: 0.1797, Test Acc: 94.26%\n", - "Epoch [128/150], Loss: 0.1559, Test Acc: 94.25%\n", - "Epoch [129/150], Loss: 0.1222, Test Acc: 94.31%\n", - "Epoch [130/150], Loss: 0.1351, Test Acc: 94.42%\n", - "Epoch [131/150], Loss: 0.1282, Test Acc: 94.38%\n", - "Epoch [132/150], Loss: 0.1631, Test Acc: 94.44%\n", - "Epoch [133/150], Loss: 0.1531, Test Acc: 94.18%\n", - "Epoch [134/150], Loss: 0.1375, Test Acc: 94.45%\n", - "Epoch [135/150], Loss: 0.1421, Test Acc: 94.43%\n", - "Epoch [136/150], Loss: 0.1712, Test Acc: 93.80%\n", - "Epoch [137/150], Loss: 0.1655, Test Acc: 94.41%\n", - "Epoch [138/150], Loss: 0.1699, Test Acc: 94.34%\n", - "Epoch [139/150], Loss: 0.1899, Test Acc: 94.36%\n", - "Epoch [140/150], Loss: 0.1488, Test Acc: 94.40%\n", - "Epoch [141/150], Loss: 0.1621, Test Acc: 94.36%\n", - "Epoch [142/150], Loss: 0.1432, Test Acc: 94.33%\n", - "Epoch [143/150], Loss: 0.1767, Test Acc: 94.27%\n", - "Epoch [144/150], Loss: 0.1421, Test Acc: 94.46%\n", - "Epoch [145/150], Loss: 0.1765, Test Acc: 94.42%\n", - "Epoch [146/150], Loss: 0.1434, Test Acc: 94.36%\n", - "Epoch [147/150], Loss: 0.1741, Test Acc: 94.39%\n", - "Epoch [148/150], Loss: 0.1763, Test Acc: 94.30%\n", - "Epoch [149/150], Loss: 0.2035, Test Acc: 94.36%\n", - "Epoch [150/150], Loss: 0.1627, Test Acc: 94.44%\n" + "Using device: cpu\n", + "Initial dataset size: 18366741\n", + "Dataset size after initial 100.0% downsampling: 18366741\n", + "Class distribution after initial downsampling - Class 0: 16624486, Class 1: 1742255\n", + "Final class distribution after balancing - Class 0: 1742255, Class 1: 1742255\n", + "Epoch [1/150], Loss: 0.3424, Test Acc: 90.87%\n", + "Epoch [2/150], Loss: 0.2061, Test Acc: 92.43%\n", + "Epoch [3/150], Loss: 0.2207, Test Acc: 93.48%\n", + "Epoch [4/150], Loss: 0.1377, Test Acc: 93.69%\n", + "Epoch [5/150], Loss: 0.2022, Test Acc: 93.73%\n", + "Epoch [6/150], Loss: 0.1242, Test Acc: 93.97%\n", + "Epoch [7/150], Loss: 0.1772, Test Acc: 93.98%\n", + "Epoch [8/150], Loss: 0.1987, Test Acc: 93.90%\n", + "Epoch [9/150], Loss: 0.1600, Test Acc: 94.15%\n", + "Epoch [10/150], Loss: 0.1941, Test Acc: 93.83%\n", + "Epoch [11/150], Loss: 0.1946, Test Acc: 94.18%\n", + "Epoch [12/150], Loss: 0.2152, Test Acc: 93.76%\n", + "Epoch [13/150], Loss: 0.1469, Test Acc: 93.85%\n", + "Epoch [14/150], Loss: 0.1494, Test Acc: 94.17%\n", + "Epoch [15/150], Loss: 0.1263, Test Acc: 94.27%\n", + "Epoch [16/150], Loss: 0.1233, Test Acc: 94.29%\n", + "Epoch [17/150], Loss: 0.1259, Test Acc: 94.25%\n", + "Epoch [18/150], Loss: 0.1444, Test Acc: 94.28%\n", + "Epoch [19/150], Loss: 0.1305, Test Acc: 94.12%\n", + "Epoch [20/150], Loss: 0.1436, Test Acc: 94.26%\n", + "Epoch [21/150], Loss: 0.1249, Test Acc: 94.34%\n", + "Epoch [22/150], Loss: 0.1338, Test Acc: 94.29%\n", + "Epoch [23/150], Loss: 0.1914, Test Acc: 93.89%\n", + "Epoch [24/150], Loss: 0.1653, Test Acc: 94.14%\n", + "Epoch [25/150], Loss: 0.1550, Test Acc: 94.09%\n", + "Epoch [26/150], Loss: 0.1382, Test Acc: 94.27%\n", + "Epoch [27/150], Loss: 0.1392, Test Acc: 94.33%\n", + "Epoch [28/150], Loss: 0.1839, Test Acc: 94.03%\n", + "Epoch [29/150], Loss: 0.1620, Test Acc: 94.34%\n", + "Epoch [30/150], Loss: 0.1780, Test Acc: 94.14%\n", + "Epoch [31/150], Loss: 0.1233, Test Acc: 94.24%\n", + "Epoch [32/150], Loss: 0.1800, Test Acc: 94.22%\n", + "Epoch [33/150], Loss: 0.1882, Test Acc: 94.24%\n", + "Epoch [34/150], Loss: 0.1088, Test Acc: 94.28%\n", + "Epoch [35/150], Loss: 0.1633, Test Acc: 94.04%\n", + "Epoch [36/150], Loss: 0.1678, Test Acc: 94.26%\n", + "Epoch [37/150], Loss: 0.1865, Test Acc: 94.42%\n", + "Epoch [38/150], Loss: 0.1413, Test Acc: 94.32%\n", + "Epoch [39/150], Loss: 0.2067, Test Acc: 94.20%\n", + "Epoch [40/150], Loss: 0.1876, Test Acc: 94.41%\n", + "Epoch [41/150], Loss: 0.1121, Test Acc: 94.38%\n", + "Epoch [42/150], Loss: 0.1637, Test Acc: 94.44%\n", + "Epoch [43/150], Loss: 0.1130, Test Acc: 94.42%\n", + "Epoch [44/150], Loss: 0.1650, Test Acc: 94.29%\n", + "Epoch [45/150], Loss: 0.1995, Test Acc: 94.35%\n", + "Epoch [46/150], Loss: 0.1533, Test Acc: 94.45%\n", + "Epoch [47/150], Loss: 0.1788, Test Acc: 94.41%\n", + "Epoch [48/150], Loss: 0.1505, Test Acc: 94.48%\n", + "Epoch [49/150], Loss: 0.1835, Test Acc: 94.36%\n", + "Epoch [50/150], Loss: 0.1037, Test Acc: 94.43%\n", + "Epoch [51/150], Loss: 0.1583, Test Acc: 94.47%\n", + "Epoch [52/150], Loss: 0.1640, Test Acc: 94.40%\n", + "Epoch [53/150], Loss: 0.1782, Test Acc: 94.37%\n", + "Epoch [54/150], Loss: 0.1510, Test Acc: 94.37%\n", + "Epoch [55/150], Loss: 0.1421, Test Acc: 94.36%\n", + "Epoch [56/150], Loss: 0.1372, Test Acc: 94.43%\n", + "Epoch [57/150], Loss: 0.2006, Test Acc: 94.14%\n", + "Epoch [58/150], Loss: 0.1841, Test Acc: 94.50%\n", + "Epoch [59/150], Loss: 0.1700, Test Acc: 94.37%\n", + "Epoch [60/150], Loss: 0.1414, Test Acc: 94.51%\n", + "Epoch [61/150], Loss: 0.1405, Test Acc: 94.44%\n", + "Epoch [62/150], Loss: 0.2052, Test Acc: 94.41%\n", + "Epoch [63/150], Loss: 0.0882, Test Acc: 94.48%\n", + "Epoch [64/150], Loss: 0.1823, Test Acc: 94.46%\n", + "Epoch [65/150], Loss: 0.1257, Test Acc: 94.45%\n", + "Epoch [66/150], Loss: 0.1584, Test Acc: 94.48%\n", + "Epoch [67/150], Loss: 0.1542, Test Acc: 94.42%\n", + "Epoch [68/150], Loss: 0.1112, Test Acc: 94.41%\n", + "Epoch [69/150], Loss: 0.1157, Test Acc: 94.55%\n", + "Epoch [70/150], Loss: 0.1528, Test Acc: 94.49%\n", + "Epoch [71/150], Loss: 0.1896, Test Acc: 94.45%\n", + "Epoch [72/150], Loss: 0.1444, Test Acc: 94.34%\n", + "Epoch [73/150], Loss: 0.2029, Test Acc: 94.46%\n", + "Epoch [74/150], Loss: 0.1632, Test Acc: 94.42%\n", + "Epoch [75/150], Loss: 0.1777, Test Acc: 94.21%\n", + "Epoch [76/150], Loss: 0.1635, Test Acc: 94.42%\n", + "Epoch [77/150], Loss: 0.1849, Test Acc: 94.38%\n", + "Epoch [78/150], Loss: 0.1142, Test Acc: 94.51%\n", + "Epoch [79/150], Loss: 0.1880, Test Acc: 94.23%\n", + "Epoch [80/150], Loss: 0.1703, Test Acc: 94.54%\n", + "Epoch [81/150], Loss: 0.1926, Test Acc: 94.57%\n", + "Epoch [82/150], Loss: 0.1641, Test Acc: 94.44%\n", + "Epoch [83/150], Loss: 0.1645, Test Acc: 94.45%\n", + "Epoch [84/150], Loss: 0.1519, Test Acc: 94.45%\n", + "Epoch [85/150], Loss: 0.1326, Test Acc: 94.48%\n", + "Epoch [86/150], Loss: 0.1182, Test Acc: 94.36%\n", + "Epoch [87/150], Loss: 0.1002, Test Acc: 94.54%\n", + "Epoch [88/150], Loss: 0.1484, Test Acc: 94.44%\n", + "Epoch [89/150], Loss: 0.1713, Test Acc: 94.24%\n", + "Epoch [90/150], Loss: 0.1931, Test Acc: 94.48%\n", + "Epoch [91/150], Loss: 0.1486, Test Acc: 94.48%\n", + "Epoch [92/150], Loss: 0.1814, Test Acc: 94.56%\n", + "Epoch [93/150], Loss: 0.1772, Test Acc: 94.32%\n", + "Epoch [94/150], Loss: 0.1537, Test Acc: 94.47%\n", + "Epoch [95/150], Loss: 0.2322, Test Acc: 94.53%\n", + "Epoch [96/150], Loss: 0.1302, Test Acc: 94.50%\n", + "Epoch [97/150], Loss: 0.1553, Test Acc: 94.57%\n", + "Epoch [98/150], Loss: 0.1736, Test Acc: 94.43%\n", + "Epoch [99/150], Loss: 0.1642, Test Acc: 94.47%\n", + "Epoch [100/150], Loss: 0.2049, Test Acc: 94.41%\n", + "Epoch [101/150], Loss: 0.1515, Test Acc: 94.57%\n", + "Epoch [102/150], Loss: 0.1156, Test Acc: 94.50%\n", + "Epoch [103/150], Loss: 0.1173, Test Acc: 94.60%\n", + "Epoch [104/150], Loss: 0.1272, Test Acc: 94.47%\n", + "Epoch [105/150], Loss: 0.1126, Test Acc: 94.47%\n", + "Epoch [106/150], Loss: 0.2333, Test Acc: 94.50%\n", + "Epoch [107/150], Loss: 0.1720, Test Acc: 94.43%\n", + "Epoch [108/150], Loss: 0.1621, Test Acc: 94.22%\n", + "Epoch [109/150], Loss: 0.1451, Test Acc: 94.58%\n", + "Epoch [110/150], Loss: 0.1747, Test Acc: 94.42%\n", + "Epoch [111/150], Loss: 0.1361, Test Acc: 94.51%\n", + "Epoch [112/150], Loss: 0.1195, Test Acc: 94.55%\n", + "Epoch [113/150], Loss: 0.1503, Test Acc: 94.42%\n", + "Epoch [114/150], Loss: 0.2048, Test Acc: 94.56%\n", + "Epoch [115/150], Loss: 