Changed svm_c_linear_dcd_trainer so that when the user forces the last weight

to 1 it automatically disables use of the bias since that is almost certainly
what the user will want to do.

dlib/svm/svm_c_linear_dcd_trainer.h

@@ -284,7 +284,7 @@ namespace dlib
                             );
 
                         // extend w by the right number of elements
-                        if (have_bias)
+                        if (have_bias && !last_weight_1)
                         {
                             // Splice some zeros into the w vector so it will have the
                             // right length.  Here we are being careful to move the bias
@@ -316,7 +316,7 @@ namespace dlib
                     index.reserve(x.size());
                     Q.reserve(x.size());
 
-                    if (have_bias)
+                    if (have_bias && !last_weight_1)
                         w.set_size(dims+1);
                     else
                         w.set_size(dims);
@@ -328,7 +328,7 @@ namespace dlib
                 {
                     Q.push_back(length_squared(x(i)));
 
-                    if (have_bias)
+                    if (have_bias && !last_weight_1)
                     {
                         index.push_back(i);
                         Q.back() += 1;
@@ -570,7 +570,7 @@ namespace dlib
                         alpha[i] = std::min(std::max(alpha[i] - G/state.Q[i], (scalar_type)0.0), C);
                         const scalar_type delta = (alpha[i]-alpha_old)*y(i);
                         add_to(w, x(i), delta);
-                        if (have_bias)
+                        if (have_bias && !last_weight_1)
                             w(w.size()-1) -= delta;
 
                         if (last_weight_1)
@@ -618,7 +618,7 @@ namespace dlib
 
             // put the solution into a decision function and then return it
             decision_function<kernel_type> df;
-            if (have_bias)
+            if (have_bias && !last_weight_1)
                 df.b = w(w.size()-1);
             else
                 df.b = 0;
@@ -639,7 +639,7 @@ namespace dlib
             const sample_type& sample
         ) const
         {
-            if (have_bias)
+            if (have_bias && !last_weight_1)
             {
                 const long w_size_m1 = w.size()-1;
                 return dlib::dot(colm(w,0,w_size_m1), sample) - w(w_size_m1);

dlib/svm/svm_c_linear_dcd_trainer_abstract.h

@@ -111,6 +111,9 @@ namespace dlib
                 - returns true if this trainer has the constraint that the last weight in
                   the learned parameter vector must be 1.  This is the weight corresponding
                   to the feature in the training vectors with the highest dimension.  
+                - Forcing the last weight to 1 also disables the bias and therefore the b
+                  field of the learned decision_function will be 0 when forces_last_weight_to_1() == true.
+                  This is true regardless of the setting of #include_bias().
         !*/
 
         void force_last_weight_to_1 (