Started Loss, done softmax, up to p.125

I've implemented alot of support functions that needs to be refactored, optimised and tested; mean.h, exponential.h, matdiv.h matsum.h matsubtract.h. Maybe we need to have a look at if matdiv/matmul should be in the same. Same with matadd/matsubtract and if some of it should be in matvec.h.

include/modules/neural_networks/activation_functions/Softmax.h

@@ -4,7 +4,11 @@
 
 #include "./utils/vector.h"
 #include "./utils/matrix.h"
-#include "./utils/random.h"
+
+#include "./numerics/max.h"
+#include "./numerics/matsubtract.h"
+#include "./numerics/exponential.h"
+#include "./numerics/matdiv.h"
 
 
 namespace neural_networks{
@@ -12,15 +16,20 @@ namespace neural_networks{
 	template <typename T>
 	struct activation_softmax{
 
+			utils::Matrix<T> exp_values;
+			utils::Matrix<T> probabilities;
 			utils::Matrix<T> outputs;
 
-	        void forward(utils::Matrix<T> inputs){
-			    //outputs = numerics::max(inputs, T{0});
-			    //outputs.print();
+	        void forward(const utils::Matrix<T> inputs){
+	
+
+			    exp_values = numerics::exponential(numerics::matsubtract(inputs, numerics::max(inputs, "rows"), "col"));
+			    probabilities = numerics::matdiv(exp_values, numerics::matsum(exp_values, "col"), "col");
+
+			    outputs = probabilities;
+
 	        }
 
-
-
 	};