0.1824, Test Acc: 94.49%\n", + "Epoch [116/150], Loss: 0.1742, Test Acc: 94.50%\n", + "Epoch [117/150], Loss: 0.1500, Test Acc: 94.53%\n", + "Epoch [118/150], Loss: 0.1526, Test Acc: 94.33%\n", + "Epoch [119/150], Loss: 0.1414, Test Acc: 94.57%\n", + "Epoch [120/150], Loss: 0.1750, Test Acc: 94.48%\n", + "Epoch [121/150], Loss: 0.1907, Test Acc: 94.57%\n", + "Epoch [122/150], Loss: 0.0992, Test Acc: 94.42%\n", + "Epoch [123/150], Loss: 0.1828, Test Acc: 94.42%\n", + "Epoch [124/150], Loss: 0.1448, Test Acc: 94.59%\n", + "Epoch [125/150], Loss: 0.1930, Test Acc: 94.46%\n", + "Epoch [126/150], Loss: 0.1551, Test Acc: 94.57%\n", + "Epoch [127/150], Loss: 0.1176, Test Acc: 94.59%\n", + "Epoch [128/150], Loss: 0.1288, Test Acc: 94.55%\n", + "Epoch [129/150], Loss: 0.1850, Test Acc: 94.52%\n", + "Epoch [130/150], Loss: 0.1316, Test Acc: 94.25%\n", + "Epoch [131/150], Loss: 0.1279, Test Acc: 94.46%\n", + "Epoch [132/150], Loss: 0.1782, Test Acc: 94.58%\n", + "Epoch [133/150], Loss: 0.1183, Test Acc: 94.51%\n", + "Epoch [134/150], Loss: 0.0878, Test Acc: 94.57%\n", + "Epoch [135/150], Loss: 0.1704, Test Acc: 94.55%\n", + "Epoch [136/150], Loss: 0.1472, Test Acc: 94.54%\n", + "Epoch [137/150], Loss: 0.1073, Test Acc: 94.49%\n", + "Epoch [138/150], Loss: 0.1131, Test Acc: 94.56%\n", + "Epoch [139/150], Loss: 0.1376, Test Acc: 94.44%\n", + "Epoch [140/150], Loss: 0.0973, Test Acc: 94.58%\n", + "Epoch [141/150], Loss: 0.1670, Test Acc: 94.59%\n", + "Epoch [142/150], Loss: 0.1520, Test Acc: 94.58%\n", + "Epoch [143/150], Loss: 0.2283, Test Acc: 94.50%\n", + "Epoch [144/150], Loss: 0.1047, Test Acc: 94.55%\n", + "Epoch [145/150], Loss: 0.2174, Test Acc: 94.55%\n", + "Epoch [146/150], Loss: 0.1195, Test Acc: 94.44%\n", + "Epoch [147/150], Loss: 0.1501, Test Acc: 94.46%\n", + "Epoch [148/150], Loss: 0.1644, Test Acc: 94.27%\n", + "Epoch [149/150], Loss: 0.1812, Test Acc: 94.33%\n", + "Epoch [150/150], Loss: 0.1857, Test Acc: 94.48%\n" ] } ], @@ -349,7 +351,7 @@ "print(f\"Using device: {device}\")\n", "\n", "# Create labels tensor\n", - "labels_tensor = 1 - torch.tensor(t5_isFake_filtered, dtype=torch.float32)\n", + "labels_tensor = 1 - torch.tensor(t3_isFake_filtered, dtype=torch.float32)\n", "\n", "# Set initial downsample fraction\n", "initial_downsample_fraction = 1.0 # Adjust this value as needed\n", @@ -374,20 +376,39 @@ " def __init__(self):\n", " super(WeightedBCELoss, self).__init__()\n", " \n", - " def forward(self, outputs, targets):\n", + " def forward(self, outputs, targets, weights):\n", " eps = 1e-7\n", - " losses = -((targets * torch.log(outputs + eps) + \n", + " losses = -(weights * (targets * torch.log(outputs + eps) + \n", " (1 - targets) * torch.log(1 - outputs + eps)))\n", " return losses.mean()\n", "\n", + "def calculate_sample_weights(t3_sim_vxy, weight_factor=1.0):\n", + " \"\"\"\n", + " Calculate sample weights giving higher importance to displaced t5's\n", + " \n", + " Args:\n", + " t3_sim_vxy: Array of t5 simulation values\n", + " weight_factor: How much more weight to give to displaced samples\n", + " \n", + " Returns:\n", + " Tensor of sample weights\n", + " \"\"\"\n", + " weights = torch.ones(len(t3_sim_vxy))\n", + " displaced_mask = t3_sim_vxy > 0.1\n", + " weights[displaced_mask] = weight_factor\n", + " return weights\n", "\n", "# Print initial dataset size\n", "print(f\"Initial dataset size: {len(labels_tensor)}\")\n", "\n", + "# Calculate sample weights\n", + "sample_weights = calculate_sample_weights(torch.tensor(np.concatenate(branches['t3_sim_vxy'])))\n", + "\n", "# Remove rows with NaN and update weights accordingly\n", "nan_mask = torch.isnan(input_features_tensor).any(dim=1) | torch.isnan(labels_tensor)\n", "filtered_inputs = input_features_tensor[~nan_mask]\n", "filtered_labels = labels_tensor[~nan_mask]\n", + "filtered_weights = sample_weights[~nan_mask]\n", "\n", "# Initial downsampling of entire dataset\n", "if initial_downsample_fraction < 1.0:\n", @@ -396,6 +417,7 @@ " indices = torch.randperm(total_samples)[:samples_to_keep]\n", " filtered_inputs = filtered_inputs[indices]\n", " filtered_labels = filtered_labels[indices]\n", + " filtered_weights = filtered_weights[indices]\n", "\n", "print(f\"Dataset size after initial {initial_downsample_fraction*100}% downsampling: {len(filtered_labels)}\")\n", "\n", @@ -413,13 +435,14 @@ "# Create balanced dataset with weights\n", "balanced_inputs = filtered_inputs[balanced_indices]\n", "balanced_labels = filtered_labels[balanced_indices]\n", + "balanced_weights = filtered_weights[balanced_indices]\n", "\n", "# Verify balanced distribution\n", "balanced_counts = torch.bincount(balanced_labels.int())\n", "print(f\"Final class distribution after balancing - Class 0: {balanced_counts[0]}, Class 1: {balanced_counts[1]}\")\n", "\n", "# Create dataset with weights\n", - "dataset = TensorDataset(balanced_inputs, balanced_labels)\n", + "dataset = TensorDataset(balanced_inputs, balanced_labels, balanced_weights)\n", "\n", "# Split into train and test sets\n", "train_size = int(0.8 * len(dataset))\n", @@ -440,7 +463,7 @@ " total = 0\n", " correct = 0\n", " with torch.no_grad():\n", - " for inputs, targets in loader:\n", + " for inputs, targets, weights in loader:\n", " inputs, targets = inputs.to(device), targets.to(device)\n", " outputs = model(inputs)\n", " predicted = outputs.squeeze() > 0.5\n", @@ -454,12 +477,12 @@ "loss_log = []\n", "\n", "for epoch in range(num_epochs):\n", - " for inputs, targets in train_loader:\n", - " inputs, targets = inputs.to(device), targets.to(device)\n", + " for inputs, targets, weights in train_loader:\n", + " inputs, targets, weights = inputs.to(device), targets.to(device), weights.to(device)\n", " \n", " # Forward pass\n", " outputs = model(inputs)\n", - " loss = loss_function(outputs.squeeze(), targets)\n", + " loss = loss_function(outputs.squeeze(), targets, weights)\n", " \n", " loss_log.append(loss.item())\n", "\n", @@ -490,22 +513,22 @@ "name": "stdout", "output_type": "stream", "text": [ - "Baseline accuracy: 0.9399382472038269\n", + "Baseline accuracy: 0.9416274428367615\n", "Feature importances:\n", - "Feature 13 importance: 0.0324\n", - "Feature 9 importance: 0.0289\n", - "Feature 8 importance: 0.0175\n", - "Feature 4 importance: 0.0169\n", - "Feature 0 importance: 0.0092\n", - "Feature 12 importance: 0.0072\n", - "Feature 3 importance: 0.0037\n", - "Feature 6 importance: 0.0036\n", - "Feature 2 importance: 0.0032\n", - "Feature 10 importance: 0.0012\n", - "Feature 5 importance: 0.0011\n", - "Feature 11 importance: 0.0002\n", - "Feature 1 importance: -0.0000\n", - "Feature 7 importance: -0.0005\n" + "Feature 13 importance: 0.0447\n", + "Feature 9 importance: 0.0308\n", + "Feature 2 importance: 0.0236\n", + "Feature 0 importance: 0.0187\n", + "Feature 4 importance: 0.0150\n", + "Feature 5 importance: 0.0130\n", + "Feature 8 importance: 0.0127\n", + "Feature 3 importance: 0.0079\n", + "Feature 12 importance: 0.0061\n", + "Feature 6 importance: 0.0057\n", + "Feature 10 importance: 0.0026\n", + "Feature 7 importance: 0.0007\n", + "Feature 11 importance: 0.0004\n", + "Feature 1 importance: 0.0000\n" ] } ], @@ -559,13 +582,13 @@ "name": "stderr", "output_type": "stream", "text": [ - "/tmp/ipykernel_1791539/52354147.py:7: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n", + "/tmp/ipykernel_234858/52354147.py:7: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n", " inputs = torch.tensor(features, dtype=torch.float32).to(device)\n" ] }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -610,105 +633,165 @@ "cell_type": "code", "execution_count": 9, "metadata": {}, + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/tmp/ipykernel_234858/52354147.py:7: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n", + " inputs = torch.tensor(features, dtype=torch.float32).to(device)\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Prediction scores for displaced tracks (t3_sim_vxy > 0.1):\n", + "Mean score: 0.7552\n", + "Median score: 0.9132\n" + ] + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "# Get model predictions\n", + "probabilities = model_outputs(filtered_inputs, model)\n", + "\n", + "# Get displaced track mask\n", + "displaced_mask = np.concatenate(branches['t3_sim_vxy'])[~nan_mask] > 0.1\n", + "\n", + "# Calculate statistics for displaced tracks\n", + "displaced_predictions = probabilities[displaced_mask]\n", + "mean_score = np.mean(displaced_predictions)\n", + "median_score = np.median(displaced_predictions)\n", + "\n", + "print(f\"Prediction scores for displaced tracks (t3_sim_vxy > 0.1):\")\n", + "print(f\"Mean score: {mean_score:.4f}\")\n", + "print(f\"Median score: {median_score:.4f}\")\n", + "\n", + "plt.hist(displaced_predictions, bins=100, histtype='step', linewidth=1.5) # Outline only, no fill\n", + "plt.yscale('log')\n", + "plt.xlabel(\"DNN Score\", fontsize=12)\n", + "plt.ylabel(\"Frequency (log scale)\", fontsize=12)\n", + "plt.title(\"DNN Score for Displaced T5s\", fontsize=14, weight='bold')\n", + "\n", + "plt.grid(visible=True, which='both', linestyle='--', linewidth=0.5)\n", + "plt.tight_layout()\n", + "plt.show()" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_layer1[32] = {\n", - "-0.1918962f, -1.0575372f, -0.8276564f, -0.0243965f, -0.1577621f, 1.0067693f, 1.5348158f, 0.4439710f, 0.0041234f, -1.1558943f, -1.4180470f, 1.0221841f, -0.0592227f, -1.2107433f, -0.2100758f, 1.2193928f, -0.3124787f, -1.9197327f, -0.8064887f, -0.2178766f, -0.0111392f, -0.1638742f, 0.0029338f, -0.0157688f, 0.2662797f, 1.8194629f, 0.8465537f, -0.7592145f, -0.8783396f, 0.5602613f, -0.0764334f, -0.8502049f };\n", + "-0.2400621f, 1.3418640f, -0.1700544f, -0.1809052f, -0.0194816f, -1.6636838f, -0.2450859f, -1.1127867f, 0.1477622f, -0.0740084f, -0.8851718f, -0.2633939f, 0.1941925f, 0.9132360f, 1.5794699f, -0.7953343f, -0.2615424f, 0.2003026f, -2.2417912f, -0.8144404f, -1.2980548f, -0.1407361f, -0.0966190f, -0.2355828f, -0.3362100f, 0.8529679f, -0.0075295f, -0.0501248f, -0.8618563f, -0.7658819f, 0.8201296f, -0.0064448f };\n", "\n", "ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_layer1[14][32] = {\n", - "{ -0.2258795f, 0.4783621f, 0.4048500f, -0.2535419f, -0.2204617f, 0.2292106f, 0.5185162f, 1.0356801f, -0.1940614f, 0.7601448f, -0.2762348f, 0.4394473f, -0.2472922f, 0.6392686f, 0.0174135f, 1.1337029f, 0.5831500f, -0.5536784f, 0.1608927f, 0.3617360f, -0.0014275f, 0.1839313f, -0.1858799f, 0.0080405f, 0.1321980f, 0.4460649f, 0.4296648f, 0.4569599f, -0.2347775f, -0.3548780f, 0.1986685f, 0.2211701f },\n", - "{ 0.1509047f, 0.0057684f, -0.0019561f, -0.1210250f, -0.2612511f, -0.0098326f, 0.0019430f, 0.0152595f, -0.2313585f, 0.0117159f, -0.0087940f, -0.0000195f, 0.2194081f, -0.0033250f, 0.1478879f, 0.0097880f, -0.0041943f, 0.0362815f, -0.0197458f, -0.0063192f, -0.0004957f, -0.0104775f, -0.2756116f, -0.0049759f, -0.0015340f, -0.0111502f, 0.0034782f, -0.0100520f, 0.0124440f, 0.0076567f, -0.0263710f, 0.0381163f },\n", - "{ 0.1448244f, 0.0803795f, 1.6524559f, 0.1457684f, -0.1350329f, -3.6343203f, -1.6798589f, -0.0142065f, 0.0835910f, 0.0497492f, 0.6378320f, -0.4540107f, -0.2349942f, -0.0257038f, -0.2397820f, -4.3412380f, 0.1279643f, 0.9793525f, -0.5925660f, -0.8705363f, 0.0134403f, -0.1410540f, -0.0032170f, 0.0107955f, -1.5223076f, -0.1977515f, -0.9901226f, -0.5315630f, 0.5766137f, 0.8363163f, -0.2219186f, -1.3835622f },\n", - "{ -0.1057612f, 0.8005342f, -1.4052893f, -0.1196175f, 0.1360446f, 1.2852736f, -4.4616752f, 0.3914140f, -0.2356829f, 1.0064709f, 1.5389245f, -3.2000272f, 0.0828165f, 0.8944567f, -0.1592458f, 0.5353487f, 0.8051971f, 0.2057788f, 1.7512299f, 0.9215795f, -0.0036782f, 2.1131773f, -0.3204709f, -0.0044941f, 0.0865040f, 0.4481153f, -1.2256490f, 1.7857696f, -0.6445597f, -0.7477201f, 0.2316555f, -0.7155212f },\n", - "{ 0.0901890f, -1.1808448f, 1.4875709f, -0.0262624f, -0.1323451f, -0.8950851f, 0.7974008f, -2.7021067f, 0.0281707f, -1.3674084f, -4.0485940f, 5.8682752f, -0.1891649f, 7.2682476f, 0.1300428f, -2.2845411f, -8.7903214f, 0.6268425f, -1.9901284f, -2.5285044f, 0.0066100f, 7.4012928f, -0.0831600f, 0.0095476f, 6.1734300f, 0.8520991f, -8.3464308f, 2.4261341f, 3.6403832f, 5.9824433f, -0.0999645f, -0.2824311f },\n", - "{ -0.0335586f, 16.4833145f, 1.4688981f, -0.1789742f, 0.2243387f, 0.1425887f, 0.1262731f, -4.6265879f, -0.0683203f, -12.3650198f, 1.3732860f, -0.2777069f, -0.0603657f, 0.8076705f, 0.0482012f, -0.7798629f, -1.6209395f, -0.0720773f, -0.3990867f, -1.0641636f, -0.0139909f, -0.5711431f, -0.0209516f, 0.0946307f, 0.0096234f, 0.7337275f, 0.4466013f, 1.5696099f, -5.6807351f, -1.6136186f, -0.2170101f, -0.9905877f },\n", - "{ -0.0965901f, 0.4846557f, -0.0652100f, -0.2353250f, -0.1215300f, -3.5360579f, -2.0285676f, -2.8374794f, -0.1559567f, 0.7083026f, -0.1030692f, -1.1805813f, 0.0706595f, -0.2951699f, -0.0989490f, -2.8418458f, 0.8427116f, -2.2046885f, -1.3581268f, 1.5271128f, -0.0045855f, 0.1438956f, -0.2343508f, -0.0425132f, 1.0763166f, 1.0795754f, 0.4891663f, 2.6029303f, -1.0363307f, 2.2133234f, -0.2510982f, 1.1560025f },\n", - "{ 0.0074010f, -0.4152716f, 3.4255989f, -0.1664258f, 0.2028382f, 1.2840286f, 1.3538148f, -1.5683322f, 0.0429628f, -0.2171310f, 2.8748064f, -0.5141454f, 0.1099870f, -0.4614673f, -0.2515875f, -0.3862101f, 1.6164609f, -1.3735241f, -0.8896182f, 1.8310896f, 0.0188789f, -0.8612124f, -0.2143756f, 0.0385537f, -8.3583250f, 1.3280702f, -7.0806746f, -0.2001765f, 0.8213225f, -2.8566475f, 0.1479015f, 0.5155560f },\n", - "{ 0.0511283f, -1.9676421f, 4.3773866f, -0.1648336f, 0.0505374f, -3.0326672f, 2.5867159f, 1.8804691f, -0.1412510f, -2.3595653f, 3.9877729f, -14.6059513f, -0.0444889f, 7.5017557f, -0.0684788f, 1.0775388f, -8.0584450f, -0.2013538f, -12.5508022f, 9.2100115f, -0.0011293f, -7.0289955f, 0.0521985f, 0.0155513f, -1.6320782f, -4.0694451f, 7.8784938f, 8.9188061f, 10.7590771f, -9.3559170f, -0.1319706f, -0.0336969f },\n", - "{ -0.1288323f, 7.9389300f, -0.4022054f, 0.0851088f, -0.2593997f, -0.3426172f, 0.3062155f, 1.9866818f, -0.1748946f, -9.7667055f, -0.6524503f, 0.3385161f, 0.1884300f, 9.0051126f, -0.2027990f, 0.3781536f, -0.1085841f, -5.2012277f, -0.5169301f, 2.4512987f, -14.8858967f, 0.8265918f, 0.1330098f, 14.9286051f, 0.4946520f, 1.2087970f, -1.8545671f, 0.7865057f, -3.1178455f, 2.2735860f, 0.1769712f, 0.7800660f },\n", - "{ -0.0041451f, -0.3216657f, 0.2317746f, -0.0770410f, -0.1181624f, -1.6830167f, 1.3180428f, 1.7278676f, -0.1011897f, -0.8716651f, 0.8045275f, 1.0895320f, 0.0222730f, -1.0280910f, 0.1943782f, -3.5416641f, -1.8209232f, -0.6835734f, 1.1706284f, -1.0224590f, -0.0421350f, 0.6345906f, -0.2030822f, -0.0540403f, 1.6704807f, 0.2792225f, 0.5046398f, 1.1982751f, 0.1087174f, -0.6976060f, 0.1717374f, 1.3827170f },\n", - "{ 0.0014786f, -1.5822506f, -1.4512368f, 0.1379846f, -0.0946356f, 1.0877693f, -1.4706703f, 0.5806997f, 0.1940563f, -1.1956979f, 0.9592837f, -0.8354632f, 0.2176267f, -0.7358339f, -0.0474411f, 0.2848974f, -0.6754610f, -0.9115687f, 2.1276686f, -1.0713644f, -0.0001056f, 1.8921490f, -0.0347501f, 0.0047789f, 1.3341833f, 1.8055003f, 1.1767987f, -3.9586561f, -0.6598469f, -0.3037908f, -0.1673333f, 0.2427263f },\n", - "{ -0.2013803f, 0.1559301f, 0.2713688f, -0.0187786f, 0.0075337f, -0.0453327f, 0.0510727f, -0.2533994f, -0.0093928f, 0.0939973f, 0.0683563f, 0.0257523f, -0.1638049f, 0.2167263f, 0.0139614f, -0.0689526f, -0.2007709f, 0.8788205f, 0.1043992f, 0.0529033f, 0.0041733f, -0.2248188f, 0.0029659f, 0.0044919f, 0.1728916f, -1.2823037f, 0.0284686f, -0.0879781f, 0.6332331f, 0.0599467f, -0.2467749f, 0.7796255f },\n", - "{ 0.1541705f, -2.8967228f, 0.2088300f, 0.0289306f, -0.1897649f, -0.1835614f, 0.1872510f, -0.5846522f, -0.0145777f, 