farrajota

local function ConvertBNcudnn2nn(net)
  local function ConvertModule(net)
    return net:replace(function(x)
        if torch.type(x) == 'cudnn.BatchNormalization' then
          return cudnn.convert(x, nn)
        else
          return x
        end
    end)
  end

farrajota

function svhn_convert_json(path)
% convert .mat to json files

%% Create .json file

if isempty(path)
    error('Must specify a path to the dataset')
end

farrajota

local function dimnarrow(x,sz,pad,dim)
    local xn = x
    for i=1,x:dim() do
        if i > dim then
            xn = xn:narrow(i,pad[i]+1,sz[i])
        end
    end
    return xn
end

farrajota

-- Use three (consecutive) points to fit a 1D parabola
-- and return its maximum value. To achieve sub-pixel
-- precision for a 2D maxima (x,y), just fit the parabola over the
-- x and y coordinates separately with two neighboring points.
local function fitParabola(x1,x2,x3,y1,y2,y3)
  local x1_sqr = x1*x1
  local x2_sqr = x2*x2
  local x3_sqr = x3*x3
  
  local div = (x1_sqr-x1*(x2+x3)+x2*x3)*(x2-x3)

farrajota

local SplitTable, parent = torch.class('nn.MySplitTable', 'nn.Module')

function SplitTable:__init(dimension, nTensors)
   parent.__init(self)
   self.dimension = dimension
   self.nTensors = nTensors
   self.joinTable = nn.JoinTable(dimension)
end

function SplitTable:getSize(input)

farrajota

The Best of the Best Practices (BOBP) Guide for Python

A "Best of the Best Practices" (BOBP) guide to developing in Python.

In General

Values

• "Build tools for others that you want to be built for you." - Kenneth Reitz
• "Simplicity is alway better than functionality." - Pieter Hintjens

farrajota

vs Questions

---

1. http://stackoverflow.com/questions/804115 (rebase vs merge).
2. https://www.atlassian.com/git/tutorials/merging-vs-rebasing (rebase vs merge)
3. https://www.atlassian.com/git/tutorials/undoing-changes/ (reset vs checkout vs revert)
4. http://stackoverflow.com/questions/2221658 (HEAD^ vs HEAD~) (See git rev-parse)
5. http://stackoverflow.com/questions/292357 (pull vs fetch)
6. http://stackoverflow.com/questions/39651 (stash vs branch)
7. http://stackoverflow.com/questions/8358035 (reset vs checkout vs revert)
8. http://stackoverflow.com/questions/5798930 (git reset vs git rm --cached)

farrajota

Missing data

A Simple Example of a Missing Data Analysis Understanding the Reasons Leading to Missing Data Ignorable Missing Data Other Types of Missing Data Processes Examining the Patterns of Missing Data Diagnosing the Randomness of the Missing Data Process

Rules of Thumb

farrajota

data science checklist

---

Step 1

---

Data loading# data science checklist

---

Step 1

farrajota

When imputation should not be used

• If data are MCAR, imputation may not be not needed.
• If missingness is due to unmeasured variables related to the dependent variable, data are MNAR and should not be imputed.
• Imputation assumes data are MAR and should not be used with sparse data. Sparse data occur when missingness is non-random, such as a shopping cart survey of items purchased (coded 1) or not purchased (coded 0), because the null response (0) is non-random, due to unmeasured factors possibly not even known to the shopper.
• Imputation should not be used to impute all the data for a subject
• Imputation should not be used for a missing value for a given observation if that observation is also missing values on predictively critical variables in the imputation model. While this is difficult to check for each value to be imputed, a table of missing value patterns will show how many cases missing on a given variable also have missing values on other variables. In some cases this may lead a researcher