thouis

import sys
import twitter
import time

# see https://python-twitter.readthedocs.io/en/latest/getting_started.html

api = twitter.Api(consumer_key=CONSUMER_KEY,
                  consumer_secret=CONSUMER_SECRET,
                  access_token_key=ACCESS_TOKEN,
                  access_token_secret=ACCESS_SECRET)

thouis

thouis

Keybase proof

I hereby claim:

• I am thouis on github.
• I am thouis (https://keybase.io/thouis) on keybase.
• I have a public key whose fingerprint is A4F6 3AF2 C0E4 8CAE 7501 2BDE 2904 0D09 3557 95A4

To claim this, I am signing this object:

thouis

def depth_to_space(x, blocksize=2):
    b, k, d, r, c = x.shape
    r1 = x.reshape((b, k // (blocksize ** 2), blocksize, blocksize, d, r, c))
    r2 = r1.transpose(0, 1, 4, 5, 2, 6, 3)
    return r2.reshape((b, k // (blocksize ** 2), d, r * blocksize, c * blocksize))

thouis

# coding=utf-8

import numpy as np
import scipy.sparse as sparse

# Evaluation code courtesy of Juan Nunez-Iglesias, taken from
# https://github.com/janelia-flyem/gala/blob/master/gala/evaluate.py

def adapted_rand(seg, gt, all_stats=False):
    """Compute Adapted Rand error as defined by the SNEMI3D contest [1]

thouis

import tensorflow as tf
from tensorflow.python import control_flow_ops

# from http://stackoverflow.com/a/34634291
def batch_norm(x, n_out, phase_train, scope='bn', affine=True):
    """
    Batch normalization on convolutional maps.
    Args:
        x:           Tensor, 4D BHWD input maps
        n_out:       integer, depth of input maps

thouis

// Use Gists to store code you would like to remember later on
console.log(window); // log the "window" object to the console

thouis

import sys
import h5py
import numpy as np

def pygame(h5, name, board, moves):
    for idx in range(board.shape[0]):
        for plane in range(board.shape[1]):
            if plane < 44:
                # different board orderings
                f = board[idx, plane, ...].T[:, ::-1]

thouis

thouis

local grad = require 'autograd'
local nn = require 'nn'

local W1 = torch.FloatTensor(1, 1, 10, 10):normal()
local pooler = nn.SpatialMaxPooling(2, 2, 2, 2)
local unpooler = nn.SpatialMaxUnpooling(pooler)

-- not sure why these lines are necessary.
pooler.indices = torch.FloatTensor()
pooler.output = torch.FloatTensor()