vene

"""
Check sampling from a sequential CRF model.
 - Code is for n_states=2 but the strategy is general.
 - TODO; cythonize or numbaize
 - TODO; write general impl for clarity
"""

# author: vlad niculae <vlad@vene.ro>
# license: MIT

vene

# compare FY entmax losses with (log)-likelihood objectives
# author: vlad niculae

import numpy as np
import torch
import matplotlib.pyplot as plt

from entmax import entmax_bisect, entmax_bisect_loss

def main(alpha=1.5):

vene

"""
Generate data triples (x, y, z) for deterministic classification p(y | x; z)

Generative story:

    Given: n_clusters; for each cluster:
      - a cluster center (mean) center[z]
      - a linear model y=sign(w[z] * x + b[z])

    pick z from uniform Categorical(n_clusters)

vene

import numpy as np


def numeric_grad(f, x, eps=1e-3):

    grad = np.zeros_like(x)

    for i in range(x.shape[0]):
        v = np.zeros_like(x)
        v[i] = 1

vene

#pragma once

/* MLFlow interaction */

#include <string>
#include <chrono>

#include <cpr/cpr.h>
#include <nlohmann/json.hpp>

vene

# Author: vlad niculae <vlad@vene.ro>
# License: 3-clause BSD

import numpy as np
import cvxpy as cx

from copt.constraint import SimplexConstraint
from copt.splitting import minimize_primal_dual

vene

# Author: Vlad Niculae <vlad@vene.ro>
# License: Simplified BSD

import numpy as np

try:
    from numba import jit
except ImportError:
    print("numba not available")
    def jit(nopython):

vene

DYNET_PATH ?= /home/vlad/code/dynet
EIGEN_PATH ?= /home/vlad/code/eigen

CC = g++
DEBUG = -g
INCLUDES = -I$(DYNET_PATH) -I$(EIGEN_PATH) 
LIBS = -L$(DYNET_PATH)/build-cuda/dynet/ 
CFLAGS = -O3 -Wall -Wno-sign-compare -Wno-int-in-bool-context -c -fmessage-length=0 $(INCLUDES) -DEIGEN_FAST_MATH -fPIC -fno-finite-math-only -Wno-missing-braces -std=c++11 -funroll-loops
LFLAGS = $(LIBS) -ldynet

vene

import numpy as np

from sklearn.datasets import load_iris
from sklearn.linear_model import LogisticRegression
from sklearn.utils import shuffle

# get shuffled iris data
X, y = load_iris(return_X_y=True)
X, y = shuffle(X, y, random_state=0)

vene

import numpy as np
from sklearn.metrics import precision_recall_curve, average_precision_score


def naive_interpolated_precision(y_true, y_scores):
    precisions, recalls, _ = precision_recall_curve(y_true, y_scores)
    interp_precisions = []

    # the final point
    precisions = precisions[:-1]