Will this project continue?

[cyxtj]

I found this project several days ago. I'm trying to find a fast framework to implement my special LSTM.
Theano is faster than tensorflow on a single GPU. But I want even faster, so I find this repo.
What worries me is that the last commit of platoon was a month ago, and the document is very limited.
So should I stick with platoon to speed up my LSTM or switch to other frameworks like torch, especially in the perspective of a long term, like more than one year?

[cshanbo]

Hi @cyxtj
Following the lstm example will help.
Experimental results show that platoon now works fine in single-node, multi-GPUs scenario. You can implement your own controller and worker according to the examples.
Good luck.