brandonwillard

import torch

import outlines.models as models

from outlines.text.generate.regex import choice
from outlines.text.generate.continuation import continuation
from outlines.text.generate.sample import greedy


def make_greedy_tracker(generator):

brandonwillard

from typing import Tuple, Union

import numpy as np


def is_basic_idx(x):
    return isinstance(x, (slice, type(None)))


def expand_indices(

brandonwillard

import numba as nb
import numpy as np
from numba.core.datamodel.models import StructModel
from numba.core.datamodel.registry import register_default
from numba.core.extending import intrinsic, overload, overload_method


@nb.njit
def add(x, y):
    return x + y

brandonwillard

import aesara
import aesara.tensor as at
import numpy as np
from aeppl.joint_logprob import factorized_joint_logprob

# aesara.config.compute_test_value = "warn"

srng = at.random.RandomStream(seed=2320)

N = 100

brandonwillard

NumPy C API Testing

brandonwillard

brandonwillard

from collections import defaultdict


import aesara.tensor as aet
from aesara import config
from aesara.compile import optdb
from aesara.graph.basic import ancestors, applys_between, graph_inputs
from aesara.graph.fg import FunctionGraph
from aesara.graph.opt import (
    EquilibriumOptimizer,

brandonwillard

import time
from datetime import date, timedelta

import arviz as az
import formulaic
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import scipy

brandonwillard

Computing log-likehood graphs for src_python[:eval never]{Scan}s

brandonwillard

import numpy as np


sigma2_true = 2.0


def simulate_regression(N=100, M=1000):
    X = np.random.normal(size=(N, M))
    beta_true = np.random.normal(size=(M,))
    y = X.dot(beta_true) + np.random.normal(0, np.sqrt(sigma2_true), size=(N,))