2.2226386f, 0.0885817f, 0.0293056f, -0.0056043f, 2.9454181f, -0.0623621f, 0.0230481f, -0.8140234f, -4.3990140f, -0.2562745f, 0.6827632f, -4.7245188f, 0.1150251f, 0.0615204f, 4.7553473f, 0.1170709f, 0.0822542f, -0.6365855f, 0.3014538f, -2.6740234f, 0.1919117f, -0.0003937f, -0.0227543f },\n", + "{ 1.2430706f, 0.2385966f, -0.1307205f, 0.0325143f, -0.1869316f, 0.3236946f, -0.0250754f, 0.5280194f, -0.1030632f, -0.1919166f, 0.5273553f, 0.0652898f, -0.3363385f, 1.7545956f, -1.3353617f, -0.5440192f, -2.4663746f, -0.2166514f, 0.5136666f, -0.3368240f, -0.0757182f, 0.0920149f, -0.5879773f, -0.0225713f, 0.0697212f, 0.8563508f, -0.0126836f, -0.1828795f, -1.0318376f, -2.1025839f, 0.2279171f, 0.0045987f },\n", + "{ -0.0346952f, -0.0168656f, 0.0407209f, -0.1482209f, 0.0585610f, -0.0068832f, -0.1967536f, 0.0088515f, 0.0117921f, -0.1666735f, 0.0043757f, 0.0192000f, 0.0101610f, -0.0203226f, -0.0722811f, -0.0180960f, 0.0279165f, -0.2452999f, -0.0029923f, -0.0095144f, 0.0082529f, 0.0223957f, 0.0012511f, -0.0040226f, -0.0002326f, 0.0050690f, -0.0015876f, -0.2731955f, -0.0103268f, -0.0168681f, 0.0081678f, -0.0030315f },\n", + "{ -1.8121948f, -0.5878808f, -0.0582871f, 0.0676443f, -0.2566810f, 1.6678520f, 0.0751771f, -0.5183609f, 0.1661583f, -0.2097989f, -0.1213966f, 0.0194388f, 0.1857277f, -0.9746100f, -0.3444054f, -0.0763516f, 1.0879602f, 0.0184965f, 1.1060841f, 0.1923849f, -0.0412517f, -0.2354108f, 1.4612926f, -0.0095714f, 0.4891190f, -2.5078576f, 0.0020186f, 0.2102930f, 4.3355784f, 1.3152916f, -0.0975863f, 0.0044517f },\n", + "{ 1.5896463f, -2.0003557f, -0.1210541f, -0.1304628f, 0.0131182f, -0.6161298f, -0.1486594f, 1.2760408f, -0.1021480f, 0.0877840f, 0.6571876f, 0.0804771f, 0.0627130f, 1.8012530f, -4.5879707f, 1.4820974f, 0.1388381f, -0.2526270f, 0.8011180f, 0.9165006f, 0.5784082f, 0.1881581f, -0.6235034f, -0.0051164f, -1.4052374f, 0.0785857f, -0.0013880f, 0.1095322f, -0.7017108f, -0.5751966f, -0.3369588f, -0.0015683f },\n", + "{ 4.6118355f, -3.6102943f, 0.3775424f, 0.2372299f, 0.0598787f, 0.5582290f, 0.2304319f, 0.9093012f, -0.3773867f, 0.0047977f, 2.6425505f, -0.1439949f, 0.3935693f, 1.3001438f, 1.4970073f, -5.4273233f, -0.0687990f, 0.0670471f, 0.6703949f, 11.1249809f, 5.9740663f, 0.2207747f, -7.4394078f, -0.2285454f, 0.2585711f, 0.1874733f, -0.0016656f, 0.0217293f, 0.9555681f, 3.0554528f, -0.5727979f, -0.0026786f },\n", + "{ -1.7840524f, -0.4664120f, -0.1709528f, -0.1229654f, 0.2472902f, 16.6959667f, 0.1021524f, 23.0225563f, -8.8298550f, 0.1945520f, 1.4932841f, -0.1441254f, 17.2260818f, 2.2682095f, -0.2566192f, -0.3763468f, -2.6818073f, -0.1413464f, -27.9071026f, 3.8443294f, 1.4446557f, -0.0005846f, 0.8942884f, 0.2102544f, -3.4054801f, 2.1674650f, 0.1294054f, 0.1876321f, 1.8387170f, 2.3066981f, -4.6334472f, 0.0203269f },\n", + "{ 0.1221364f, -0.9206535f, -0.0578602f, 0.0086757f, -0.2020053f, 1.2470760f, 0.0963023f, 0.8218359f, -0.3108872f, -0.0679653f, 0.8242676f, -0.0567979f, -0.4664726f, 3.0569484f, 2.4042845f, -0.8774980f, 1.5614197f, -0.0359291f, 1.8720486f, 0.0217175f, -0.7438901f, -0.1705937f, 0.6922700f, -0.0708772f, 2.1580865f, -4.4101124f, 0.0070985f, 0.2220936f, 3.2375884f, -1.9589092f, -0.6771650f, -0.0064689f },\n", + "{ -1.5325402f, -1.3310759f, 0.1091106f, -0.0624776f, -0.1820637f, 1.2152895f, 0.0591179f, -2.6746740f, -0.1679525f, -0.2618925f, -0.0701432f, -0.0717427f, -0.3926733f, 1.0515925f, 1.2097669f, -0.8023561f, 0.8519148f, 0.2428202f, -2.4016669f, -0.6443134f, -0.8455796f, 0.0883630f, -0.5246723f, -0.2599009f, 2.5098257f, 2.5735190f, -0.0004903f, 0.0228214f, 0.3101828f, -2.6110868f, -0.1449447f, 0.0112252f },\n", + "{ -2.4570007f, 4.1306930f, -0.0709054f, -0.2456777f, -0.0434773f, 0.2500325f, -0.2469379f, 6.4086423f, 0.9810975f, 0.2620402f, 1.1104311f, -0.0850174f, 1.4035217f, -10.0576115f, -2.8340843f, -10.1643343f, -1.2882394f, -0.1628990f, 8.2278786f, 2.6009655f, 10.5084915f, 0.1247435f, 14.2758970f, -0.2018164f, -5.3260231f, -5.9585323f, 0.0054128f, 0.1511814f, -0.6517492f, 2.2414865f, -0.1060251f, -0.0213229f },\n", + "{ -3.5239015f, 0.4428157f, 0.0098384f, -0.1035724f, 0.2192848f, 10.4552908f, -0.1394745f, 3.0283575f, -15.7956038f, 0.2377471f, -10.4040318f, 0.0577844f, 9.0448751f, 0.9642384f, 2.6048126f, 0.5453808f, 1.8676072f, -0.2262070f, -0.2392199f, -1.4581424f, 2.0151219f, 0.0250153f, 0.2735855f, 0.1904391f, -2.4851513f, -0.4783628f, 17.5952053f, -0.2103962f, -2.4205782f, -2.6922755f, -2.0195608f, -17.3869209f },\n", + "{ 1.5699111f, 1.6051487f, -0.1158549f, -0.0307793f, 0.1400554f, 1.2221471f, 0.2116016f, 1.9285921f, 0.0042934f, 0.2085318f, 0.4481271f, 0.1907820f, 0.6049834f, -0.0403695f, -1.6741180f, 1.4565691f, 0.8936112f, -0.1921556f, 2.4673624f, 2.3142762f, 0.8624949f, 0.1622195f, -0.0171864f, 0.0495124f, 1.2759757f, -0.5775546f, -0.0131285f, 0.0019629f, 1.2494990f, -0.0695108f, 2.5645378f, -0.0251958f },\n", + "{ 0.8203233f, -0.7310246f, 0.0924949f, 0.0314029f, 0.0004415f, 1.4434688f, 0.0189398f, -3.1009448f, -0.2930109f, -0.2146460f, -0.7676558f, -0.2426467f, -1.5872091f, 0.4062420f, 0.6988549f, 1.1018623f, 1.6228473f, 0.0036644f, -1.7428403f, -0.6364639f, -0.4087792f, -0.0277658f, 0.5860791f, -0.2002079f, 1.7377874f, 1.4471169f, -0.0070140f, -0.0075029f, 0.7090668f, 0.5809853f, 1.8224814f, 0.0000271f },\n", + "{ -0.0271162f, 0.0158998f, 0.0030754f, -0.1410241f, -0.2236365f, 0.2567714f, -0.0673556f, 0.2602976f, -0.0035925f, -0.0680699f, 0.1860849f, -0.0224016f, 0.0799600f, -1.4051689f, 0.1068115f, 0.3181221f, 0.2607351f, -0.2092566f, 0.4402137f, 0.1669762f, 0.7319549f, -0.0637127f, 0.0924880f, -0.1975695f, 0.1949479f, -0.1818466f, 0.0019299f, -0.2326310f, -0.1074452f, 0.7910834f, -0.4338400f, 0.0037428f },\n", + "{ -0.8883290f, -0.0982433f, -0.1428742f, -0.1160625f, 0.2654903f, 6.9522190f, 0.1999476f, 4.6878695f, 4.9569860f, -0.1421416f, 6.2829461f, 0.1076571f, -2.4888971f, 0.2070822f, 1.0991781f, -0.1491540f, -0.2840673f, 0.1541903f, -3.4112997f, 0.9145759f, 1.2377257f, 0.1603911f, 0.3144833f, -0.2557842f, -1.3645093f, 0.6859801f, 5.6290560f, 0.0619957f, -0.5346277f, -0.8084556f, -4.3686385f, -5.5265436f },\n", "};\n", "\n", "ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_layer2[32] = {\n", - "0.5612384f, -0.0400684f, -0.1890610f, -0.8038151f, -0.0184527f, -0.8351601f, -1.7656847f, -0.4417709f, -0.3462953f, 1.0924704f, 0.1019267f, 0.0497286f, 0.3448936f, -0.0442495f, -0.1294845f, -0.1740453f, -0.6256254f, 1.0588725f, 0.1306455f, 0.0451779f, 1.2896419f, -0.0145429f, 0.2775581f, -0.6205941f, 0.0369313f, -0.0632537f, -0.1257888f, -0.2130138f, 0.0593540f, 0.8294140f, -0.5174863f, -0.0208223f };\n", + "-1.2246032f, -0.1255217f, -1.0217633f, -0.1814615f, -0.2436912f, -0.9311994f, -0.2956236f, -0.0172577f, -0.2968712f, 0.0741151f, -0.0555606f, 0.1839059f, -0.4699410f, -1.0013667f, 0.5347722f, 1.2293311f, 0.5103592f, 1.4865664f, -0.1879897f, -0.1940634f, -0.0728773f, -0.1775148f, -0.2175526f, 0.7983139f, 1.2735363f, -0.0693470f, 0.0497536f, -0.1538924f, 0.4001399f, -0.2371438f, -0.1275322f, 0.8652419f };\n", "\n", "ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_layer2[32][32] = {\n", - "{ 0.0802436f, -0.1068745f, -0.1320603f, 0.0222937f, 0.0260173f, -0.0999878f, 0.0045460f, 0.0257662f, -0.1026595f, 0.2246336f, -0.0252555f, 0.0313543f, -0.2062458f, 0.1684278f, -0.0972477f, -0.2176001f, -0.1525148f, -0.0692302f, 0.1213232f, -0.1288088f, -0.1729289f, -0.0895176f, -0.0199082f, -0.0172326f, 0.0126052f, 0.0598741f, 0.0853796f, -0.1326686f, 0.0753646f, -0.0108036f, -0.1034372f, -0.0026847f },\n", - "{ -0.5740064f, -0.1143318f, 0.1388070f, 1.3980483f, 0.0779263f, -1.2151324f, 2.2568495f, -0.5645000f, 2.2761426f, -0.2111119f, -0.3852313f, -1.2643801f, -0.8515005f, -0.1352748f, -0.1276971f, -0.9761337f, -1.1221037f, -0.1373484f, -6.1303482f, -0.6726473f, -3.5227783f, 0.3276748f, -0.8825266f, -0.0658490f, -0.1995521f, 