amachine.am_random 

 1from scipy.special import softmax
 2from scipy.stats import beta
 3import numpy as np
 4
 5def exp_uniform_blend(n: int, alpha: float) -> list[float]:
 6    exp_probs = np.exp(-np.arange(n))
 7    exp_probs /= exp_probs.sum()
 8    uniform = np.ones(n) / n
 9    return ((1 - alpha) * exp_probs + alpha * uniform).tolist()
10
11def beta_dist( n, a, b):
12	points = np.linspace(0, 1, n + 2 )[ 1:-1] 
13	probs  = beta.pdf(points, a, b)
14	return probs / probs.sum()
15
16def dirichlet_dist(n, alpha):
17	return np.random.dirichlet(np.ones(n) * alpha)
18
19def uniform_dist( n ):
20	return np.ones(n) / n
def exp_uniform_blend(n: int, alpha: float) -> list[float]: 
 6def exp_uniform_blend(n: int, alpha: float) -> list[float]:
 7    exp_probs = np.exp(-np.arange(n))
 8    exp_probs /= exp_probs.sum()
 9    uniform = np.ones(n) / n
10    return ((1 - alpha) * exp_probs + alpha * uniform).tolist()
def beta_dist(n, a, b): 
12def beta_dist( n, a, b):
13	points = np.linspace(0, 1, n + 2 )[ 1:-1] 
14	probs  = beta.pdf(points, a, b)
15	return probs / probs.sum()
def dirichlet_dist(n, alpha): 
17def dirichlet_dist(n, alpha):
18	return np.random.dirichlet(np.ones(n) * alpha)
def uniform_dist(n): 
20def uniform_dist( n ):
21	return np.ones(n) / n