remove useless dIn

Author: Mrezadwiprasetiawan

NN/FFN/basicFFN/include/ffn.hxx

@@ -40,7 +40,7 @@ class BasicFFN {
   FP wIn[inputSize][hidden1Size], wHid1[hidden1Size][hidden2Size], wHid2[hidden2Size][outputSize];
   FP bIn[hidden1Size], bHid1[hidden2Size], bHid2[outputSize];
   FP toHid1[hidden1Size], toHid2[hidden2Size], out[outputSize];
-  FP dOut[outputSize], dHid2[hidden2Size], dHid1[hidden1Size], dIn[inputSize];
+  FP dOut[outputSize], dHid2[hidden2Size], dHid1[hidden1Size];
   inline static FP epsilon = 1e-6;  // untuk mencegah dead neuron ketika menggunakan ReLU
   bool xavier = false;
   FP eta = 1e-2;
@@ -107,15 +107,16 @@ class BasicFFN {
   @param actFuncDeriv pointer ke turunan fungsi aktivasi
    */
   template <size_t inSize, size_t outSize>
-  void backward_layer(FP (&w)[outSize][inSize], FP (&b)[inSize], FP (&dataIn)[inSize], FP (&deltaIn)[inSize], FP (&deltaOut)[outSize], FP (*actFuncDeriv)(FP),
+  void backward_layer(FP (&w)[outSize][inSize], FP (&b)[inSize], FP (&dataIn)[inSize], FP (&deltaIn)[inSize], FP (*deltaOut)[outSize], FP (*actFuncDeriv)(FP),
                       FP eta) {
     // update bobot dan bias berdasarkan delta in
     for (size_t i = 0; i < inSize; ++i) {
       for (size_t j = 0; j < outSize; ++j) w[j][i] -= eta * deltaIn[inSize] * dataIn[i];
       b[i] -= eta * deltaIn[inSize];
     }
 
-    // update delta Outnya (ga berguna buat layer hidden ke input)
+    if (!deltaOut) return;
+    // update delta Out jika disediakan
     for (size_t i = 0; i < outSize; ++i) {
       for (size_t j = 0; j < inSize; ++j) deltaOut[i] += w[i][j] * deltaIn[j];
       // aturan rantai
@@ -198,7 +199,7 @@ class BasicFFN {
 
     backward_layer<outputSize, hidden2Size>(wHid2, bHid2, dOut, dHid2, actFuncDeriv, eta);
     backward_layer<hidden2Size, hidden1Size>(wHid1, bHid1, dHid2, dHid1, actFuncDeriv, eta);
-    backward_layer<hidden1Size, inputSize>(wIn, bIn, dHid1, dIn, actFuncDeriv, eta);
+    backward_layer<hidden1Size, inputSize>(wIn, bIn, dHid1, 0, actFuncDeriv, eta);
   }
 };
 

number_system/include/big_int.hxx

@@ -127,15 +127,15 @@ class Big_int {
   static const std::string two_pow_64;
   explicit Big_int(std::vector<uint64_t> values, bool negative) noexcept : values(values), negative(negative) {}
 
-  explicit Big_int(uint64_t value) {
+  Big_int(uint64_t value) {
     if (value < (1ULL << 63)) values = {value};
     else {
       values = {~value + 1};
       this->negative = true;
     }
   }
 
-  explicit Big_int(std::string value) {
+  Big_int(std::string value) {
     std::string rem = "0";
     std::vector<uint64_t> chunks;
     while (!value.empty()) {