100 is not a hardcoded dimension any more

HDC_library.py

@@ -106,8 +106,8 @@ def train_HDC_RFF(n_class, N_train, Y_train_init, HDC_cont_train, gamma, D_b):
         alpha = np.linalg.solve(Beta,L) #alpha here is the whole v vector from the slides
 
         # Get HDC prototype for class cla, still in floating point
-        final_HDC_centroid = np.zeros(100)
-        final_HDC_centroid_q = np.zeros(100)
+        final_HDC_centroid = np.zeros(np.shape(HDC_cont_train[0]))
+        final_HDC_centroid_q = np.zeros(np.shape(HDC_cont_train[0]))
 
         for i in range(N_train):
             final_HDC_centroid = final_HDC_centroid + Y_train[i]*alpha[i]*HDC_cont_train[i] #this is mu(vector) from the slides
@@ -160,7 +160,7 @@ def evaluate_F_of_x(Nbr_of_trials, HDC_cont_all, LABELS, beta_, bias_, gamma, al
         HDC_cont_train_cpy = HDC_cont_train_ * 1
         # Apply cyclic accumulation with biases and accumulation speed beta_
         cyclic_accumulation_train = HDC_cont_train_cpy % (2 ** B_cnt)
-        HDC_cont_train_cyclic = np.zeros((cyclic_accumulation_train.shape[0],100))
+        HDC_cont_train_cyclic = np.zeros((cyclic_accumulation_train.shape[0],HDC_cont_all.shape[1]))
 
         for row in range(cyclic_accumulation_train.shape[0]):
             cyclic_accumulation_train_vector = np.array(cyclic_accumulation_train[row])
@@ -194,7 +194,7 @@ def evaluate_F_of_x(Nbr_of_trials, HDC_cont_all, LABELS, beta_, bias_, gamma, al
         # Apply cyclic accumulation with biases and accumulation speed beta_
 
         cyclic_accumulation_test = HDC_cont_test_cpy % (2 ** B_cnt)
-        HDC_cont_test_cyclic = np.zeros((cyclic_accumulation_test.shape[0],100))
+        HDC_cont_test_cyclic = np.zeros((cyclic_accumulation_test.shape[0],HDC_cont_all.shape[1]))
         for row in range(cyclic_accumulation_test.shape[0]):
             cyclic_accumulation_test_vector = np.array(cyclic_accumulation_test[row])
 

playground_testing.py

@@ -90,15 +90,15 @@ def test_XOR(in1,in2,desired,dim):
 
 dataset_path = 'WISCONSIN/data.csv' 
 ##################################   
-imgsize_vector = 30 #Each input vector has 30 features
+imgsize_vector = 100 #Each input vector has 30 features
 n_class = 2
-D_b = 4 #We target 4-bit HDC prototypes
+D_b = 8 #We target 4-bit HDC prototypes
 B_cnt = 8
 maxval = 256 #The input features will be mapped from 0 to 255 (8-bit)
-D_HDC = 100 #HDC hypervector dimension
+D_HDC = 300 #HDC hypervector dimension
 portion = 0.6 #We choose 60%-40% split between train and test sets
-Nbr_of_trials = 1 #Test accuracy averaged over Nbr_of_trials runs
-N_tradeof_points = 40 #Number of tradeoff points - use 100 
+Nbr_of_trials = 8 #Test accuracy averaged over Nbr_of_trials runs
+N_tradeof_points = 100 #Number of tradeoff points - use 100 
 N_fine = int(N_tradeof_points*0.4) #Number of tradeoff points in the "fine-grain" region - use 30
 #Initialize the sparsity-accuracy hyperparameter search
 lambda_fine = np.linspace(-0.2, 0.2, N_tradeof_points-N_fine)