Skip to content
No description, website, or topics provided.
Branch: master
Clone or download
Pull request Compare This branch is 1 commit ahead of shawngraham:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
images
Affinity Propagation.ipynb
Getting Started.ipynb
README.md
classify_images.py
cluster_vectors.py
download.png
find-similar-images.ipynb
install.R
requirements.txt
runtime.txt
terminalprompt.png
toolbar.png

README.md

Identifying-Similar-Images-with-TensorFlow

You can launch this jupyter notebook in an executable environment by clicking on the button below.

Binder

DOI

It's based on Douglas Duhaime's tutorial identifying similar images with tensorflow which is very clearly laid out and explained. In this notebook, I've compressed the steps. The 'images' directory only has 25 images in it for demo purposes. In Duhaime's post, the images folder (that you grab with wget has about 2000 images in it).

UPDATE the requirements.txt file should list the python bits and pieces need to be installed, so that should save us from having to !pip install. There is a package called RPy2 that can be loaded this way, that then lets us do bits and bobs of R mixed in with python - see this post.

to get the R kernel so that I can have just an R notebook we need the runtime.txt file. Inside that, we have a single line that says r-2018-04-01 (where the date stamp comes from this).

to preload and R packages, put the name of the package in the install.R file in normal R syntax, like so: install.packages("ggplot2").

And there we have it! Sometimes the binder fails to launch. Just reload. Keep reloading.

You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.