losses.py 文件源码

def categorical_crossentropy_3d(y_true, y_predicted):
    """
    Computes categorical cross-entropy loss for a softmax distribution in a hot-encoded 3D array
    with shape (num_samples, num_classes, dim1, dim2, dim3)

    Parameters
    ----------
    y_true : keras.placeholder [batches, dim0,dim1,dim2]
        Placeholder for data holding the ground-truth labels encoded in a one-hot representation
    y_predicted : keras.placeholder [batches,channels,dim0,dim1,dim2]
        Placeholder for data holding the softmax distribution over classes

    Returns
    -------
    scalar
        Categorical cross-entropy loss value
    """
    y_true_flatten = K.flatten(y_true)
    y_pred_flatten = K.flatten(y_predicted)
    y_pred_flatten_log = -K.log(y_pred_flatten + K.epsilon())
    num_total_elements = K.sum(y_true_flatten)
    # cross_entropy = K.dot(y_true_flatten, K.transpose(y_pred_flatten_log))
    cross_entropy = tf.reduce_sum(tf.multiply(y_true_flatten, y_pred_flatten_log))
    mean_cross_entropy = cross_entropy / (num_total_elements + K.epsilon())
    return mean_cross_entropy