one part

Author: JPLAY0

assignment1/cs231n/classifiers/linear_svm.py

@@ -37,9 +37,15 @@ def svm_loss_naive(W, X, y, reg):
             margin = scores[j] - correct_class_score + 1  # note delta = 1
             if margin > 0:
                 loss += margin
+                dW[:, j] += X[i]
+                dW[:, y[i]] -= X[i]
 
-    # Right now the loss is a sum over all training examples, but we want it
+                # Right now the loss is a sum over all training examples, but we want it
     # to be an average instead so we divide by num_train.
+    # X.T (D,N) loss (N,C)
+    # dW = (X.T @ loss) / num_train + reg * W * W
+    dW /= num_train
+    dW += reg * W
     loss /= num_train
 
     # Add regularization to the loss.
@@ -55,8 +61,6 @@ def svm_loss_naive(W, X, y, reg):
     #############################################################################
     # *****START OF YOUR CODE (DO NOT DELETE/MODIFY THIS LINE)*****
 
-    pass
-
     # *****END OF YOUR CODE (DO NOT DELETE/MODIFY THIS LINE)*****
 
     return loss, dW