pytorch MultivariateNormal backward error with respect to covariance

Question

Here is a simple piece of code to learn covariance of 2D gaussian.

import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import