0.4330961f, -0.0200665f, 0.1250316f, -0.1700243f, 1.1084150f, 1.6254629f, 0.3429218f },\n", - "{ 1.1285430f, -0.2066509f, -0.0835447f, 1.6354681f, -0.1452742f, -0.4063630f, 0.9329628f, -0.3926201f, 0.1437214f, 0.2107810f, 1.6312121f, 1.3499900f, 0.3157536f, -0.0890818f, 0.1018713f, 0.1947485f, 0.2054133f, -0.2302379f, -0.1797780f, -0.5907550f, 0.9738173f, 0.6502790f, 0.2307118f, -0.2878186f, 0.7077893f, -0.2159140f, -0.0506834f, -0.1203295f, 0.0040051f, 0.9362236f, -0.8645781f, -5.5910249f },\n", - "{ 0.1194026f, -0.0770950f, 0.0977710f, -0.0434290f, -0.0754589f, -0.0162118f, 0.1764810f, -0.0389150f, 0.1742229f, -0.0958543f, -0.1752356f, -0.0456456f, 0.1090995f, -0.0889283f, 0.0165932f, 0.0426875f, 0.0146571f, 0.0522842f, -0.0530897f, 0.0810161f, 0.0913054f, 0.0682782f, 0.0375430f, 0.0441132f, 0.0616512f, 0.1282314f, -0.0735286f, 0.0566166f, 0.1289222f, 0.0390186f, 0.0632017f, 0.0279914f },\n", - "{ -0.1397050f, 0.1469029f, -0.1035157f, -0.1537557f, 0.0723362f, 0.1360317f, 0.1632015f, -0.0519030f, 0.1664423f, -0.1674401f, 0.1221813f, 0.1218546f, 0.0581016f, -0.0302506f, 0.1745771f, 0.0057055f, -0.0728391f, -0.1049056f, -0.1592961f, 0.0329875f, 0.0789358f, -0.0462632f, 0.0748179f, 0.0180050f, 0.0803397f, -0.1458464f, -0.1164891f, -0.0082622f, 0.0343072f, 0.0366396f, -0.1715891f, -0.1089048f },\n", - "{ 0.5968121f, -0.1619606f, -0.1518290f, 0.7109355f, -0.2739125f, -1.2199029f, -0.0848204f, -0.9633313f, 1.5520710f, 0.9341283f, 0.2848818f, 1.7079298f, -0.0926562f, 0.0886664f, 0.0828461f, 2.1451118f, 1.2788274f, -0.1305077f, -0.1538649f, 0.6987577f, -1.6689410f, -0.0606114f, 0.8084964f, -0.4838576f, 0.0563342f, 0.7853296f, -0.0552194f, -0.1284064f, -0.1123770f, 1.0149344f, 0.1754804f, 0.2576492f },\n", - "{ -2.3144670f, -0.0794675f, -0.1847414f, -0.1791797f, -0.0605982f, -3.8482890f, 0.9826989f, -1.9203123f, -0.8249935f, -0.6718075f, -1.6555610f, 2.0798690f, 1.5665153f, -0.2050105f, -0.1993937f, 1.1070000f, 0.0836252f, 1.6307852f, 0.2448520f, -2.1885965f, -0.2890749f, 2.9986718f, 2.8526762f, -0.7994567f, -0.3042918f, 0.0960785f, 0.0377693f, -0.0463109f, -0.0870433f, -0.8293707f, -1.3597184f, 1.3583397f },\n", - "{ -0.3422491f, -0.0783274f, -0.1543446f, 0.0939283f, -0.3309467f, 0.6933140f, 0.6562154f, -0.6217518f, 0.7983661f, -0.1371530f, 0.5118276f, 0.7320337f, 0.5217202f, -0.0545936f, -0.1052059f, -0.1444394f, 0.3567903f, 1.2509772f, -0.6311634f, -0.7454629f, -1.0077031f, 0.8453025f, -0.2257671f, -0.2347097f, -0.3497045f, -0.1478627f, -0.0333489f, -0.2663794f, -0.0317280f, -0.3712497f, 0.8136677f, -2.5842018f },\n", - "{ 0.0997253f, 0.0620212f, 0.0460688f, 0.1300038f, 0.1323504f, -0.1669361f, 0.0732264f, -0.0860083f, -0.0762404f, -0.1628099f, -0.0881021f, 0.1323460f, -0.0460503f, 0.1401906f, 0.0602387f, 0.1474468f, 0.0935555f, -0.1531849f, 0.1754870f, 0.1422898f, 0.1418013f, -0.0302381f, -0.1058683f, -0.0258421f, -0.0110884f, 0.1016085f, 0.0643939f, -0.0729882f, -0.0572575f, -0.0603358f, 0.0267279f, 0.0611828f },\n", - "{ -0.6703482f, -0.0389322f, -0.1896956f, 1.0978996f, -0.1892595f, -2.0757272f, 1.7673731f, -0.7337908f, 2.1260054f, -0.6923556f, -0.6350248f, -1.8122731f, -1.2794832f, 0.0727381f, -0.1702943f, -1.0052446f, -1.1308233f, -0.1550428f, -5.3850775f, -0.5972114f, -0.3461530f, -0.4958812f, -1.2339170f, 0.0774786f, -0.8347382f, 0.2885106f, -0.1653138f, -0.0159324f, -0.1990542f, 1.6405339f, 1.7268503f, 0.8528640f },\n", - "{ 0.0038449f, -0.2388104f, -0.1149939f, -1.2830256f, -0.0052918f, 0.3492746f, 0.3040406f, 0.6817812f, 0.8697712f, 0.4297135f, -1.2266355f, -0.6838435f, 0.4784633f, -0.2080981f, 0.1026195f, 0.4500216f, -0.4817615f, -1.5641849f, 1.6826199f, -0.6830738f, 0.3725868f, 0.4554783f, -0.0576175f, 1.0322572f, 1.9159367f, -0.3336473f, 0.1027016f, -0.2515334f, -0.1622610f, -1.8702512f, -0.7941946f, -2.9838107f },\n", - "{ -0.2906766f, -0.0472568f, -0.2609593f, -1.5798510f, -0.1301123f, -1.1353090f, -2.5198724f, 1.5027611f, 1.2625716f, 3.2891662f, -2.3402910f, -0.0245398f, -9.8846655f, -0.2448200f, -0.0981539f, -0.2132508f, 0.7027491f, -3.4207478f, 1.3422097f, 4.3238688f, -1.1800685f, -2.7913725f, 1.8557802f, 5.5698090f, 0.2008359f, 0.2571939f, -0.0491005f, -0.1192926f, -0.0141392f, 0.1872108f, -0.8192848f, 0.6364858f },\n", - "{ 0.0671812f, -0.0234023f, 0.1400131f, 0.0778011f, 0.1308578f, -0.1675161f, 0.0237332f, 0.0215410f, 0.1514422f, 0.0736446f, 0.0181612f, -0.0219220f, -0.0099684f, 0.0102909f, 0.1243076f, 0.0897413f, -0.0682666f, 0.0389046f, 0.1245468f, 0.0098897f, -0.0425716f, 0.1595688f, 0.0397469f, -0.0664724f, -0.1641733f, 0.0605745f, 0.1712325f, 0.1596854f, 0.0224220f, -0.1328268f, -0.1743169f, 0.0608113f },\n", - "{ -1.1302176f, -0.0267920f, -0.2404846f, 0.6629997f, 0.1574058f, 4.7085104f, 0.5321816f, 0.4496275f, -0.2375129f, 1.2458097f, 1.8413728f, -1.8281459f, 2.3066695f, -0.2194766f, -0.0977243f, -0.4933192f, 0.5315795f, -0.8266373f, -1.6038543f, -0.6524141f, -0.2899630f, -4.3681431f, 0.2379541f, -0.3776518f, 1.5330435f, 0.3396196f, -0.0837295f, -0.0557159f, 0.1054505f, 1.1391810f, 0.6029354f, 0.9084895f },\n", - "{ -0.0898137f, 0.0973598f, -0.1583033f, -0.1445000f, 0.0044857f, -0.1162652f, -0.1396598f, 0.1172944f, 0.0714815f, -0.1043701f, 0.0497041f, 0.0381792f, 0.1395592f, -0.0310079f, -0.1161861f, -0.0776435f, 0.0771079f, -0.0311096f, -0.1498033f, 0.0838249f, 0.1343088f, 0.1152903f, 0.0749620f, -0.1248982f, -0.0346379f, -0.0780947f, -0.0730843f, 0.1654648f, -0.1482577f, -0.0118278f, 0.1078758f, -0.1372479f },\n", - "{ -0.7470447f, -0.0269854f, 0.0793572f, -0.7553226f, -0.0888210f, 1.8323143f, 0.2497575f, -10.9856701f, -0.7587091f, 1.1695130f, -0.0198075f, -0.8016124f, -0.2920889f, -0.1624553f, -0.0925370f, -1.3623806f, -1.3067284f, 0.2490501f, 0.0292220f, -0.2309522f, 0.0131977f, -0.7858045f, 0.1076498f, 0.6439033f, -0.1686209f, -0.8376603f, 0.0470587f, 0.0656591f, 0.0300643f, 1.4373403f, 0.2995886f, -1.3597747f },\n", - "{ -0.8746193f, -0.0201447f, -0.1603909f, -0.9373384f, 0.0254792f, 0.7993639f, 0.1087815f, -0.4859151f, -1.9559643f, -2.9621441f, -1.7735658f, -2.2227323f, -0.3725011f, 0.0284650f, -0.2266676f, -0.7529051f, -0.4477961f, 0.2071678f, 2.2770953f, 1.1170574f, -0.7023426f, 0.6896675f, -1.2416189f, -1.0012786f, -1.8752691f, -0.0359559f, -0.0125555f, -0.0457818f, -0.0775177f, 0.6092747f, 0.6639680f, 1.0951738f },\n", - "{ -1.7250717f, -0.0309622f, 0.1054536f, 2.0844676f, -0.2957073f, -0.0859962f, -1.5239947f, -0.2195731f, -1.5450290f, 0.4916542f, -1.3940116f, -0.5043938f, -0.7850562f, -0.1393226f, 0.0421535f, -0.2805571f, 0.5149463f, -0.4160199f, 1.3064783f, 0.3980424f, 0.1537250f, 0.2291165f, -1.5061877f, -0.0060016f, 0.1236818f, -0.1598873f, 0.0182953f, -0.0649871f, -0.1681922f, -0.8940877f, -0.6040494f, -0.5640311f },\n", - "{ 0.1721584f, -0.0760296f, 0.0164177f, 1.6552234f, -0.2948481f, -0.8678465f, 0.9449416f, -2.5188146f, -0.3830298f, -0.9600880f, 1.3968240f, 0.4318309f, 1.2421557f, -0.2275233f, -0.2093997f, -0.4352153f, -0.2142241f, -1.6719555f, -1.5515612f, -1.5296252f, 0.6067497f, 2.3952096f, -1.4322679f, -1.7078539f, -0.0283693f, -0.0217880f, -0.0268028f, -0.0223932f, -0.0441584f, 0.2260311f, -0.3117388f, 0.0257804f },\n", - "{ -0.3736052f, -0.1436419f, -0.1791924f, 0.6735403f, -0.1233307f, 0.9371016f, -0.5032559f, 0.0065924f, 0.2331814f, 0.5494700f, 1.4163370f, 0.7603047f, 1.8672758f, -0.1058595f, -0.0214494f, -0.1112118f, 0.4055682f, -0.2838995f, -1.8654461f, -0.6464235f, 0.3173830f, 0.1275925f, -0.0674378f, -0.1359007f, 0.8348094f, 0.2912215f, -0.1404384f, -0.2908648f, -0.1660577f, 0.4171410f, -0.5499387f, -0.9420084f },\n", - "{ -10.1053333f, -0.1126838f, 0.0914457f, 2.1485925f, -0.2125181f, -0.6991695f, -1.5648814f, 3.5901635f, 0.4265155f, 2.2461650f, 1.5659872f, -4.5079947f, -1.8503289f, 0.0036438f, -0.0552862f, 6.9531665f, 7.5919185f, -5.2455883f, 1.9158086f, -0.4571260f, -8.5806217f, 3.1702795f, 