Блог

 def svm_loss_naive(W, X, y, reg):
  """
  Structured SVM loss function, naive implementation (with loops).

  Inputs have dimension D, there are C classes, and we operate on minibatches
  of N examples. When you implment the regularization over W, please DO NOT
  multiply the regularization term by 1/2 (no coefficient).

  Inputs:
  - W: A PyTorch tensor of shape (D, C) containing weights.
  - X: A PyTorch tensor of shape (N, D) containing a minibatch of data.
  - y: A PyTorch tensor of shape (N,) containing training labels; y[i] = c means
    that X[i] has label c, where 0 <= c < C.
  - reg: (float) regularization strength

  Returns a tuple of:
  - loss as torch scalar
  - gradient of loss with respect to weights W; a tensor of same shape as W
  """
  dW = torch.zeros_like(W) # initialize the gradient as zero

  # compute the loss and the gradient
  num_classes = W.shape[1]
  num_train = X.shape[0]
  loss = 0.0
  # print(W.t().shape)
  # print(X[0].shape)
  # print(W.t(х).mv(X[0]).shape)
  for i in range(num_train):
    scores = W.t().mv(X[i])
    correct_class_score = scores[y[i]]
    for j in range(num_classes):
      if j == y[i]:
        continue
      margin = scores[j] - correct_class_score   1 # note delta = 1
      if margin > 0:
        loss  = margin
        #######################################################################
        # TODO:                                                               #
        # Compute the gradient of the loss function and store it dW. (part 1) #
        # Rather than first computing the loss and then computing the         #
        # derivative, it is simple to compute the derivative at the same time #
        # that the loss is being computed.                                    #
        #######################################################################
        # Replace "pass" statement with your code
        dW[:, j]  = X[i]  
        dW[:, y[i]]-= X[i]
    #######################################################################
        #                       END OF YOUR CODE                              #
        #######################################################################
 

Задача состоит в том, чтобы вычислить градиент функции потерь. Я не мог решить это сам, поэтому нашел решение на github. Многие подходы были точно такими же, как и описанный выше. Даже сейчас я все еще не понимаю этого. Небольшая помощь, пожалуйста, будет оценена по достоинству.