A couple weeks ago I was fairly new to neural networks, and tried a bunch of frameworks.
Most of them had unclear documentation, were hard to install (multiple build issues, undeclared depencies, etc) and simply did not produce any results. Why? Because I messed up, but as someone who isn't an expert at neural networks there is absolutely no way to find out what's going wrong.
Keras was easy to install, had pretty good documentation, and produced results immediately, with further tweaking improving them. I can highly recommend it, and it's nice to see that it's now available for R as well.
[+] [-] nightcracker|8 years ago|reply
Most of them had unclear documentation, were hard to install (multiple build issues, undeclared depencies, etc) and simply did not produce any results. Why? Because I messed up, but as someone who isn't an expert at neural networks there is absolutely no way to find out what's going wrong.
Keras was easy to install, had pretty good documentation, and produced results immediately, with further tweaking improving them. I can highly recommend it, and it's nice to see that it's now available for R as well.