4.6932554f, 0.9312032f, -0.2189541f, -35.8472443f, -0.0437394f, 0.0193345f, -0.0759754f, -0.8393922f, -8.7812824f, 1.2664827f },\n", - "{ -0.4285544f, -0.0112712f, -0.1986143f, -0.6809276f, -0.2643681f, 0.8245937f, -0.1275320f, 1.8360085f, -1.0321223f, 1.0097107f, -0.4381949f, -0.0334159f, -0.1864567f, 0.0119804f, -0.1997918f, -0.3271705f, -0.0386158f, 1.0598557f, 0.9783336f, 1.9556490f, -0.2163550f, -2.8529193f, 0.6785525f, 1.3684491f, 1.2711524f, -0.1338721f, -0.0004762f, -0.0270106f, -0.1851630f, 0.1504625f, 0.1159483f, 1.3600112f },\n", - "{ 0.1121258f, 0.0640089f, -0.1267886f, 0.0409490f, -0.1376833f, -0.0064835f, 0.1204074f, -0.0622866f, 0.1238085f, 0.0468351f, -0.0761505f, 0.0465574f, 0.0356306f, -0.0866522f, 0.1416768f, 0.1586308f, 0.1173862f, -0.0805042f, 0.1809647f, -0.1865381f, -0.1134762f, -0.1078163f, 0.1637889f, 0.3176169f, 0.1323416f, 0.0248524f, -0.0535714f, 0.1136851f, 0.0925486f, 0.0475878f, -0.0636055f, -0.0584198f },\n", - "{ -14.6186295f, -0.1993816f, 0.1559632f, 2.3735945f, -0.0626924f, -8.2991180f, 0.2650338f, 3.5373747f, 1.6012716f, 2.8416808f, 1.1672604f, -4.3095527f, -1.1607343f, 0.0817207f, 0.1491408f, 7.4242001f, 7.2067833f, -4.3056359f, 0.8474119f, -0.9166479f, 0.6955321f, 2.0489492f, 4.0285969f, 0.0758955f, -1.0957255f, -38.1951637f, -0.0826333f, -0.1540980f, -0.2915384f, 1.0447723f, -9.4216700f, 0.7907819f },\n", - "{ 1.0854393f, -0.2057757f, -0.2333630f, -2.8073475f, -0.0092103f, -0.0027702f, 0.0400727f, 1.1957222f, 2.1408370f, 2.4174137f, -4.4619761f, -2.1267519f, -2.3692110f, -0.0215195f, 0.0940003f, 1.5145874f, 0.4090779f, 1.0525370f, 0.7181456f, 2.2475765f, -4.0287180f, 0.2974668f, 1.9458562f, 0.4356304f, -1.8651968f, 0.1091178f, -0.0625651f, 0.0101182f, -0.1866876f, 2.0775406f, -0.0924411f, -0.9942529f },\n", - "{ 0.6088188f, -0.1557965f, -0.1098123f, -2.0238423f, -0.0628158f, 1.8817046f, -2.8116503f, -1.9974719f, -2.8561919f, -1.5064508f, -0.5036830f, 0.0460033f, -1.2808908f, -0.2553190f, -0.2242925f, 0.3286708f, 0.6866547f, -0.8440441f, -3.4105279f, 1.5247352f, 0.5397177f, -0.5575525f, -0.4773433f, -0.8334278f, -1.8625259f, -0.5315101f, -0.1528733f, -0.1654517f, -0.1824549f, -0.9775680f, -2.9137912f, -0.4873172f },\n", - "{ -3.4014626f, -0.0843555f, 0.0550500f, 1.7411346f, -0.1923270f, 0.9657442f, -2.5730155f, -0.4918169f, 0.7225664f, -1.8889602f, 3.7809584f, 1.7930225f, -1.6396630f, -0.2156895f, -0.1968789f, -0.4836586f, -1.0472771f, 3.1005175f, -3.8917956f, -4.2640653f, 0.9576553f, -0.5442499f, -0.7966165f, -7.2177167f, 3.5426965f, 0.3307767f, -0.1710942f, -0.1729530f, -0.0668134f, -0.4916894f, 1.9444537f, 0.5960958f },\n", - "{ -0.9639474f, -0.0439117f, 0.0067572f, -0.8821743f, 0.1859997f, -0.7608641f, 1.5826563f, -1.4137437f, -0.3524639f, 0.4026311f, -0.0760814f, 0.7603636f, 2.5665376f, 0.0416676f, -0.1360063f, -0.6073055f, -1.3086349f, 1.9021115f, 1.3700111f, -0.9890064f, -0.1628284f, -1.3746781f, 0.2909067f, 0.3049279f, 0.8926148f, -0.4370743f, -0.0742112f, -0.0174381f, 0.0485841f, -0.0802396f, 3.5397923f, -0.6531604f },\n", - "{ 0.0491318f, -0.0591713f, 0.0498613f, 1.1349666f, -0.0889358f, -1.5782114f, 0.4471801f, 1.2610462f, -0.4626705f, 0.9874090f, 1.3419105f, 0.9187180f, 0.3291495f, -0.0520998f, -0.0861611f, 0.0016949f, -0.3544937f, -0.4579405f, -0.3237021f, -0.2511895f, -3.9122045f, -0.4933105f, 1.1840253f, 0.0769721f, 1.6942110f, -0.0282394f, -0.0797413f, -0.0965752f, -0.1654065f, 0.6972140f, 0.0176173f, 0.2189070f },\n", - "{ -1.1536870f, -0.1555014f, -0.3485503f, -1.5355002f, -0.2638741f, 1.3639680f, 1.6288967f, 0.6296268f, -0.5865426f, -0.5947766f, -1.4163440f, -1.5454646f, -0.7648816f, -0.1768368f, -0.1375362f, 0.3310686f, -0.2621009f, 0.6404898f, 1.0589499f, 0.2490305f, 0.3255049f, -0.1451550f, -0.3024975f, 0.8173084f, -2.1893589f, -0.0099213f, -0.1596652f, 0.1068924f, 0.0260035f, 1.4544648f, -0.7779374f, -0.7673930f },\n", - "{ 0.0435760f, -0.0246287f, -0.1238125f, -0.0320143f, 0.1576336f, 0.0702628f, -0.1744860f, -0.1600130f, -0.0870369f, 0.0819791f, 0.0445500f, 0.0019862f, 0.1529912f, 0.0886548f, 0.0478378f, -0.0487068f, 0.0401030f, 0.0869702f, 0.1359502f, 0.1713400f, -0.1529109f, 0.1224933f, 0.0072967f, -0.0784628f, 0.1494704f, -0.1378658f, -0.0747428f, -0.0850826f, -0.0940957f, 0.0511818f, 0.1520686f, -0.0680635f },\n", - "{ 1.3802143f, 0.0130436f, -0.0118297f, 0.9667702f, -0.4034582f, 0.3238442f, 0.4695910f, -0.3288718f, 1.1822759f, 1.5390285f, 0.2593471f, -0.0801804f, 0.4593041f, 0.0215906f, -0.2340092f, 0.4232760f, -0.0674659f, 0.7093915f, -0.0001982f, 0.2270681f, -0.8337699f, -0.2296698f, 0.9161285f, 0.3906515f, -0.1553026f, 0.2775815f, -0.0620411f, -0.1116518f, -0.0008762f, -0.3016884f, 1.4699873f, 0.6260686f },\n", + "{ 0.5089083f, 0.0575017f, 0.5365443f, -0.0749042f, 0.0996821f, -0.0870541f, 1.4083110f, 0.0042451f, 0.0506413f, 0.1160910f, -0.0628334f, 0.5274143f, 0.4513455f, -0.9582062f, 0.0673435f, 0.6348773f, -1.9705150f, -0.2561726f, -0.1166724f, 0.0368685f, -0.1610932f, 0.0683969f, 0.0069520f, -0.1350110f, 0.9181439f, 0.1249722f, 1.0909826f, -0.0840018f, 0.9957948f, -0.0971043f, -0.0137849f, -2.2041860f },\n", + "{ 0.6033787f, 0.0869169f, -0.0859117f, 0.0315756f, -0.1130356f, 0.7097771f, -0.1339651f, -0.0328042f, 1.0109031f, 0.2830297f, -0.0936190f, -0.6122728f, 0.3918095f, 1.2013817f, 0.4156620f, -0.6251741f, 1.1518294f, 1.6751471f, -0.2515613f, -0.0678125f, -0.1258993f, 0.0016276f, 0.0489553f, -0.3969306f, -1.0053991f, -0.1408774f, -0.5742593f, -0.3888345f, -1.0801376f, 0.1232338f, -0.2474749f, 1.3247567f },\n", + "{ -0.0069317f, -0.0813263f, 0.0308648f, 0.0865811f, -0.1700420f, 0.0967914f, -0.0156711f, -0.1217743f, -0.0792132f, 0.0874347f, -0.0772177f, -0.0141818f, -0.0828527f, 0.1910101f, 0.0311023f, 0.0512783f, -0.1374564f, -0.1192217f, 0.0168840f, -0.0671096f, -0.1756282f, 0.0961502f, 0.1194936f, 0.1121689f, 0.0702861f, 0.0227058f, -0.1065478f, 0.0892736f, 0.1295921f, -0.0824243f, 0.1052793f, 0.1367913f },\n", + "{ 0.0218436f, 0.0658335f, -0.1404412f, 0.1147386f, 0.1404460f, -0.1633583f, -0.1220772f, 0.1565208f, -0.0028298f, -0.0484391f, 0.1510092f, 0.0730970f, 0.0835523f, -0.1401114f, 0.1184470f, 0.1755763f, -0.0048258f, 0.0622622f, -0.0341085f, -0.0806764f, -0.1538521f, -0.0015015f, 0.1527237f, -0.1283317f, -0.1138941f, 0.0758406f, 0.1322748f, 0.0583293f, -0.1526688f, -0.0575901f, 0.0629805f, 0.1006109f },\n", + "{ -0.0993613f, 0.1225954f, 0.0214997f, -0.1216850f, -0.1446034f, -0.1262178f, 0.1148179f, -0.0104865f, 0.1139638f, 0.1162356f, 0.0760315f, 0.0347920f, 0.1263798f, -0.0845871f, -0.0076240f, 0.0147776f, 0.0706741f, -0.0541775f, 0.1547825f, -0.0731929f, -0.1654783f, 0.1004926f, 0.0731769f, -0.0515466f, 0.0012489f, 0.1251491f, 0.0608163f, -0.1210160f, -0.0371475f, 0.0280545f, 0.0563840f, 0.0180449f },\n", + "{ 0.8394753f, 0.0863981f, 0.6413645f, 0.0409160f, -0.0078189f, -3.7878563f, 0.2678362f, 0.0196527f, 1.7210206f, -1.6028817f, -0.2068004f, -2.3214684f, -1.0184679f, -1.5257946f, 0.2028781f, -2.7971203f, 0.5778140f, -0.2566824f, -0.0615155f, -0.1043012f, -0.1660616f, -0.1092668f, -0.0927776f, 1.8451736f, 0.4113030f, -0.1081991f, 0.9670511f, 0.0100726f, -2.8746226f, -0.1742988f, -0.1429935f, 1.5318521f },\n", + "{ -0.0845136f, 0.0384511f, -0.1385615f, -0.0871150f, 0.0897845f, 0.0878664f, -0.0010738f, -0.1549659f, 0.0356569f, -0.1536330f, -0.0262665f, -0.1161826f, 0.1544266f, 0.0787109f, -0.0407829f, -0.0264572f, -0.0792578f, -0.0405977f, 0.1106679f, 0.1500933f, -0.0898741f, -0.1199529f, -0.1163880f, -0.1057636f, 0.0889505f, 0.0825375f, 0.1614031f, 0.0067508f, 0.0352729f, 0.0078936f, 0.1533775f, 0.0849639f },\n", + "{ 2.7207191f, -0.0055598f, 2.5277357f, -1.4293425f, 0.1117166f, 0.8231006f, -2.0332973f, -0.0739002f, 1.8611740f, -8.6415014f, -0.0123378f, -0.3534637f, -1.7967031f, 2.9673553f, 0.2152573f, -4.9613218f, 0.3553480f, -0.1516881f, 0.0313580f, 0.0298279f, 