how-to

I recently got myself to start using Python on Windows, whereas till very recently I had been working on Python only from Ubuntu.

I am sure I am late in realizing this, but installing Tensorflow was just so easy!

If you’ve tried installing Tensorflow for Windows when it was first introduced, and gave up back then – try again. The method I’d recommend would be using Anaconda Navigator from where you first open a terminal (figure below). You may notice that I already have a tensorflow environment set up, since I am writing this post after installation.

Once you have terminal open, create a conda environment named tensorflow by invoking the following command, with your python version:

C:> conda create -n tensorflow python=3.6

That’s all! You should now have tensorflow ready to use.

For more details, you could always go here. Otherwise, the screenshot below gives a sense of what it takes.

I recently bought a new laptop and began installing essential software all over again, including R of course! And I wanted all the libraries that I had installed in my previous laptop. Instead of installing libraries one by one all over again, I did the following:

Step 1: Save a list of packages installed in your old computing device (from your old device).

installed <- as.data.frame(installed.packages()) write.csv(installed, 'installed_previously.csv')

This saves information on installed packages in a csv file named installed_previously.csv. Now copy or e-mail this file to your new device and access it from your working directory in R.

Step 2: Create a list of libraries from your old list that were not already installed when you freshly download R (from your new device).

installedPreviously <- read.csv('installed_previously.csv') baseR <- as.data.frame(installed.packages()) toInstall <- setdiff(installedPreviously, baseR)

We now have a list of libraries that were installed in your previous computer in addition to the R packages already installed when you download R. So you now go ahead and install these libraries.

Step 3: Download this list of libraries.

install.packages(toInstall)

That’s it. Save yourself the trouble installing packages one-by-one all over again.