daviesl

import torch

# Set the seed for reproducibility
torch.manual_seed(42)

# True target distribution probabilities
true_probs = torch.tensor([0.2, 0.5, 0.3])

def f(x):
    # Minus sign is important to ensure we are minimizing