0.1744945f, -0.0543368f, -0.0657357f, -0.3597513f, 1.8521340f, -0.0200091f, 1.6491756f, -0.1347934f, -8.2300911f, 0.0531232f, 0.0599637f, -0.1086621f },\n", + "{ -0.7867953f, -0.0110963f, -0.4613801f, 1.0450928f, 0.1429149f, 1.4774159f, -1.1104021f, 0.0747843f, -0.8985512f, -1.1148651f, 0.0115854f, 0.1383941f, 0.1892344f, -0.4564212f, -0.1128876f, -6.8694520f, -0.1744848f, 1.5719544f, -0.1184832f, 0.0280010f, -0.0129338f, -0.0419609f, -0.1565810f, -1.3889359f, -2.2191308f, 0.0424039f, -1.2463529f, -0.0588384f, -0.4484319f, -0.1165283f, 0.0796688f, -0.5284064f },\n", + "{ 0.0652653f, 0.0615156f, -0.0795163f, 0.0001713f, 0.1648186f, -0.0528994f, -0.0847025f, -0.0793956f, 0.1112951f, -0.1075590f, -0.0029891f, -0.1321859f, -0.0573127f, -0.1392062f, 0.0507489f, -0.0275592f, 0.1043706f, 0.1051407f, 0.1235337f, -0.0951271f, 0.0715892f, -0.0868704f, -0.0582845f, -0.0601897f, 0.1471449f, 0.0802297f, 0.1092225f, -0.1009769f, 0.1374661f, -0.1123024f, -0.1594280f, 0.1239780f },\n", + "{ 0.4587068f, -0.0936088f, 0.1091407f, -1.4789468f, -0.2378854f, -0.7586868f, 1.2026873f, -0.1347813f, 0.6108205f, -0.9265009f, -0.1364171f, 0.4698360f, 0.1535055f, -0.2266157f, 0.0333197f, 1.4881195f, -0.1966134f, -0.9363742f, -0.2550847f, 0.0512618f, 0.0478554f, -0.0761714f, 0.0852998f, 0.7968215f, 0.3493969f, -0.2375832f, 0.6220873f, -0.1845089f, 1.1993877f, -0.0197335f, -0.0138845f, 1.9340528f },\n", + "{ -0.1209233f, 0.1684445f, -0.1172153f, 0.0205280f, -0.1180427f, -0.0904171f, -0.1664954f, -0.0941068f, 0.0728728f, 0.1295698f, -0.1708585f, -0.1739318f, 0.0748263f, -0.1401377f, 0.1161512f, 0.1253984f, 0.1506475f, 0.0483222f, -0.0940781f, -0.1165934f, 0.1754140f, 0.1053056f, -0.1382262f, 0.1219736f, -0.1636132f, 0.1426921f, -0.0591449f, 0.0103367f, -0.1190467f, -0.0959214f, -0.0777284f, -0.1641069f },\n", + "{ -0.6834289f, -0.0725848f, -0.2700611f, 1.0437120f, 0.0824887f, -2.7070620f, 0.4225597f, -0.0604267f, -0.6852378f, 0.1944695f, 0.0911727f, 0.1900434f, -0.0984350f, 1.9884139f, 0.4437184f, 0.7536621f, 0.0569126f, 0.6176490f, -0.0592906f, -0.0017330f, 0.0474772f, 0.0168128f, 0.0533080f, -1.2100418f, -2.4400961f, -0.0830669f, -1.3237737f, 0.1108096f, -1.6645818f, -0.1670439f, -0.0161379f, -0.3699313f },\n", + "{ -0.2601279f, -0.0627888f, -2.3651617f, -1.1256676f, -0.0283316f, -6.7830796f, -3.6895976f, -0.1580759f, -1.4195836f, -6.0885744f, 0.0452110f, -5.3680663f, -2.2806647f, -5.1564946f, 0.0099778f, 2.9737246f, -3.1940920f, -0.6476036f, 0.1829322f, 0.0246909f, -0.1631330f, -0.0444936f, -0.2123401f, -4.5203066f, -0.0740298f, 0.0058150f, -0.7009476f, 0.1273159f, 2.3706875f, 0.0853582f, -0.2045061f, 2.0988390f },\n", + "{ 0.0851754f, -0.1730167f, -2.2383544f, -1.4635299f, -0.0050372f, -1.0768518f, 1.2955434f, -0.1811388f, 2.4874403f, -0.0573546f, -0.0442298f, -1.1430494f, -0.1609198f, 0.5267517f, 0.1293965f, -0.7079992f, 0.9464460f, 1.9218533f, -0.1341864f, -0.0830647f, -0.0308418f, -0.0314987f, -0.1568387f, -0.5759755f, -3.5660131f, -0.2200986f, 1.9044633f, -0.0455861f, 1.2135637f, -0.0406553f, -0.0549359f, -3.2340167f },\n", + "{ -0.3601795f, -0.0950057f, 1.2506303f, 0.0705646f, 0.0159455f, 0.1998335f, -0.9806644f, -0.0993264f, 0.2361836f, 0.2085727f, -0.1145219f, -0.4826447f, 0.0773994f, 0.0078874f, -0.0499348f, 0.0593346f, 2.1682246f, 0.8369263f, -0.2426349f, 0.0915828f, 0.0177455f, -0.1256301f, -0.1294592f, -0.7195153f, -2.4313867f, -0.2160123f, -1.9654233f, -0.3661108f, 0.0221972f, 0.0420728f, -0.1662701f, 2.2909124f },\n", + "{ 0.5229083f, -0.0859756f, -1.2948138f, 0.2981660f, -0.0616795f, 0.5993323f, 1.2229108f, -0.1168597f, -0.1483506f, 0.2205500f, -0.1778468f, 0.7693940f, -0.1189292f, 0.4196700f, 0.5927172f, 0.8341553f, 1.0524089f, -1.6661072f, -0.1706631f, -0.1182255f, -0.1298270f, 0.1067912f, -0.0550823f, -0.3989949f, 1.0862428f, 0.0692725f, 0.0220655f, -0.0182227f, -0.3141328f, 0.0763854f, -0.2399873f, 0.3735673f },\n", + "{ -0.0493408f, -0.0620185f, -0.1440084f, 0.1366522f, -0.1421596f, -0.1071353f, 0.1569152f, -0.1714708f, -0.0085891f, -0.1523663f, -0.0666709f, 0.0920407f, -0.0104404f, 0.1303387f, -0.1666427f, -0.0643643f, -0.1648330f, -0.0403286f, 0.0011041f, 0.1209591f, 0.0040688f, -0.0941681f, -0.0029900f, -0.0372638f, 0.0528831f, -0.0957583f, -0.0537955f, -0.0797893f, -0.1704947f, -0.0559995f, 0.0740464f, -0.0752997f },\n", + "{ 1.5719815f, -0.1644378f, 1.8601693f, 5.5668306f, -0.1303247f, 5.0759544f, -1.4812335f, 0.0062226f, 0.5767426f, 5.0568337f, -0.1407256f, 4.7255578f, -0.6883666f, 2.1262701f, 0.3128934f, 2.2308426f, -1.1327085f, 3.4682004f, -0.2262406f, -0.1728720f, 0.1500425f, -0.1453467f, 0.0638414f, 0.0793114f, -0.3400198f, -0.2237183f, 0.1568154f, -0.0257598f, -0.4672260f, -0.1525187f, -0.2054953f, 0.1591575f },\n", + "{ -0.4987546f, -0.0995077f, 2.7520263f, 0.4053383f, -0.1276614f, 0.0818262f, -0.9757426f, -0.0043751f, -1.8990395f, -1.7651865f, -0.0606477f, -0.6438876f, -0.1125105f, -0.6256728f, -0.0335146f, 0.7741688f, 0.3912064f, -0.9055670f, -0.0045660f, 0.0077818f, -0.1204589f, 0.0773467f, 0.0537127f, -1.7081609f, 0.7387068f, -0.0725944f, -0.1396644f, 0.1628178f, -0.7278796f, -0.1286270f, -0.1079767f, -0.2916186f },\n", + "{ 0.1261214f, -0.0106476f, -1.0569049f, 0.2244065f, -0.1248134f, 0.9224349f, 0.1777439f, -0.1238517f, 0.3914232f, 0.1859665f, -0.1939506f, -0.4722152f, 0.2143378f, 0.8231044f, -0.2002590f, -1.3563371f, -0.2053423f, 1.3496532f, -0.1801796f, -0.0217177f, 0.0810915f, -0.1319678f, -0.1299806f, 0.9309435f, 1.2306161f, -0.1924531f, 1.8588091f, -0.2374240f, -1.2482435f, 0.0170704f, 0.0808160f, -1.6042936f },\n", + "{ 0.0353613f, -0.0415041f, 0.0430906f, 0.0391508f, 0.1654718f, 0.1647980f, 0.1752385f, 0.0376194f, 0.0297110f, -0.1648499f, -0.1170499f, 0.0424385f, -0.1728038f, 0.0867688f, 0.1044088f, -0.0208259f, -0.0972596f, 0.0016974f, -0.0276596f, -0.1079382f, -0.1497553f, 0.0368193f, -0.1487449f, -0.0142207f, -0.1384017f, -0.1133966f, 0.1205297f, -0.1694416f, -0.0848187f, 0.0165042f, -0.0311819f, -0.0107586f },\n", + "{ 0.3395178f, -0.0544951f, 0.4140292f, 0.4284469f, -0.1532507f, -0.8734673f, 1.3085660f, -0.0968414f, -0.9431682f, 0.0212970f, -0.0573725f, 0.2512161f, 0.1806573f, 1.9569112f, -0.2206540f, -1.4288682f, 2.3636749f, -2.0449739f, -0.0088101f, -0.0307277f, -0.1746318f, 0.0359623f, -0.1769885f, -0.1383351f, -3.5865929f, 0.0237848f, 0.2055962f, -0.2543508f, 0.9893691f, -0.1158237f, -0.1074060f, -0.1367110f },\n", + "{ -0.0818370f, -0.0691959f, 0.1008360f, -0.0867900f, 0.0706240f, 0.0653615f, -0.1036123f, -0.0565025f, 0.0877569f, -0.1702891f, -0.0305304f, 0.0311808f, -0.0784205f, 0.1343800f, 0.0457310f, 0.0330426f, -0.1380319f, -0.1204205f, 0.0351189f, 0.1660071f, -0.1655417f, -0.1318305f, -0.1387838f, -0.1377107f, -0.1507244f, -0.0132244f, -0.0144924f, 0.0727539f, -0.0959819f, -0.0384600f, 0.0876117f, 0.0915885f },\n", + "{ -0.2913065f, -0.1268295f, 0.0402893f, 0.5167278f, -0.0104059f, 0.2077989f, 0.3679934f, -0.2081723f, 0.1353243f, 0.8197210f, -0.1013957f, 1.1287574f, 0.1784925f, 1.4747944f, 0.0095530f, -0.2737869f, 0.7512702f, -0.4536139f, -0.2225167f, -0.1508588f, -0.1516107f, -0.0296172f, 0.0093031f, -0.0913578f, -0.1219817f, -0.0470407f, 0.2288801f, -0.1066827f, 0.9788207f, -0.1060857f, -0.2672491f, -1.1543201f },\n", + "{ 0.8206128f, 0.0830891f, 0.6098021f, 0.8114421f, 0.0100041f, 0.1802981f, 1.4186163f, -0.0978715f, 1.4950410f, -0.0608325f, -0.0877750f, -2.8319721f, 0.4774732f, -0.0360274f, -0.4008928f, -1.3351992f, -0.6958030f, 0.4174623f, -0.1046506f, -0.0839466f, -0.0734599f, -0.0768588f, -0.1111481f, 1.5469090f, -2.7853961f, -0.1235103f, 2.0516131f, -0.1304165f, -0.6423004f, -0.0017117f, -0.0383703f, -1.2066915f },\n", + "{ 5.9012880f, -0.1430433f, -4.8122344f, 1.3565518f, 0.0676391f, -6.5592794f, -10.6126184f, -0.0076572f, 3.2418084f, 0.8378894f, -0.0120687f, -2.6607361f, -32.6262512f, 0.5281777f, -40.5600204f, 