Hi Rajani,

I'd suggest: https://github.com/apache/incubator-mxnet

and a tutorial https://www.r-bloggers.com/image-recognition-tutorial-in-r-using-deep-convolutional-neural-networks-mxnet-package/

Hi Rajani,

I'd suggest: https://github.com/apache/incubator-mxnet

and a tutorial https://www.r-bloggers.com/image-recognition-tutorial-in-r-using-deep-convolutional-neural-networks-mxnet-package/

Dear Dr Kamath,

I entirely agree with Dr Maskeliunas' recommendations. In addition, the following website might be useful too:

http://www.rblog.uni-freiburg.de/2017/02/07/deep-learning-in-r/

Yours sincerely,

Rainer

Hi!

In my opinion, Keras with back end tensorflow support and MXNet would be the best available R packages (for deep learning API compatible with R ) to deal with CNN.

https://keras.rstudio.com/

https://mxnet.incubator.apache.org/api/r/index.html

Along side of CPU processing they both support faster GPU processing.

Here is one nice book " Deep Learning with R " written by " François Chollet with J. J. Allaire " which will be release on upcoming January 2018. I believe contents of this book would be enough to learn deep learning in R with Keras and Tensorflow implementation.

https://www.manning.com/books/deep-learning-with-r

You can get a nice preview of one of its chapter's implementation with image classification in the following website

https://www.r-bloggers.com/image-classification-on-small-datasets-with-keras/

Hoping it will help you

Thank you all, I implemented CNN with mxnet in R

https://www.r-bloggers.com/convolutional-neural-networks-in-r/

https://www.kaggle.com/cd12631/using-r-and-tensorflow-to-build-cnn

Please follow the above links