ciela

ciela

import time
from concurrent.futures import ProcessPoolExecutor, as_completed
import random

import jsonlines as jl


class MyExecutor:
    def write(self, text: str):
        time.sleep(random.randint(0, 2))

ciela

from typing import Tuple

import torch
import torch.nn as nn


class GradientReversalFunction(torch.autograd.Function):
    
    @staticmethod
    def forward(ctx, input_forward: torch.Tensor, scale: torch.Tensor) -> torch.Tensor:

ciela

import collection
log.debug(len(train_loader))
labels = []
for i, data in enumerate(train_loader):
    _, label = data
    labels.append(label.item())
labels_counter = collections.Counter(labels)
log.debug(labels_counter)
labels = []
log.debug(len(val_loader))

ciela

import torch
from torch import nn as nn
from torch import autograd as ag

data = ag.Variable(torch.Tensor(torch.randn([100, 100])))
labels = ag.Variable(torch.Tensor(torch.randn([100, 100])))

multi_label_soft_margin = nn.MultiLabelSoftMarginLoss()
print(multi_label_soft_margin(data, labels))

ciela

# MultiLabelSoftMarginLoss only
ml_criterion = nn.MultiLabelSoftMarginLoss()
## torch.randn
data, labels = Variable(torch.randn([1, 5])), Variable(torch.randn([1, 5]))
print(data.data, labels.data)
print(ml_criterion(data, labels))
## fixed FloatTensor
data, labels = Variable(torch.FloatTensor([1, 50, 100, 50, 1])), Variable(torch.FloatTensor([0, 0, 1, 0, 0]))
print(data.data, labels.data)
print(ml_criterion(data, labels))

ciela

#!/usr/bin/env python3
# coding: utf-8
# プログラミングのための確率統計 p.280 例題8.2 の numpy での実装
import numpy as np


# データ群
x1 = np.array([0, 5])
x2 = np.array([0, -5])
x3 = np.array([4, 3])

ciela

import sys
import numpy as np
import numpy.random as rnd
import matplotlib.pyplot as plt


A = 0.7
SEED = 10000
TRIAL = 100

ciela

#!/usr/bin/env python3
# coding: utf-8

import sys
import numpy as np


def make_data(n):
    """
    優先順位が行ごとにシャッフルされたn次正方行列を生成する

ciela

#!/usr/bin/env python3
# coding: utf-8

import numpy as np


NUM = 100000 # 試行回数
DOORS = np.array([1, 2, 3]) # ドア