0.2489026f, -1.4531211f, -3.3126616f, -0.2288216f, 0.0087201f, 0.1374615f, 0.1106128f, 0.0246236f, -22.0268345f, 0.0629736f, 0.1088031f, 4.0246558f, 0.0067829f, -3.1944098f, -0.0843966f, -0.1402911f, 0.2893855f },\n", + "{ 0.0687621f, -0.0233423f, -0.0049267f, 0.0329577f, -0.0040116f, -0.0038231f, -0.1479600f, 0.0412774f, -0.1182764f, -0.0800665f, 0.0088746f, -0.0386593f, 0.1165797f, -0.0171098f, 0.0882285f, -0.0606064f, -0.0594507f, -0.0446539f, -0.1713613f, -0.0783319f, -0.1286740f, 0.0856650f, -0.1753173f, -0.0516984f, -0.1444318f, 0.1162534f, 0.0405000f, -0.0671118f, -0.0576360f, -0.1409870f, -0.1164895f, -0.0137712f },\n", + "{ -1.3168329f, -0.0033607f, 1.7238024f, -2.6184859f, -0.1436181f, -0.4960048f, 0.3308013f, 0.0842941f, -0.8149745f, -0.5922119f, 0.1075923f, -0.8475354f, -4.2522564f, -0.6004850f, -1.7454799f, 0.7431618f, -1.3471994f, 1.8992596f, 0.0583799f, -0.0095420f, -0.1535583f, -0.0602523f, -0.1145862f, -0.9051632f, 0.4375299f, -0.0222844f, -0.7205035f, -0.3361300f, 1.7331023f, -0.1060063f, -0.1233467f, -0.9474893f },\n", + "{ 0.0834905f, -0.0987103f, -1.4479494f, 1.1476132f, -0.1930190f, -0.8689866f, -0.8929498f, -0.1202456f, -0.4120998f, -0.3361138f, -0.1739257f, 0.7572783f, 0.2644251f, -1.1061769f, -0.0444610f, 0.5947868f, -1.8700641f, -0.8042451f, 0.0032584f, -0.0918714f, -0.0878237f, -0.0561467f, -0.0953190f, -1.0795568f, -1.2196006f, -0.1444945f, -1.2117988f, -0.1134432f, 0.3373002f, -0.0059311f, 0.0073372f, -0.0959612f },\n", + "{ 0.4720845f, -0.1032461f, -0.2123445f, -1.0046439f, -0.1504408f, 0.7121164f, 0.3517790f, -0.1829089f, 0.8960306f, -0.2978761f, -0.1084950f, -0.3584599f, -0.1990081f, 0.5987743f, -0.5727913f, 0.0202107f, -0.9536695f, 0.3441935f, 0.0361146f, -0.2192044f, -0.0706470f, 0.0332405f, -0.0892546f, 1.4020443f, 0.8471918f, -0.1363626f, 0.3598230f, 0.0537083f, -0.2526802f, 0.0390054f, -0.0971618f, -0.1145800f },\n", + "{ 6.6881781f, 0.0808753f, -4.0573001f, 5.1219244f, -0.1900990f, -0.9485528f, -8.2378635f, -0.0019337f, 8.3292599f, -5.9191675f, 0.0768199f, -1.6080343f, -38.4854393f, 2.5592017f, -35.8240967f, -2.1315293f, 1.3917850f, 3.9621065f, -0.0458908f, 0.1264065f, 0.0298109f, 0.0203746f, 0.0184819f, -14.0934982f, 1.2857219f, 0.0492928f, 4.1220269f, -0.3709270f, 2.5108135f, -0.0593165f, 0.0093613f, -1.2152934f },\n", "};\n", "\n", "ALPAKA_STATIC_ACC_MEM_GLOBAL const float bias_output_layer[1] = {\n", - "0.7275639f };\n", + "1.4821327f };\n", "\n", "ALPAKA_STATIC_ACC_MEM_GLOBAL const float wgtT_output_layer[32][1] = {\n", - "{ 1.4243358f },\n", - "{ 0.0335807f },\n", - "{ 0.0551641f },\n", - "{ -0.9836086f },\n", - "{ -0.0249541f },\n", - "{ -1.5375688f },\n", - "{ -0.7714168f },\n", - "{ -0.9649364f },\n", - "{ -1.1769278f },\n", - "{ 1.3249911f },\n", - "{ -1.6541473f },\n", - "{ 1.4079021f },\n", - "{ -0.8831168f },\n", - "{ 0.0122874f },\n", - "{ 0.0511134f },\n", - "{ -2.6734750f },\n", - "{ 2.8394303f },\n", - "{ 0.9675560f },\n", - "{ -1.4186903f },\n", - "{ -2.0796514f },\n", - "{ -1.7693948f },\n", - "{ -0.8502544f },\n", - "{ -1.5927037f },\n", - "{ -1.1028550f },\n", - "{ 0.8137528f },\n", - "{ 6.3073616f },\n", - "{ 0.1059108f },\n", - "{ -0.0468376f },\n", - "{ 0.1322162f },\n", - "{ 0.7481517f },\n", - "{ -1.2260461f },\n", - "{ -0.9095332f },\n", + "{ 2.0100033f },\n", + "{ 0.1341594f },\n", + "{ -1.1707165f },\n", + "{ -1.1725436f },\n", + "{ -0.1255062f },\n", + "{ -1.1738188f },\n", + "{ 1.0822003f },\n", + "{ 0.1121197f },\n", + "{ -1.0663537f },\n", + "{ -0.7722268f },\n", + "{ 0.0201395f },\n", + "{ 1.4703517f },\n", + "{ 2.3982928f },\n", + "{ -0.5911243f },\n", + "{ 3.9915426f },\n", + "{ -0.9631299f },\n", + "{ -0.6016800f },\n", + "{ 1.3677936f },\n", + "{ 0.0113799f },\n", + "{ 0.0772981f },\n", + "{ 0.0646394f },\n", + "{ 0.0517149f },\n", + "{ 0.0408024f },\n", + "{ 1.2203258f },\n", + "{ -0.6350328f },\n", + "{ -0.0750008f },\n", + "{ -1.5730240f },\n", + "{ -0.0382495f },\n", + "{ -1.2106268f },\n", + "{ 0.0631248f },\n", + "{ 0.0212301f },\n", + "{ -0.8234047f },\n", "};\n", "\n" ] @@ -749,7 +832,7 @@ }, { "cell_type": "code", - "execution_count": 10, + "execution_count": 11, "metadata": {}, "outputs": [], "source": [ @@ -767,12 +850,44 @@ }, { "cell_type": "code", - "execution_count": 11, + "execution_count": 12, + "metadata": {}, + "outputs": [], + "source": [ + "t3_pt = np.concatenate(branches['t3_radius']) * 2 * (2.99792458e-3 * 3.8) / 2" + ] + }, + { + "cell_type": "code", + "execution_count": 13, "metadata": {}, "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "pt: 0 to 5\n", + "93% Retention Cut: {0.3181, 0.3705, 0.4841, 0.6139, 0.6397, 0.7068, 0.8456, 0.8847, 0.9407, 0.9588} Mean: 0.6763\n", + "98% Retention Cut: {0.076, 0.0876, 0.1308, 0.1834, 0.2712, 0.3317, 0.5011, 0.5748, 0.8011, 0.8586} Mean: 0.3816\n", + "99% Retention Cut: {0.0387, 0.0395, 0.0599, 0.083, 0.1433, 0.163, 0.2927, 0.3243, 0.638, 0.7512} Mean: 0.2533\n", + "99.5% Retention Cut: {0.0185, 0.0183, 0.0277, 0.0385, 0.0745, 0.0757, 0.1232, 0.1528, 0.4176, 0.5605} Mean: 0.1507\n", + "99.9% Retention Cut: {0.0029, 0.0029, 0.0046, 0.0079, 0.0103, 0.0108, 0.0126, 0.0282, 0.0942, 0.1336} Mean: 0.0308\n" + ] + }, + { + "data": { + "image/png": "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", "text/plain": [ "
" ] @@ -785,10 +900,12 @@ "output_type": "stream", "text": [ "\n", - "pt: 0 to inf\n", - "93% Retention Cut: {0.3115, 0.3408, 0.4742, 0.569, 0.5812, 0.6033, 0.7483, 0.7834, 0.8547, 0.8994} Mean: 0.6166\n", - "98% Retention Cut: {0.0642, 0.0673, 0.1201, 0.1563, 0.2093, 0.2102, 0.3688, 0.4391, 0.6223, 0.7282} Mean: 0.2986\n", - "99% Retention Cut: {0.024, 0.0267, 0.052, 0.0658, 0.093, 0.0968, 0.1913, 0.2443, 0.4012, 0.5449} Mean: 0.174\n" + "pt: 5 to inf\n", + "93% Retention Cut: {0.0282, 0.0323, 0.0696, 0.2575, 0.315, 0.3752, 0.6781, 0.593, 0.6475, 0.4725} Mean: 0.3469\n", + "98% Retention Cut: {0.0045, 0.0058, 0.008, 0.0194, 0.048, 0.0823, 0.4164, 0.2597, 0.2001, 0.0529} Mean: 0.1097\n", + "99% Retention Cut: {0.0019, 0.0024, 0.0047, 0.0107, 0.0107, 0.0316, 0.2533, 0.1114, 0.0501, 0.0407} Mean: 0.0518\n", + "99.5% Retention Cut: {0.0007, 0.0016, 0.003, 0.0047, 0.0043, 0.008, 0.1114, 0.0296, 0.024, 0.0144} Mean: 0.0202\n", + "99.9% Retention Cut: {0.0002, 0.0003, 0.001, 0.0009, 0.0001, 0.0003, 0.0084, 0.0085, 0.0044, 0.0066} Mean: 0.0031\n" ] } ], @@ -821,8 +938,10 @@ " Dictionary containing branch data\n", " \"\"\"\n", " # Filter data based on pt bin\n", - " abs_eta = eta_list[0][full_tracks]\n", - " predictions_filtered = predictions[full_tracks]\n", + " abs_eta = eta_list[0][full_tracks & (t3_pt > pt_min) & \n", + " (t3_pt <= pt_max)]\n", + " predictions_filtered = predictions[full_tracks & (t3_pt > pt_min) & \n", + " (t3_pt <= pt_max)]\n", " \n", " # Dictionary to store cut values for different percentiles\n", " cut_values = {p: [] for p in percentiles}\n", @@ -886,52 +1005,12 @@ " eta_list, predictions, full_tracks, branches)\n", "\n", "# Example call:\n", - "percentiles = [93, 98, 99]\n", - "pt_bins = [0, np.inf]\n", + "percentiles = [93, 98, 99, 99.5, 99.9]\n", + "pt_bins = [0, 5, np.inf]\n", "eta_bin_edges = np.arange(0, 2.75, 0.25)\n", "analyze_pt_bins(pt_bins, percentiles, eta_bin_edges, eta_list, predictions, full_tracks, branches)" ] }, - { - "cell_type": "code", - "execution_count": 12, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "(2707725,)" - ] - }, - "execution_count": 12, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "np.shape(predictions[predictions > 0.215])" - ] - }, - { - "cell_type": "code", - "execution_count": 13, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "(15131951,)" - ] - }, - "execution_count": 13, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "np.shape(predictions)" - ] - }, { "cell_type": "code", "execution_count": null,