I know some popular libraries among researcher focusing on deep reinforcement learning such as TensorFlow, Theano, Caffe, Torch and CNTK. I usually work on both CPU and GPU. I don't need much detail configuration and customization, but computational time is important. Can anyone recommend which library is best for my case?
Thank you in advance
I'd recommend Tensorflow. It has several levels of abstraction so you can pick the one that suits you best.
Thank you for all of your answer.
As for now, I am considering TensorFlow + Keras.
Since computation is concerned, I would recommend TensorFlow because it supports distributed computing and multiple GPUs.
You can search for OpenAI gym for some guide to this though it is linux based.
You can also look into MXNet. It is an open source deep learning library from Apache foundation https://mxnet.apache.org/
When compared to Tensorflow, MXNet is tested to be faster with far less resource utilization and it is also supported by Keras (https://medium.com/@julsimon/keras-shoot-out-tensorflow-vs-mxnet-51ae2b30a9c0). Also, you are not locked to Python API as in TF, you can use Python/Scala/R/Julia/C++/Perl APIs to write code. Further, it supported by AWS as their preffered DL framework (https://www.infoworld.com/article/3144025/cloud-computing/why-amazon-picked-mxnet-for-deep-learning.html)
- Badges
- Science topic