speedcell4

import torch
from torch import Tensor
from torch.testing import assert_close


def fft(tensor: Tensor, pi: Tensor = torch.acos(torch.tensor(-1.))) -> Tensor:
    n = tensor.size()[-1]
    index = torch.arange(n, dtype=torch.float32)
    fourier = torch.exp(index[:, None] * index[None, :] * pi * -2j / n)

speedcell4

import math

import torch
from torch import Tensor
from torch import nn
from torch.nn import functional as F
from torch.nn import init


class ComplexLinear(nn.Module):

speedcell4

# CUDA
export CUDA_HOME=/usr/local/cuda-9.0
export CUDA_ROOT=/usr/local/cuda-9.0
export CUDA_PATH=/usr/local/cuda-9.0
export PATH=${CUDA_HOME}/bin:${PATH}
export LIBRARY_PATH=${CUDA_HOME}/lib64:${LIBRARY_PATH}
export LD_LIBRARY_PATH=${CUDA_HOME}/lib64:${LD_LIBRARY_PATH}

# cuDNN

speedcell4

speedcell4

pip
wheel
setuptools
cython
numpy
scipy
matplotlib
pandas
gensim
sklearn

speedcell4

% lstlisting coq style (inspired from https://github.com/cmc333333/forensics-thesis)
% lstlisting coq style (inspired from a file of Assia Mahboubi)
%
\lstdefinelanguage{coq}{ 
%
% Anything betweeen $ becomes LaTeX math mode
mathescape=true,
%
% Comments may or not include Latex commands
texcl=false, 

speedcell4

import chainer.functions as F
import numpy as np
from chainer import Variable, Function
from chainer import cuda


class Scatter(Function):
    def __init__(self, row: int, col: int):
        self.row = row
        self.col = col

speedcell4

#include <iostream>

using namespace std;

int binary_search(int *a, int value, int left, int right) {
  if (left >= right) return -1;
  else {
    int mid = (left + right) / 2;
    if (a[mid] == value) return mid;
    else if (a[mid] > value) return binary_search(a, value, left, mid);

speedcell4

import tensorflow as tf


class Summary(object):
    def __init__(self, logdir: str):
        self._placeholders = {}
        self._observations = {}

        self._session = tf.Session()
        self._writer = tf.summary.FileWriter(logdir=logdir)

speedcell4

from chainer import Chain, report


class MultiTaskClassifier(Chain):
    def __init__(self, task1, task2, loss1, loss2, acc1, acc2, lambda_=1.0):
        """
        :param task1: `Chain` of task 1
        :param task2: `Chain` of task 2
        :param loss1: loss function of task1
        :param loss2: loss function of task2